• Posts tagged "itemCF"

Blog Archives

Mahout推荐算法API详解

Hadoop家族系列文章,主要介绍Hadoop家族产品,常用的项目包括Hadoop, Hive, Pig, HBase, Sqoop, Mahout, Zookeeper, Avro, Ambari, Chukwa,新增加的项目包括,YARN, Hcatalog, Oozie, Cassandra, Hama, Whirr, Flume, Bigtop, Crunch, Hue等。

从2011年开始,中国进入大数据风起云涌的时代,以Hadoop为代表的家族软件,占据了大数据处理的广阔地盘。开源界及厂商,所有数据软件,无一不向Hadoop靠拢。Hadoop也从小众的高富帅领域,变成了大数据开发的标准。在Hadoop原有技术基础之上,出现了Hadoop家族产品,通过“大数据”概念不断创新,推出科技进步。

作为IT界的开发人员,我们也要跟上节奏,抓住机遇,跟着Hadoop一起雄起!

关于作者:

  • 张丹(Conan), 程序员Java,R,PHP,Javascript
  • weibo:@Conan_Z
  • blog: http://blog.fens.me
  • email: bsspirit@gmail.com

转载请注明出处:
http://blog.fens.me/mahout-recommendation-api

mahout-Recommendation

前言

用Mahout来构建推荐系统,是一件既简单又困难的事情。简单是因为Mahout完整地封装了“协同过滤”算法,并实现了并行化,提供非常简单的API接口;困难是因为我们不了解算法细节,很难去根据业务的场景进行算法配置和调优。

本文将深入算法API去解释Mahout推荐算法底层的一些事。

目录

  1. Mahout推荐算法介绍
  2. 算法评判标准:召回率与准确率
  3. Recommender.java的API接口
  4. 测试程序:RecommenderTest.java
  5. 基于用户的协同过滤算法UserCF
  6. 基于物品的协同过滤算法ItemCF
  7. SlopeOne算法
  8. KNN Linear interpolation item–based推荐算法
  9. SVD推荐算法
  10. Tree Cluster-based 推荐算法
  11. Mahout推荐算法总结

1. Mahout推荐算法介绍

Mahoutt推荐算法,从数据处理能力上,可以划分为2类:

  • 单机内存算法实现
  • 基于Hadoop的分步式算法实现

1). 单机内存算法实现

单机内存算法实现:就是在单机下运行的算法,是由cf.taste项目实现的,像我的们熟悉的UserCF,ItemCF都支持单机内存运行,并且参数可以灵活配置。单机算法的基本实例,请参考文章:用Maven构建Mahout项目

单机内存算法的问题在于,受限于单机的资源。对于中等规模的数据,像1G,10G的数据量,有能力进行计算,但是超过100G的数据量,对于单机来说是不可能完成的任务。

2). 基于Hadoop的分步式算法实现

基于Hadoop的分步式算法实现:就是把单机内存算法并行化,把任务分散到多台计算机一起运行。Mahout提供了ItemCF基于Hadoop并行化算法实现。基于Hadoop的分步式算法实现,请参考文章:
Mahout分步式程序开发 基于物品的协同过滤ItemCF

分步式并行算法的问题在于,如何让单机算法并行化。在单机算法中,我们只需要考虑算法,数据结构,内存,CPU就够了,但是分步式算法还要额外考虑很多的情况,比如多节点的数据合并,数据排序,网路通信的效率,节点宕机重算,数据分步式存储等等的很多问题。

2. 算法评判标准:召回率(recall)与查准率(precision)

Mahout提供了2个评估推荐器的指标,查准率和召回率(查全率),这两个指标是搜索引擎中经典的度量方法。

precision_recall


         相关 不相关
检索到     A    C
未检索到   B    D
  • A:检索到的,相关的 (搜到的也想要的)
  • B:未检索到的,但是相关的 (没搜到,然而实际上想要的)
  • C:检索到的,但是不相关的 (搜到的但没用的)
  • D:未检索到的,也不相关的 (没搜到也没用的)

被检索到的越多越好,这是追求“查全率”,即A/(A+B),越大越好。
被检索到的,越相关的越多越好,不相关的越少越好,这是追求“查准率”,即A/(A+C),越大越好。

在大规模数据集合中,这两个指标是相互制约的。当希望索引出更多的数据的时候,查准率就会下降,当希望索引更准确的时候,会索引更少的数据。

3. Recommender的API接口

1). 系统环境:

  • Win7 64bit
  • Java 1.6.0_45
  • Maven 3
  • Eclipse Juno Service Release 2
  • Mahout 0.8
  • Hadoop 1.1.2

2). Recommender接口文件:
org.apache.mahout.cf.taste.recommender.Recommender.java

mahout-Recommender-class

接口中方法的解释:

  • recommend(long userID, int howMany): 获得推荐结果,给userID推荐howMany个Item
  • recommend(long userID, int howMany, IDRescorer rescorer): 获得推荐结果,给userID推荐howMany个Item,可以根据rescorer对结构重新排序。
  • estimatePreference(long userID, long itemID): 当打分为空,估计用户对物品的打分
  • setPreference(long userID, long itemID, float value): 赋值用户,物品,打分
  • removePreference(long userID, long itemID): 删除用户对物品的打分
  • getDataModel(): 提取推荐数据

通过Recommender接口,我可以猜出核心算法,应该会在子类的estimatePreference()方法中进行实现。

3). 通过继承关系到Recommender接口的子类:

mahout-Recommender-hierarchy

推荐算法实现类:

  • GenericUserBasedRecommender: 基于用户的推荐算法
  • GenericItemBasedRecommender: 基于物品的推荐算法
  • KnnItemBasedRecommender: 基于物品的KNN推荐算法
  • SlopeOneRecommender: Slope推荐算法
  • SVDRecommender: SVD推荐算法
  • TreeClusteringRecommender:TreeCluster推荐算法

下面将分别介绍每种算法的实现。

4. 测试程序:RecommenderTest.java

测试数据集:item.csv


1,101,5.0
1,102,3.0
1,103,2.5
2,101,2.0
2,102,2.5
2,103,5.0
2,104,2.0
3,101,2.5
3,104,4.0
3,105,4.5
3,107,5.0
4,101,5.0
4,103,3.0
4,104,4.5
4,106,4.0
5,101,4.0
5,102,3.0
5,103,2.0
5,104,4.0
5,105,3.5
5,106,4.0

测试程序:org.conan.mymahout.recommendation.job.RecommenderTest.java


package org.conan.mymahout.recommendation.job;

import java.io.IOException;
import java.util.List;

import org.apache.mahout.cf.taste.common.TasteException;
import org.apache.mahout.cf.taste.eval.RecommenderBuilder;
import org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;
import org.apache.mahout.cf.taste.model.DataModel;
import org.apache.mahout.cf.taste.recommender.RecommendedItem;
import org.apache.mahout.common.RandomUtils;

public class RecommenderTest {

    final static int NEIGHBORHOOD_NUM = 2;
    final static int RECOMMENDER_NUM = 3;

    public static void main(String[] args) throws TasteException, IOException {
        RandomUtils.useTestSeed();
        String file = "datafile/item.csv";
        DataModel dataModel = RecommendFactory.buildDataModel(file);
        slopeOne(dataModel);
    }

    public static void userCF(DataModel dataModel) throws TasteException{}
    public static void itemCF(DataModel dataModel) throws TasteException{}
    public static void slopeOne(DataModel dataModel) throws TasteException{}

    ...

每种算法都一个单独的方法进行算法测试,如userCF(),itemCF(),slopeOne()….

5. 基于用户的协同过滤算法UserCF

基于用户的协同过滤,通过不同用户对物品的评分来评测用户之间的相似性,基于用户之间的相似性做出推荐。简单来讲就是:给用户推荐和他兴趣相似的其他用户喜欢的物品。

举例说明:

image015

基于用户的 CF 的基本思想相当简单,基于用户对物品的偏好找到相邻邻居用户,然后将邻居用户喜欢的推荐给当前用户。计算上,就是将一个用户对所有物品的偏好作为一个向量来计算用户之间的相似度,找到 K 邻居后,根据邻居的相似度权重以及他们对物品的偏好,预测当前用户没有偏好的未涉及物品,计算得到一个排序的物品列表作为推荐。图 2 给出了一个例子,对于用户 A,根据用户的历史偏好,这里只计算得到一个邻居 – 用户 C,然后将用户 C 喜欢的物品 D 推荐给用户 A。

上文中图片和解释文字,摘自: https://www.ibm.com/developerworks/cn/web/1103_zhaoct_recommstudy2/

算法API: org.apache.mahout.cf.taste.impl.recommender.GenericUserBasedRecommender


  @Override
  public float estimatePreference(long userID, long itemID) throws TasteException {
    DataModel model = getDataModel();
    Float actualPref = model.getPreferenceValue(userID, itemID);
    if (actualPref != null) {
      return actualPref;
    }
    long[] theNeighborhood = neighborhood.getUserNeighborhood(userID);
    return doEstimatePreference(userID, theNeighborhood, itemID);
  }

 protected float doEstimatePreference(long theUserID, long[] theNeighborhood, long itemID) throws TasteException {
    if (theNeighborhood.length == 0) {
      return Float.NaN;
    }
    DataModel dataModel = getDataModel();
    double preference = 0.0;
    double totalSimilarity = 0.0;
    int count = 0;
    for (long userID : theNeighborhood) {
      if (userID != theUserID) {
        // See GenericItemBasedRecommender.doEstimatePreference() too
        Float pref = dataModel.getPreferenceValue(userID, itemID);
        if (pref != null) {
          double theSimilarity = similarity.userSimilarity(theUserID, userID);
          if (!Double.isNaN(theSimilarity)) {
            preference += theSimilarity * pref;
            totalSimilarity += theSimilarity;
            count++;
          }
        }
      }
    }
    // Throw out the estimate if it was based on no data points, of course, but also if based on
    // just one. This is a bit of a band-aid on the 'stock' item-based algorithm for the moment.
    // The reason is that in this case the estimate is, simply, the user's rating for one item
    // that happened to have a defined similarity. The similarity score doesn't matter, and that
    // seems like a bad situation.
    if (count <= 1) {
      return Float.NaN;
    }
    float estimate = (float) (preference / totalSimilarity);
    if (capper != null) {
      estimate = capper.capEstimate(estimate);
    }
    return estimate;
  }

测试程序:


    public static void userCF(DataModel dataModel) throws TasteException {
        UserSimilarity userSimilarity = RecommendFactory.userSimilarity(RecommendFactory.SIMILARITY.EUCLIDEAN, dataModel);
        UserNeighborhood userNeighborhood = RecommendFactory.userNeighborhood(RecommendFactory.NEIGHBORHOOD.NEAREST, userSimilarity, dataModel, NEIGHBORHOOD_NUM);
        RecommenderBuilder recommenderBuilder = RecommendFactory.userRecommender(userSimilarity, userNeighborhood, true);

        RecommendFactory.evaluate(RecommendFactory.EVALUATOR.AVERAGE_ABSOLUTE_DIFFERENCE, recommenderBuilder, null, dataModel, 0.7);
        RecommendFactory.statsEvaluator(recommenderBuilder, null, dataModel, 2);

        LongPrimitiveIterator iter = dataModel.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = recommenderBuilder.buildRecommender(dataModel).recommend(uid, RECOMMENDER_NUM);
            RecommendFactory.showItems(uid, list, true);
        }
    }

程序输出:


AVERAGE_ABSOLUTE_DIFFERENCE Evaluater Score:1.0
Recommender IR Evaluator: [Precision:0.5,Recall:0.5]
uid:1,(104,4.333333)(106,4.000000)
uid:2,(105,4.049678)
uid:3,(103,3.512787)(102,2.747869)
uid:4,(102,3.000000)

用R语言重写UserCF的实现,请参考文章:用R解析Mahout用户推荐协同过滤算法(UserCF)

6. 基于物品的协同过滤算法ItemCF

基于item的协同过滤,通过用户对不同item的评分来评测item之间的相似性,基于item之间的相似性做出推荐。简单来讲就是:给用户推荐和他之前喜欢的物品相似的物品。

举例说明:

image017

基于物品的 CF 的原理和基于用户的 CF 类似,只是在计算邻居时采用物品本身,而不是从用户的角度,即基于用户对物品的偏好找到相似的物品,然后根据用户的历史偏好,推荐相似的物品给他。从计算的角度看,就是将所有用户对某个物品的偏好作为一个向量来计算物品之间的相似度,得到物品的相似物品后,根据用户历史的偏好预测当前用户还没有表示偏好的物品,计算得到一个排序的物品列表作为推荐。图 3 给出了一个例子,对于物品 A,根据所有用户的历史偏好,喜欢物品 A 的用户都喜欢物品 C,得出物品 A 和物品 C 比较相似,而用户 C 喜欢物品 A,那么可以推断出用户 C 可能也喜欢物品 C。

上文中图片和解释文字,摘自: https://www.ibm.com/developerworks/cn/web/1103_zhaoct_recommstudy2/

算法API: org.apache.mahout.cf.taste.impl.recommender.GenericItemBasedRecommender


  @Override
  public float estimatePreference(long userID, long itemID) throws TasteException {
    PreferenceArray preferencesFromUser = getDataModel().getPreferencesFromUser(userID);
    Float actualPref = getPreferenceForItem(preferencesFromUser, itemID);
    if (actualPref != null) {
      return actualPref;
    }
    return doEstimatePreference(userID, preferencesFromUser, itemID);
  }

protected float doEstimatePreference(long userID, PreferenceArray preferencesFromUser, long itemID)
    throws TasteException {
    double preference = 0.0;
    double totalSimilarity = 0.0;
    int count = 0;
    double[] similarities = similarity.itemSimilarities(itemID, preferencesFromUser.getIDs());
    for (int i = 0; i < similarities.length; i++) {
      double theSimilarity = similarities[i];
      if (!Double.isNaN(theSimilarity)) {
        // Weights can be negative!
        preference += theSimilarity * preferencesFromUser.getValue(i);
        totalSimilarity += theSimilarity;
        count++;
      }
    }
    // Throw out the estimate if it was based on no data points, of course, but also if based on
    // just one. This is a bit of a band-aid on the 'stock' item-based algorithm for the moment.
    // The reason is that in this case the estimate is, simply, the user's rating for one item
    // that happened to have a defined similarity. The similarity score doesn't matter, and that
    // seems like a bad situation.
    if (count <= 1) {
      return Float.NaN;
    }
    float estimate = (float) (preference / totalSimilarity);
    if (capper != null) {
      estimate = capper.capEstimate(estimate);
    }
    return estimate;
  }

测试程序:


    public static void itemCF(DataModel dataModel) throws TasteException {
        ItemSimilarity itemSimilarity = RecommendFactory.itemSimilarity(RecommendFactory.SIMILARITY.EUCLIDEAN, dataModel);
        RecommenderBuilder recommenderBuilder = RecommendFactory.itemRecommender(itemSimilarity, true);

        RecommendFactory.evaluate(RecommendFactory.EVALUATOR.AVERAGE_ABSOLUTE_DIFFERENCE, recommenderBuilder, null, dataModel, 0.7);
        RecommendFactory.statsEvaluator(recommenderBuilder, null, dataModel, 2);

        LongPrimitiveIterator iter = dataModel.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = recommenderBuilder.buildRecommender(dataModel).recommend(uid, RECOMMENDER_NUM);
            RecommendFactory.showItems(uid, list, true);
        }
    }

程序输出:


AVERAGE_ABSOLUTE_DIFFERENCE Evaluater Score:0.8676552772521973
Recommender IR Evaluator: [Precision:0.5,Recall:1.0]
uid:1,(105,3.823529)(104,3.722222)(106,3.478261)
uid:2,(106,2.984848)(105,2.537037)(107,2.000000)
uid:3,(106,3.648649)(102,3.380000)(103,3.312500)
uid:4,(107,4.722222)(105,4.313953)(102,4.025000)
uid:5,(107,3.736842)

7. SlopeOne算法

这个算法在mahout-0.8版本中,已经被@Deprecated。

SlopeOne是一种简单高效的协同过滤算法。通过均差计算进行评分。SlopeOne论文下载(PDF)

1). 举例说明:
用户X,Y,Z,对于物品A,B进行打分,如下表,求Z对B的打分是多少?

slopeone

Slope one算法认为:平均值可以代替某两个未知个体之间的打分差异,事物A对事物B的平均差是:((5 - 4) + (4 - 2)) / 2 = 1.5,就得到Z对B的打分是,3-1.5 = 1.5。

Slope one算法将用户的评分之间的关系看作简单的线性关系:

Y = mX + b

2). 平均加权计算:
用户X,Y,Z,对于物品A,B,C进行打分,如下表,求Z对A的打分是多少?

slopeone2

  • 1. 计算A和B的平均差, ((5-3)+(3-4))/2=0.5
  • 2. 计算A和C的平均差, (5-2)/1=3
  • 3. Z对A的评分,通过AB得到, 2+0.5=2.5
  • 4. Z对A的评分,通过AC得到,5+3=8
  • 5. 通过加权平均计算Z对A的评分:A和B都有评价的用户数为2,A和C都有评价的用户数为1,权重为别是2和1, (2*2.5+1*8)/(2+1)=13/3=4.33

通过这种简单的方式,我们可以快速计算出一个评分项,完成推荐过程!

算法API: org.apache.mahout.cf.taste.impl.recommender.slopeone.SlopeOneRecommender


@Override
  public float estimatePreference(long userID, long itemID) throws TasteException {
    DataModel model = getDataModel();
    Float actualPref = model.getPreferenceValue(userID, itemID);
    if (actualPref != null) {
      return actualPref;
    }
    return doEstimatePreference(userID, itemID);
  }
  
  private float doEstimatePreference(long userID, long itemID) throws TasteException {
    double count = 0.0;
    double totalPreference = 0.0;
    PreferenceArray prefs = getDataModel().getPreferencesFromUser(userID);
    RunningAverage[] averages = diffStorage.getDiffs(userID, itemID, prefs);
    int size = prefs.length();
    for (int i = 0; i < size; i++) {
      RunningAverage averageDiff = averages[i];
      if (averageDiff != null) {
        double averageDiffValue = averageDiff.getAverage();
        if (weighted) {
          double weight = averageDiff.getCount();
          if (stdDevWeighted) {
            double stdev = ((RunningAverageAndStdDev) averageDiff).getStandardDeviation();
            if (!Double.isNaN(stdev)) {
              weight /= 1.0 + stdev;
            }
            // If stdev is NaN, then it is because count is 1. Because we're weighting by count,
            // the weight is already relatively low. We effectively assume stdev is 0.0 here and
            // that is reasonable enough. Otherwise, dividing by NaN would yield a weight of NaN
            // and disqualify this pref entirely
            // (Thanks Daemmon)
          }
          totalPreference += weight * (prefs.getValue(i) + averageDiffValue);
          count += weight;
        } else {
          totalPreference += prefs.getValue(i) + averageDiffValue;
          count += 1.0;
        }
      }
    }
    if (count <= 0.0) {
      RunningAverage itemAverage = diffStorage.getAverageItemPref(itemID);
      return itemAverage == null ? Float.NaN : (float) itemAverage.getAverage();
    } else {
      return (float) (totalPreference / count);
    }
  }

测试程序:


    public static void slopeOne(DataModel dataModel) throws TasteException {
        RecommenderBuilder recommenderBuilder = RecommendFactory.slopeOneRecommender();

        RecommendFactory.evaluate(RecommendFactory.EVALUATOR.AVERAGE_ABSOLUTE_DIFFERENCE, recommenderBuilder, null, dataModel, 0.7);
        RecommendFactory.statsEvaluator(recommenderBuilder, null, dataModel, 2);

        LongPrimitiveIterator iter = dataModel.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = recommenderBuilder.buildRecommender(dataModel).recommend(uid, RECOMMENDER_NUM);
            RecommendFactory.showItems(uid, list, true);
        }
    }

程序输出:


AVERAGE_ABSOLUTE_DIFFERENCE Evaluater Score:1.3333333333333333
Recommender IR Evaluator: [Precision:0.25,Recall:0.5]
uid:1,(105,5.750000)(104,5.250000)(106,4.500000)
uid:2,(105,2.286115)(106,1.500000)
uid:3,(106,2.000000)(102,1.666667)(103,1.625000)
uid:4,(105,4.976859)(102,3.509071)

8. KNN Linear interpolation item–based推荐算法

这个算法在mahout-0.8版本中,已经被@Deprecated。

算法来自论文:
This algorithm is based in the paper of Robert M. Bell and Yehuda Koren in ICDM '07.

(TODO未完)

算法API: org.apache.mahout.cf.taste.impl.recommender.knn.KnnItemBasedRecommender


@Override
  protected float doEstimatePreference(long theUserID, PreferenceArray preferencesFromUser, long itemID)
    throws TasteException {
    
    DataModel dataModel = getDataModel();
    int size = preferencesFromUser.length();
    FastIDSet possibleItemIDs = new FastIDSet(size);
    for (int i = 0; i < size; i++) {
      possibleItemIDs.add(preferencesFromUser.getItemID(i));
    }
    possibleItemIDs.remove(itemID);
    
    List mostSimilar = mostSimilarItems(itemID, possibleItemIDs.iterator(),
      neighborhoodSize, null);
    long[] theNeighborhood = new long[mostSimilar.size() + 1];
    theNeighborhood[0] = -1;
  
    List usersRatedNeighborhood = Lists.newArrayList();
    int nOffset = 0;
    for (RecommendedItem rec : mostSimilar) {
      theNeighborhood[nOffset++] = rec.getItemID();
    }
    
    if (!mostSimilar.isEmpty()) {
      theNeighborhood[mostSimilar.size()] = itemID;
      for (int i = 0; i < theNeighborhood.length; i++) {
        PreferenceArray usersNeighborhood = dataModel.getPreferencesForItem(theNeighborhood[i]);
        int size1 = usersRatedNeighborhood.isEmpty() ? usersNeighborhood.length() : usersRatedNeighborhood.size();
        for (int j = 0; j < size1; j++) {
          if (i == 0) {
            usersRatedNeighborhood.add(usersNeighborhood.getUserID(j));
          } else {
            if (j >= usersRatedNeighborhood.size()) {
              break;
            }
            long index = usersRatedNeighborhood.get(j);
            if (!usersNeighborhood.hasPrefWithUserID(index) || index == theUserID) {
              usersRatedNeighborhood.remove(index);
              j--;
            }
          }
        }
      }
    }

    double[] weights = null;
    if (!mostSimilar.isEmpty()) {
      weights = getInterpolations(itemID, theNeighborhood, usersRatedNeighborhood);
    }
    
    int i = 0;
    double preference = 0.0;
    double totalSimilarity = 0.0;
    for (long jitem : theNeighborhood) {
      
      Float pref = dataModel.getPreferenceValue(theUserID, jitem);
      
      if (pref != null) {
        double weight = weights[i];
        preference += pref * weight;
        totalSimilarity += weight;
      }
      i++;
      
    }
    return totalSimilarity == 0.0 ? Float.NaN : (float) (preference / totalSimilarity);
  }
  
}

测试程序:


    public static void itemKNN(DataModel dataModel) throws TasteException {
        ItemSimilarity itemSimilarity = RecommendFactory.itemSimilarity(RecommendFactory.SIMILARITY.EUCLIDEAN, dataModel);
        RecommenderBuilder recommenderBuilder = RecommendFactory.itemKNNRecommender(itemSimilarity, new NonNegativeQuadraticOptimizer(), 10);

        RecommendFactory.evaluate(RecommendFactory.EVALUATOR.AVERAGE_ABSOLUTE_DIFFERENCE, recommenderBuilder, null, dataModel, 0.7);
        RecommendFactory.statsEvaluator(recommenderBuilder, null, dataModel, 2);

        LongPrimitiveIterator iter = dataModel.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = recommenderBuilder.buildRecommender(dataModel).recommend(uid, RECOMMENDER_NUM);
            RecommendFactory.showItems(uid, list, true);
        }
    }

程序输出:


AVERAGE_ABSOLUTE_DIFFERENCE Evaluater Score:1.5
Recommender IR Evaluator: [Precision:0.5,Recall:1.0]
uid:1,(107,5.000000)(104,3.501168)(106,3.498198)
uid:2,(105,2.878995)(106,2.878086)(107,2.000000)
uid:3,(103,3.667444)(102,3.667161)(106,3.667019)
uid:4,(107,4.750247)(102,4.122755)(105,4.122709)
uid:5,(107,3.833621)

9. SVD推荐算法

(TODO未完)

算法API: org.apache.mahout.cf.taste.impl.recommender.svd.SVDRecommender


@Override
  public float estimatePreference(long userID, long itemID) throws TasteException {
    double[] userFeatures = factorization.getUserFeatures(userID);
    double[] itemFeatures = factorization.getItemFeatures(itemID);
    double estimate = 0;
    for (int feature = 0; feature < userFeatures.length; feature++) {
      estimate += userFeatures[feature] * itemFeatures[feature];
    }
    return (float) estimate;
  }

测试程序:


    public static void svd(DataModel dataModel) throws TasteException {
        RecommenderBuilder recommenderBuilder = RecommendFactory.svdRecommender(new ALSWRFactorizer(dataModel, 10, 0.05, 10));

        RecommendFactory.evaluate(RecommendFactory.EVALUATOR.AVERAGE_ABSOLUTE_DIFFERENCE, recommenderBuilder, null, dataModel, 0.7);
        RecommendFactory.statsEvaluator(recommenderBuilder, null, dataModel, 2);

        LongPrimitiveIterator iter = dataModel.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = recommenderBuilder.buildRecommender(dataModel).recommend(uid, RECOMMENDER_NUM);
            RecommendFactory.showItems(uid, list, true);
        }
    }

程序输出:


AVERAGE_ABSOLUTE_DIFFERENCE Evaluater Score:0.09990564982096355
Recommender IR Evaluator: [Precision:0.5,Recall:1.0]
uid:1,(104,4.032909)(105,3.390885)(107,1.858541)
uid:2,(105,3.761718)(106,2.951908)(107,1.561116)
uid:3,(103,5.593422)(102,2.458930)(106,-0.091259)
uid:4,(105,4.068329)(102,3.534025)(107,0.206257)
uid:5,(107,0.105169)

10. Tree Cluster-based 推荐算法

这个算法在mahout-0.8版本中,已经被@Deprecated。

(TODO未完)

算法API: org.apache.mahout.cf.taste.impl.recommender.TreeClusteringRecommender


  @Override
  public float estimatePreference(long userID, long itemID) throws TasteException {
    DataModel model = getDataModel();
    Float actualPref = model.getPreferenceValue(userID, itemID);
    if (actualPref != null) {
      return actualPref;
    }
    buildClusters();
    List topRecsForUser = topRecsByUserID.get(userID);
    if (topRecsForUser != null) {
      for (RecommendedItem item : topRecsForUser) {
        if (itemID == item.getItemID()) {
          return item.getValue();
        }
      }
    }
    // Hmm, we have no idea. The item is not in the user's cluster
    return Float.NaN;
  }

测试程序:


    public static void treeCluster(DataModel dataModel) throws TasteException {
        UserSimilarity userSimilarity = RecommendFactory.userSimilarity(RecommendFactory.SIMILARITY.LOGLIKELIHOOD, dataModel);
        ClusterSimilarity clusterSimilarity = RecommendFactory.clusterSimilarity(RecommendFactory.SIMILARITY.FARTHEST_NEIGHBOR_CLUSTER, userSimilarity);
        RecommenderBuilder recommenderBuilder = RecommendFactory.treeClusterRecommender(clusterSimilarity, 10);

        RecommendFactory.evaluate(RecommendFactory.EVALUATOR.AVERAGE_ABSOLUTE_DIFFERENCE, recommenderBuilder, null, dataModel, 0.7);
        RecommendFactory.statsEvaluator(recommenderBuilder, null, dataModel, 2);

        LongPrimitiveIterator iter = dataModel.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = recommenderBuilder.buildRecommender(dataModel).recommend(uid, RECOMMENDER_NUM);
            RecommendFactory.showItems(uid, list, true);
        }
    }

程序输出:


AVERAGE_ABSOLUTE_DIFFERENCE Evaluater Score:NaN
Recommender IR Evaluator: [Precision:NaN,Recall:0.0]

11. Mahout推荐算法总结

算法及适用场景:

recommender-intro

算法评分的结果:

recommender-score

通过对上面几种算法的一平分比较:itemCF,itemKNN,SVD的Rrecision,Recall的评分值是最好的,并且itemCF和SVD的AVERAGE_ABSOLUTE_DIFFERENCE是最低的,所以,从算法的角度知道了,哪个算法是更准确的或者会索引到更多的数据集。

另外的一些因素:

  • 1. 这3个指标,并不能直接决定计算结果一定itemCF,SVD好
  • 2. 各种算法的参数我们并没有调优
  • 3. 数据量和数据分布,是影响算法的评分

程序源代码下载

https://github.com/bsspirit/maven_mahout_template/tree/mahout-0.8/src/main/java/org/conan/mymahout/recommendation/job

转载请注明出处:
http://blog.fens.me/mahout-recommendation-api

打赏作者

Mahout分步式程序开发 基于物品的协同过滤ItemCF

Hadoop家族系列文章,主要介绍Hadoop家族产品,常用的项目包括Hadoop, Hive, Pig, HBase, Sqoop, Mahout, Zookeeper, Avro, Ambari, Chukwa,新增加的项目包括,YARN, Hcatalog, Oozie, Cassandra, Hama, Whirr, Flume, Bigtop, Crunch, Hue等。

从2011年开始,中国进入大数据风起云涌的时代,以Hadoop为代表的家族软件,占据了大数据处理的广阔地盘。开源界及厂商,所有数据软件,无一不向Hadoop靠拢。Hadoop也从小众的高富帅领域,变成了大数据开发的标准。在Hadoop原有技术基础之上,出现了Hadoop家族产品,通过“大数据”概念不断创新,推出科技进步。

作为IT界的开发人员,我们也要跟上节奏,抓住机遇,跟着Hadoop一起雄起!

关于作者:

  • 张丹(Conan), 程序员Java,R,PHP,Javascript
  • weibo:@Conan_Z
  • blog: http://blog.fens.me
  • email: bsspirit@gmail.com

转载请注明出处:
http://blog.fens.me/hadoop-mahout-mapreduce-itemcf/

mahout-hadoop-itemcf

前言

Mahout是Hadoop家族一员,从血缘就继承了Hadoop程序的特点,支持HDFS访问和MapReduce分步式算法。随着Mahout的发展,从0.7版本开始,Mahout做了重大的升级。移除了部分算法的单机内存计算,只支持基于Hadoop的MapReduce平行计算。

从这点上,我们能看出Mahout走向大数据,坚持并行化的决心!相信在Hadoop的大框架下,Mahout最终能成为一个大数据的明星产品!

目录

  1. Mahout开发环境介绍
  2. Mahout基于Hadoop的分步环境介绍
  3. 用Mahout实现协同过滤ItemCF
  4. 模板项目上传github

1. Mahout开发环境介绍

用Maven构建Mahout项目 文章中,我们已经配置好了基于Maven的Mahout的开发环境,我们将继续完成Mahout的分步式的程序开发。

本文的mahout版本为0.8。

开发环境:

  • Win7 64bit
  • Java 1.6.0_45
  • Maven 3
  • Eclipse Juno Service Release 2
  • Mahout 0.8
  • Hadoop 1.1.2

找到pom.xml,修改mahout版本为0.8

<mahout.version>0.8</mahout.version>

然后,下载依赖库。

~ mvn clean install

由于 org.conan.mymahout.cluster06.Kmeans.java 类代码,是基于mahout-0.6的,所以会报错。我们可以先注释这个文件。

2. Mahout基于Hadoop的分步环境介绍

hadoop-mahout-cluster-dev

如上图所示,我们可以选择在win7中开发,也可以在linux中开发,开发过程我们可以在本地环境进行调试,标配的工具都是Maven和Eclipse。

Mahout在运行过程中,会把MapReduce的算法程序包,自动发布的Hadoop的集群环境中,这种开发和运行模式,就和真正的生产环境差不多了。

3. 用Mahout实现协同过滤ItemCF

实现步骤:

  • 1. 准备数据文件: item.csv
  • 2. Java程序:HdfsDAO.java
  • 3. Java程序:ItemCFHadoop.java
  • 4. 运行程序
  • 5. 推荐结果解读

1). 准备数据文件: item.csv
上传测试数据到HDFS,单机内存实验请参考文章:用Maven构建Mahout项目


~ hadoop fs -mkdir /user/hdfs/userCF
~ hadoop fs -copyFromLocal /home/conan/datafiles/item.csv /user/hdfs/userCF

~ hadoop fs -cat /user/hdfs/userCF/item.csv
1,101,5.0
1,102,3.0
1,103,2.5
2,101,2.0
2,102,2.5
2,103,5.0
2,104,2.0
3,101,2.5
3,104,4.0
3,105,4.5
3,107,5.0
4,101,5.0
4,103,3.0
4,104,4.5
4,106,4.0
5,101,4.0
5,102,3.0
5,103,2.0
5,104,4.0
5,105,3.5
5,106,4.0

2). Java程序:HdfsDAO.java
HdfsDAO.java,是一个HDFS操作的工具,用API实现Hadoop的各种HDFS命令,请参考文章:Hadoop编程调用HDFS

我们这里会用到HdfsDAO.java类中的一些方法:


        HdfsDAO hdfs = new HdfsDAO(HDFS, conf);
        hdfs.rmr(inPath);
        hdfs.mkdirs(inPath);
        hdfs.copyFile(localFile, inPath);
        hdfs.ls(inPath);
        hdfs.cat(inFile);

3). Java程序:ItemCFHadoop.java
用Mahout实现分步式算法,我们看到Mahout in Action中的解释。

aglorithm_2

实现程序:


package org.conan.mymahout.recommendation;

import org.apache.hadoop.mapred.JobConf;
import org.apache.mahout.cf.taste.hadoop.item.RecommenderJob;
import org.conan.mymahout.hdfs.HdfsDAO;

public class ItemCFHadoop {

    private static final String HDFS = "hdfs://192.168.1.210:9000";

    public static void main(String[] args) throws Exception {
        String localFile = "datafile/item.csv";
        String inPath = HDFS + "/user/hdfs/userCF";
        String inFile = inPath + "/item.csv";
        String outPath = HDFS + "/user/hdfs/userCF/result/";
        String outFile = outPath + "/part-r-00000";
        String tmpPath = HDFS + "/tmp/" + System.currentTimeMillis();

        JobConf conf = config();
        HdfsDAO hdfs = new HdfsDAO(HDFS, conf);
        hdfs.rmr(inPath);
        hdfs.mkdirs(inPath);
        hdfs.copyFile(localFile, inPath);
        hdfs.ls(inPath);
        hdfs.cat(inFile);

        StringBuilder sb = new StringBuilder();
        sb.append("--input ").append(inPath);
        sb.append(" --output ").append(outPath);
        sb.append(" --booleanData true");
        sb.append(" --similarityClassname org.apache.mahout.math.hadoop.similarity.cooccurrence.measures.EuclideanDistanceSimilarity");
        sb.append(" --tempDir ").append(tmpPath);
        args = sb.toString().split(" ");

        RecommenderJob job = new RecommenderJob();
        job.setConf(conf);
        job.run(args);

        hdfs.cat(outFile);
    }

    public static JobConf config() {
        JobConf conf = new JobConf(ItemCFHadoop.class);
        conf.setJobName("ItemCFHadoop");
        conf.addResource("classpath:/hadoop/core-site.xml");
        conf.addResource("classpath:/hadoop/hdfs-site.xml");
        conf.addResource("classpath:/hadoop/mapred-site.xml");
        return conf;
    }
}

RecommenderJob.java,实际上就是封装了,上面整个图的分步式并行算法的执行过程!如果没有这层封装,我们需要自己去实现图中8个步骤MapReduce算法。

关于上面算法的深度剖析,请参考文章:R实现MapReduce的协同过滤算法

4). 运行程序
控制台输出:


Delete: hdfs://192.168.1.210:9000/user/hdfs/userCF
Create: hdfs://192.168.1.210:9000/user/hdfs/userCF
copy from: datafile/item.csv to hdfs://192.168.1.210:9000/user/hdfs/userCF
ls: hdfs://192.168.1.210:9000/user/hdfs/userCF
==========================================================
name: hdfs://192.168.1.210:9000/user/hdfs/userCF/item.csv, folder: false, size: 229
==========================================================
cat: hdfs://192.168.1.210:9000/user/hdfs/userCF/item.csv
1,101,5.0
1,102,3.0
1,103,2.5
2,101,2.0
2,102,2.5
2,103,5.0
2,104,2.0
3,101,2.5
3,104,4.0
3,105,4.5
3,107,5.0
4,101,5.0
4,103,3.0
4,104,4.5
4,106,4.0
5,101,4.0
5,102,3.0
5,103,2.0
5,104,4.0
5,105,3.5
5,106,4.0SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
2013-10-14 10:26:35 org.apache.hadoop.util.NativeCodeLoader 
警告: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
2013-10-14 10:26:35 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:35 org.apache.hadoop.io.compress.snappy.LoadSnappy 
警告: Snappy native library not loaded
2013-10-14 10:26:36 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0001
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:36 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:36 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:36 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:36 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:36 org.apache.hadoop.io.compress.CodecPool getCompressor
信息: Got brand-new compressor
2013-10-14 10:26:36 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0001_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:36 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0001_m_000000_0' done.
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:36 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:36 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:36 org.apache.hadoop.io.compress.CodecPool getDecompressor
信息: Got brand-new decompressor
2013-10-14 10:26:36 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 42 bytes
2013-10-14 10:26:36 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0001_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:36 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0001_r_000000_0 is allowed to commit now
2013-10-14 10:26:36 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0001_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/preparePreferenceMatrix/itemIDIndex
2013-10-14 10:26:36 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:36 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0001_r_000000_0' done.
2013-10-14 10:26:37 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:37 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0001
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息: Counters: 19
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=187
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=3287330
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=916
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=3443292
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=645
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=229
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=46
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Map input records=21
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=14
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=84
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=376569856
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=116
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=21
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=7
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=7
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=7
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=7
2013-10-14 10:26:37 org.apache.hadoop.mapred.Counters log
信息:     Map output records=21
2013-10-14 10:26:37 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:37 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0002
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:37 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:37 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:37 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:37 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:37 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0002_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:37 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0002_m_000000_0' done.
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:37 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:37 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:37 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 68 bytes
2013-10-14 10:26:37 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0002_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:37 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0002_r_000000_0 is allowed to commit now
2013-10-14 10:26:37 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0002_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/preparePreferenceMatrix/userVectors
2013-10-14 10:26:37 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:37 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0002_r_000000_0' done.
2013-10-14 10:26:38 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:38 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0002
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息: Counters: 20
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:   org.apache.mahout.cf.taste.hadoop.item.ToUserVectorsReducer$Counters
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     USERS=5
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=288
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=6574274
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=1374
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=6887592
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=1120
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=229
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=72
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Map input records=21
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=42
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=63
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=575930368
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=116
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=0
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=21
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=5
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=0
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=5
2013-10-14 10:26:38 org.apache.hadoop.mapred.Counters log
信息:     Map output records=21
2013-10-14 10:26:38 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:38 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0003
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:38 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:38 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:38 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:38 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:38 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0003_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:38 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0003_m_000000_0' done.
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:38 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:38 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:38 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 89 bytes
2013-10-14 10:26:38 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0003_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:38 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0003_r_000000_0 is allowed to commit now
2013-10-14 10:26:38 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0003_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/preparePreferenceMatrix/ratingMatrix
2013-10-14 10:26:38 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:38 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0003_r_000000_0' done.
2013-10-14 10:26:39 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:39 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0003
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息: Counters: 21
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=335
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:   org.apache.mahout.cf.taste.hadoop.preparation.ToItemVectorsMapper$Elements
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     USER_RATINGS_NEGLECTED=0
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     USER_RATINGS_USED=21
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=9861349
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=1950
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=10331958
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=1751
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=288
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=93
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Map input records=5
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=14
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=336
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=775290880
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=157
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=21
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=7
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=7
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=7
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=7
2013-10-14 10:26:39 org.apache.hadoop.mapred.Counters log
信息:     Map output records=21
2013-10-14 10:26:39 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:39 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0004
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:39 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:39 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:39 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:39 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:39 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0004_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:39 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0004_m_000000_0' done.
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:39 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:39 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:39 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 118 bytes
2013-10-14 10:26:39 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0004_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:39 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0004_r_000000_0 is allowed to commit now
2013-10-14 10:26:39 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0004_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/weights
2013-10-14 10:26:39 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:39 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0004_r_000000_0' done.
2013-10-14 10:26:40 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:40 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0004
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息: Counters: 20
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=381
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=13148476
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=2628
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=13780408
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=2551
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=335
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:   org.apache.mahout.math.hadoop.similarity.cooccurrence.RowSimilarityJob$Counters
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     ROWS=7
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=122
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Map input records=7
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=16
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=516
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=974651392
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=158
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=24
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=8
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=8
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=8
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=5
2013-10-14 10:26:40 org.apache.hadoop.mapred.Counters log
信息:     Map output records=24
2013-10-14 10:26:40 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:40 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0005
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:40 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:40 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:40 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:40 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:40 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0005_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:40 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0005_m_000000_0' done.
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:40 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:40 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:40 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 121 bytes
2013-10-14 10:26:40 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0005_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:40 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0005_r_000000_0 is allowed to commit now
2013-10-14 10:26:40 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0005_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/pairwiseSimilarity
2013-10-14 10:26:40 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:40 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0005_r_000000_0' done.
2013-10-14 10:26:41 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:41 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0005
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息: Counters: 21
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=392
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=16435577
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=3488
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=17230010
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=3408
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=381
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:   org.apache.mahout.math.hadoop.similarity.cooccurrence.RowSimilarityJob$Counters
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     PRUNED_COOCCURRENCES=0
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     COOCCURRENCES=57
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=125
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Map input records=5
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=14
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=744
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=1174011904
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=129
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=21
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=7
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=7
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=7
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=7
2013-10-14 10:26:41 org.apache.hadoop.mapred.Counters log
信息:     Map output records=21
2013-10-14 10:26:41 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:41 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0006
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:41 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:41 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:41 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:41 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:41 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0006_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:41 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0006_m_000000_0' done.
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:41 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:41 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:41 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 158 bytes
2013-10-14 10:26:41 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0006_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:41 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0006_r_000000_0 is allowed to commit now
2013-10-14 10:26:41 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0006_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/similarityMatrix
2013-10-14 10:26:41 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:41 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0006_r_000000_0' done.
2013-10-14 10:26:42 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:42 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0006
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息: Counters: 19
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=554
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=19722740
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=4342
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=20674772
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=4354
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=392
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=162
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Map input records=7
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=14
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=599
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=1373372416
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=140
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=25
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=7
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=7
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=7
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=7
2013-10-14 10:26:42 org.apache.hadoop.mapred.Counters log
信息:     Map output records=25
2013-10-14 10:26:42 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:42 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:42 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0007
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0007_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:42 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0007_m_000000_0' done.
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:42 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0007_m_000001_0 is done. And is in the process of commiting
2013-10-14 10:26:42 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0007_m_000001_0' done.
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:42 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 2 sorted segments
2013-10-14 10:26:42 org.apache.hadoop.io.compress.CodecPool getDecompressor
信息: Got brand-new decompressor
2013-10-14 10:26:42 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 2 segments left of total size: 233 bytes
2013-10-14 10:26:42 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0007_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:42 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0007_r_000000_0 is allowed to commit now
2013-10-14 10:26:42 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0007_r_000000_0' to hdfs://192.168.1.210:9000/tmp/1381717594500/partialMultiply
2013-10-14 10:26:42 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:42 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0007_r_000000_0' done.
2013-10-14 10:26:43 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:43 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0007
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息: Counters: 19
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=572
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=34517913
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=8751
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=36182630
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=7934
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=0
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=241
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Map input records=12
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=56
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=453
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=2558459904
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=665
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=0
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=28
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=7
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=0
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=7
2013-10-14 10:26:43 org.apache.hadoop.mapred.Counters log
信息:     Map output records=28
2013-10-14 10:26:43 org.apache.hadoop.mapreduce.lib.input.FileInputFormat listStatus
信息: Total input paths to process : 1
2013-10-14 10:26:43 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Running job: job_local_0008
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:43 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: io.sort.mb = 100
2013-10-14 10:26:43 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: data buffer = 79691776/99614720
2013-10-14 10:26:43 org.apache.hadoop.mapred.MapTask$MapOutputBuffer 
信息: record buffer = 262144/327680
2013-10-14 10:26:43 org.apache.hadoop.mapred.MapTask$MapOutputBuffer flush
信息: Starting flush of map output
2013-10-14 10:26:43 org.apache.hadoop.mapred.MapTask$MapOutputBuffer sortAndSpill
信息: Finished spill 0
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0008_m_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:43 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0008_m_000000_0' done.
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task initialize
信息:  Using ResourceCalculatorPlugin : null
2013-10-14 10:26:43 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:43 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Merging 1 sorted segments
2013-10-14 10:26:43 org.apache.hadoop.mapred.Merger$MergeQueue merge
信息: Down to the last merge-pass, with 1 segments left of total size: 206 bytes
2013-10-14 10:26:43 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task done
信息: Task:attempt_local_0008_r_000000_0 is done. And is in the process of commiting
2013-10-14 10:26:43 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: 
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task commit
信息: Task attempt_local_0008_r_000000_0 is allowed to commit now
2013-10-14 10:26:43 org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter commitTask
信息: Saved output of task 'attempt_local_0008_r_000000_0' to hdfs://192.168.1.210:9000/user/hdfs/userCF/result
2013-10-14 10:26:43 org.apache.hadoop.mapred.LocalJobRunner$Job statusUpdate
信息: reduce > reduce
2013-10-14 10:26:43 org.apache.hadoop.mapred.Task sendDone
信息: Task 'attempt_local_0008_r_000000_0' done.
2013-10-14 10:26:44 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息:  map 100% reduce 100%
2013-10-14 10:26:44 org.apache.hadoop.mapred.JobClient monitorAndPrintJob
信息: Job complete: job_local_0008
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息: Counters: 19
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:   File Output Format Counters 
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Bytes Written=217
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:   FileSystemCounters
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_READ=26299802
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_READ=7357
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     FILE_BYTES_WRITTEN=27566408
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     HDFS_BYTES_WRITTEN=6269
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:   File Input Format Counters 
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Bytes Read=572
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:   Map-Reduce Framework
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Map output materialized bytes=210
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Map input records=7
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Reduce shuffle bytes=0
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Spilled Records=42
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Map output bytes=927
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Total committed heap usage (bytes)=1971453952
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     SPLIT_RAW_BYTES=137
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Combine input records=0
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Reduce input records=21
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Reduce input groups=5
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Combine output records=0
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Reduce output records=5
2013-10-14 10:26:44 org.apache.hadoop.mapred.Counters log
信息:     Map output records=21
cat: hdfs://192.168.1.210:9000/user/hdfs/userCF/result//part-r-00000
1	[104:1.280239,106:1.1462644,105:1.0653841,107:0.33333334]
2	[106:1.560478,105:1.4795978,107:0.69935876]
3	[103:1.2475469,106:1.1944525,102:1.1462644]
4	[102:1.6462644,105:1.5277859,107:0.69935876]
5	[107:1.1993587]

5). 推荐结果解读
我们可以把上面的日志分解析成3个部分解读

  • a. 初始化环境
  • b. 算法执行
  • c. 打印推荐结果

a. 初始化环境
出初HDFS的数据目录和工作目录,并上传数据文件。


Delete: hdfs://192.168.1.210:9000/user/hdfs/userCF
Create: hdfs://192.168.1.210:9000/user/hdfs/userCF
copy from: datafile/item.csv to hdfs://192.168.1.210:9000/user/hdfs/userCF
ls: hdfs://192.168.1.210:9000/user/hdfs/userCF
==========================================================
name: hdfs://192.168.1.210:9000/user/hdfs/userCF/item.csv, folder: false, size: 229
==========================================================
cat: hdfs://192.168.1.210:9000/user/hdfs/userCF/item.csv

b. 算法执行
分别执行,上图中对应的8种MapReduce算法。
Job complete: job_local_0001
Job complete: job_local_0002
Job complete: job_local_0003
Job complete: job_local_0004
Job complete: job_local_0005
Job complete: job_local_0006
Job complete: job_local_0007
Job complete: job_local_0008

c. 打印推荐结果

方便我们看到计算后的推荐结果


cat: hdfs://192.168.1.210:9000/user/hdfs/userCF/result//part-r-00000
1	[104:1.280239,106:1.1462644,105:1.0653841,107:0.33333334]
2	[106:1.560478,105:1.4795978,107:0.69935876]
3	[103:1.2475469,106:1.1944525,102:1.1462644]
4	[102:1.6462644,105:1.5277859,107:0.69935876]
5	[107:1.1993587]

4. 模板项目上传github

https://github.com/bsspirit/maven_mahout_template/tree/mahout-0.8

大家可以下载这个项目,做为开发的起点。


~ git clone https://github.com/bsspirit/maven_mahout_template
~ git checkout mahout-0.8

我们完成了基于物品的协同过滤分步式算法实现,下面将继续介绍Mahout的Kmeans的分步式算法实现,请参考文章:Mahout分步式程序开发 聚类Kmeans

转载请注明出处:
http://blog.fens.me/hadoop-mahout-mapreduce-itemcf/

打赏作者