基于Mahout的图书推荐系统

一、 用Maven搭建Mahout的开发环境

package com.panguoyuan.mahout.itemcf;

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

import org.apache.mahout.cf.taste.common.TasteException;
import org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;
import org.apache.mahout.cf.taste.impl.model.file.FileDataModel;
import org.apache.mahout.cf.taste.impl.neighborhood.NearestNUserNeighborhood;
import org.apache.mahout.cf.taste.impl.recommender.GenericUserBasedRecommender;
import org.apache.mahout.cf.taste.impl.similarity.EuclideanDistanceSimilarity;
import org.apache.mahout.cf.taste.model.DataModel;
import org.apache.mahout.cf.taste.recommender.RecommendedItem;
import org.apache.mahout.cf.taste.recommender.Recommender;
import org.apache.mahout.cf.taste.similarity.UserSimilarity;
public class UserCF {
    final static int NEIGHBORHOOD_NUM = 2;
    final static int RECOMMENDER_NUM = 3;
    public static void main(String[] args) throws IOException, TasteException {
        String file = "inputdata/item.csv";
        DataModel model = new FileDataModel(new File(file));
        UserSimilarity user = new EuclideanDistanceSimilarity(model);
        NearestNUserNeighborhood neighbor = new NearestNUserNeighborhood(NEIGHBORHOOD_NUM, user, model);
        Recommender r = new GenericUserBasedRecommender(model, neighbor, user);
        LongPrimitiveIterator iter = model.getUserIDs();
        while (iter.hasNext()) {
            long uid = iter.nextLong();
            List list = r.recommend(uid, RECOMMENDER_NUM);
            System.out.printf("uid:%s", uid);
            for (RecommendedItem ritem : list) {
                System.out.printf("(%s,%f)", ritem.getItemID(), ritem.getValue());
            }
            System.out.println();
        }
    }
}

（8）在eclipse里运行结果如下


二、用案例的数据集，基于Mahout，任选一种算法，对任意一个女性用户进行协同过滤推荐，并解释推荐结果是否合理，解释过程可以写成一文档说明。
1、选择基于用户的协同过滤算法:UserCF
2、算法模型：DataModel+UserSimilarity+UserNeighborhood+UserBasedRecommender

package com.panguoyuan.mahout.itemcf;

import java.io.File;
import java.util.List;

import org.apache.mahout.cf.taste.impl.model.file.FileDataModel;
import org.apache.mahout.cf.taste.impl.neighborhood.NearestNUserNeighborhood;
import org.apache.mahout.cf.taste.impl.recommender.GenericUserBasedRecommender;
import org.apache.mahout.cf.taste.impl.similarity.PearsonCorrelationSimilarity;
import org.apache.mahout.cf.taste.model.DataModel;
import org.apache.mahout.cf.taste.neighborhood.UserNeighborhood;
import org.apache.mahout.cf.taste.recommender.RecommendedItem;
import org.apache.mahout.cf.taste.recommender.UserBasedRecommender;
import org.apache.mahout.cf.taste.similarity.UserSimilarity;

public class BasedUserBookRecommender2 {

    public static void main(String[] args) throws Exception {
        long userId = 188;
        //构建数据模型
        DataModel model = new FileDataModel(new File("inputdata/rating.csv"));
        //创建相似度
        UserSimilarity itemSimilarity = new PearsonCorrelationSimilarity(model);
        //UserSimilarity itemSimilarity = new EuclideanDistanceSimilarity(model);
        //GenericUserSimilarity genericItemSimilarity = new GenericUserSimilarity(itemSimilarity, model);
        //构建近邻算法
        UserNeighborhood neighborhood = new NearestNUserNeighborhood(3, itemSimilarity, model);
        //构建推荐模型
        UserBasedRecommender userBasedRecommender = new GenericUserBasedRecommender(model, neighborhood, itemSimilarity);
        //计算并返回图书推荐结果
        List recommendations = userBasedRecommender.recommend(188, 5);
        //打印推荐结果
        showItems(userId, recommendations, true);

    }

    public static void showItems(long uid,
            List recommendations, boolean skip) {
        if (skip || recommendations.size() > 0) {
            System.out.printf("userId:%s,", uid);
            for (RecommendedItem r : recommendations) {
                System.out.printf("(%s,%f)", r.getItemID(), r.getValue());
            }
            System.out.println();
        }
    }
}

4、输出结果

userId:188,(885,9.500000)(396,7.000000)(688,6.000000)

5、用R语言对推荐结果进行人工分析
（1）导入分析数据（rating.csv为评分数据，user.csv为用户信息）

ratings=read.csv("F:\workspace1\mahout\inputdata\rating.csv",FALSE)
users=read.csv("F:\workspace1\mahout\inputdata\user.csv",FALSE)

（2）修改列名

ratings=data.frame('userid'=ratings$V1,'bookid'=ratings$V2,'grade'=ratings$V3)
users=data.frame('userid'=users$V1,'sex'=users$V2,'age'=users$V3)

（3）查看用户188都看了哪些书

> ratings[c(ratings$userid==188),]
userid bookid grade
3760    188    798     6
3761    188    653     3
3762    188    426     6
3763    188    742     7
3764    188    549     2
3765    188    520     8
3766    188    312     2
3767    188    213    10
3768    188    954     5
3769    188    121    10
3770    188    204     9
3771    188    684     3
3772    188    493     4
3773    188    452     1
3774    188    622     3
3775    188    298     8

（4）图书885推荐分数最高，下面查看该图书有哪些人评过分

ratings[c(ratings$bookid==885),]
userid bookid grade
182       9    885     8
1225     60    885    10
3691    184    885     9

（5）查看这用户9,用户60,用户184，用户188的信息

> users[c(9,60,184,188),]  userid sex age
9        9   M  50
60      60   F  49
184    184   M  27
188    188   F  24

（6）查看这用户9，用户60，用户184与用户188都共同看了哪些图书

> rating188=ratings[which(ratings$userid==188),]
>  rating9=ratings[which(ratings$userid==9),]
>  rating60=ratings[which(ratings$userid==60),]
> rating184=ratings[which(ratings$userid==184),]
> intersect(rating188$bookid,rating9$bookid)
integer(0)
> intersect(rating188$bookid,rating60$bookid)
[1] 312 298
> intersect(rating188$bookid,rating184$bookid)
[1] 121 684

    从上面可以看出用户188与用户60共同看了312和298这两本书，与用户184共同看了121和684这两本书，他们都有共同的偏好，所以给用户188推荐图书885是合理的。

三、增加过滤条件，排除男性，只保留对女性用户的推荐评分
 选用的算法模型为：FileDataModel+EuclideanDistanceSimilarity+GenericItemBasedRecommender

package com.panguoyuan.mahout.itemcf;

import java.io.BufferedReader;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
import java.util.HashSet;
import java.util.List;
import java.util.Set;

import org.apache.mahout.cf.taste.common.TasteException;
import org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;
import org.apache.mahout.cf.taste.impl.model.file.FileDataModel;
import org.apache.mahout.cf.taste.impl.recommender.GenericItemBasedRecommender;
import org.apache.mahout.cf.taste.impl.similarity.EuclideanDistanceSimilarity;
import org.apache.mahout.cf.taste.impl.similarity.GenericItemSimilarity;
import org.apache.mahout.cf.taste.model.DataModel;
import org.apache.mahout.cf.taste.recommender.IDRescorer;
import org.apache.mahout.cf.taste.recommender.ItemBasedRecommender;
import org.apache.mahout.cf.taste.recommender.RecommendedItem;
import org.apache.mahout.cf.taste.similarity.ItemSimilarity;

public class BookFilterGenderRecommender3 {

    public static void main(String[] args) throws Exception {
        DataModel model = new FileDataModel(new File("inputdata/rating.csv"));
        ItemSimilarity otherSimilarity = new EuclideanDistanceSimilarity(model);
        GenericItemSimilarity similarity = new GenericItemSimilarity(otherSimilarity, model);
        ItemBasedRecommender recommender = new GenericItemBasedRecommender(model, similarity);
        filterRecommender(188, recommender, model);
    }

    public static void showItems(long uid,List recommendations, boolean skip) {
        if (skip || recommendations.size() > 0) {
            System.out.printf("userId:%s,", uid);
            for (RecommendedItem r : recommendations) {
                System.out.printf("Item:(%s,%f)", r.getItemID(), r.getValue());
                System.out.println();
            }

        }
    }
   
     /**
     * 对用户性别进行过滤
     */
    public static void filterRecommender(long uid, ItemBasedRecommender recommender, DataModel dataModel) throws TasteException, IOException {
        Set userids = getMale("datafile/book/user.csv");

        //计算男性用户打分过的图书
        Set bookids = new HashSet();
        for (long uids : userids) {
            LongPrimitiveIterator iter = dataModel.getItemIDsFromUser(uids).iterator();
            while (iter.hasNext()) {
                long bookid = iter.next();
                bookids.add(bookid);
            }
        }

        IDRescorer rescorer = new FilterRescorer(bookids);
        List list = recommender.recommend(uid, 10, rescorer);
        showItems(uid, list, false);
    }

    /**
     * 返回所有男性id
     */
    public static Set getMale(String file) throws IOException {
        BufferedReader br = new BufferedReader(new FileReader(new File(file)));
        Set userids = new HashSet();
        String s = null;
        while ((s = br.readLine()) != null) {
            String[] cols = s.split(",");
            if (cols[1].equals("M")) {
                userids.add(Long.parseLong(cols[0]));
            }
        }
        br.close();
        return userids;
    }
}

/**
* 对结果重计算
*/
class FilterRescorer implements IDRescorer {
    final private Set userids;

    public FilterRescorer(Set userids) {
        this.userids = userids;
    }

    @Override
    public double rescore(long id, double originalScore) {
        return isFiltered(id) ? Double.NaN : originalScore;
    }

    @Override
    public boolean isFiltered(long id) {
        return userids.contains(id);
    }
}

3、打印推荐结果

SLF4J: 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.
userId:188,Item:(365,8.800000)
Item:(725,8.583333)
Item:(427,8.000000)
Item:(403,7.987013)
Item:(734,7.676371)
Item:(256,7.533333)
Item:(300,7.428571)
Item:(743,7.333333)
Item:(356,6.875000)
Item:(579,6.777778)

4、人工对数据进行分析

（1）查看图书365都有哪些用户评过分

> ratings[c(ratings$bookid==365),]
     userid bookid grade
1046     51    365     9
2206    111    365     9
2632    134    365     4
> users[c(51,111,134),]
    userid sex age
51      51   F  18
111    111   F  40
134    134   F  74

（2）利用intersect函数把用户188与25，45，65这三个用户共同评分过的图书汇集出来

说明：intersect(A,B)是一个数据框都在A和B这些行

>rating188=ratings[which(ratings$userid==188),]
>rating51=ratings[which(ratings$userid==51),]
>rating111=ratings[which(ratings$userid==111),]
>rating134=ratings[which(ratings$userid==134),]
> intersect(rating188$bookid,rating51$bookid)
integer(0)> intersect(rating188$bookid,rating134$bookid)
[1] 204
> intersect(rating188$bookid,rating111$bookid)
[1] 742

（3）从上面可以看出用户188与用户134共同看了204图书，与111共同看了742图书

> rating188
     userid bookid grade
3760    188    798     6
3761    188    653     3
3762    188    426     6
3763    188    742     7
3764    188    549     2
3765    188    520     8
3766    188    312     2
3767    188    213    10
3768    188    954     5
3769    188    121    10
3770    188    204     9
3771    188    684     3
3772    188    493     4
3773    188    452     1
3774    188    622     3
3775    188    298     8

综上所述把图书365推荐给用户188是合理的。