通过mapReduce实现基于项目的协同过滤推荐
- 需求介绍
协同过滤推荐网上有很多种介绍,我这里主要介绍的是基于项目的协同过滤。基于项目的协同过滤推荐基于这样的假设:一个用户会喜欢他之前喜欢的项目相似的项目。因此,基于项目的协同过滤推荐关键在于计算物品之间的相似度。 - 数据介绍
我选用的数据集合如下:
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
3,101,2.0
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
这个数据集合总共有三列,分别是用户的id,项目的id,用户对项目的评分。我想通过这些数据计算出推荐结果。 - 环境介绍
我的系统是Windows7,用的,软件如下:
Eclipse Java EE IDE for Web Developers.
Version: Mars.2 Release (4.5.2)
Build id: 20160218-0600
JDK1.7
Maven3.3.9
Hadoop2.6
开发环境我这里就不介绍了,大家感兴趣的网上搜搜配置一下。 - 实现步骤介绍
1、我先构建一个maven的项目,配置hadoop依赖的jar包,主要步骤如下:
打开ecplise,在下面右键,选择new,选择Project,弹出一个New Project界面,在输入框中输入maven,出现如下图所示的项目,,点击next。。构建maven项目这里就不详细介绍了,网上有很多。这里主要介绍pom.xml配置。我主要配置了如下内容:
hadoop-common
hadoop-hdfs
hadoop-mapreduce-client-core
hadoop-mapreduce-client-jobclient
hadoop-mapreduce-client-common
结果如下:
2、通过WinSCP把uItemTable.csv数据上传到分布式服务器端的Downloads文件夹,通过命令:
./hadoop -put ~/Downloads hdfs://master:9000/data
效果如下图所示:
3、读取hdfs对应数据,按用户分组,计算所有物品出现的组合列表,得到用户对物品的评分矩阵。
3.1、构建推荐入口类,Recommend.java内容如下:
package recommand.hadoop.phl;
import java.io.IOException;
import java.net.URISyntaxException;
import java.util.HashMap;
import java.util.Map;
import java.util.regex.Pattern;
/**
<ul><li>推荐入口</li>
<li>@author root
*
*/
public class Recommend {
public static final String HDFS = "hdfs://10.10.44.92:9000";
public static final Pattern DELIMITER = Pattern.compile("[\t,]");
public static void main(String[] args) throws ClassNotFoundException, IOException, URISyntaxException, InterruptedException {
Map<String,String> path = new HashMap<String,String>();
path.put("Step1Input", HDFS + "/data/uItemTable.csv");
path.put("Step1Output", HDFS + "/output/step1");
path.put("Step2Input", path.get("Step1Output"));
path.put("Step2Output", HDFS + "/output/step2");
path.put("Step3Input1", path.get("Step1Output"));
path.put("Step3Output1", HDFS + "/output/step3_1");
path.put("Step3Input2", path.get("Step2Output"));
path.put("Step3Output2", HDFS + "/output/step3_2");
path.put("Step4Input1", path.get("Step3Output1"));
path.put("Step4Input2", path.get("Step3Output2"));
path.put("Step4Output", HDFS + "/output/step4");
Step1.runStep1(path);
Step2.runStep2(path);
Step3.runStep3_1(path);
Step3.runStep3_2(path);
Step4.runStep4(path);
System.exit(0);
}
}
3.2、管理hdfs的类,HDFSDao.java内容如下:
package recommand.hadoop.phl;
import java.io.IOException;
import java.net.URI;
import java.net.URISyntaxException;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.mapreduce.Job;
import sun.tools.tree.ThisExpression;
public class HDFSDao {
//HDFS访问地址
private static final String HDFS = "hdfs://10.10.44.92:9000/";
//hdfs路径
private String hdfsPath;
//Hadoop系统配置
private Configuration conf;
//运行job的名字
private String name;
public HDFSDao(Configuration conf) {
this.conf = conf;
}
public HDFSDao( Configuration conf,String name) {
this.conf = conf;
this.name = name;
}
public HDFSDao(String hdfs, Configuration conf,String name) {
this.hdfsPath = hdfs;
this.conf = conf;
this.name = name;
}
public Job conf() throws IOException{
Job job = Job.getInstance(this.conf,this.name);
return job;
}
public void rmr(String outUrl) throws IOException, URISyntaxException {
FileSystem fileSystem = FileSystem.get(new URI(outUrl),this.conf);
if(fileSystem.exists(new Path(outUrl))){
fileSystem.delete(new Path(outUrl), true);
System.out.println("outUrl "+outUrl);
}
}
}
3.3、按用户分组,计算所有物品出现的组合列表,得到用户对物品的评分矩阵,Step1.java,内容如下:
package recommand.hadoop.phl;
import java.io.IOException;
import java.net.URISyntaxException;
import java.util.Map;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
/**</li>
<li>按用户分组,计算所有物品出现的组合列表,得到用户对物品的评分矩阵</li>
<li>@author root
*
*/
public class Step1 {
public static class myMapper extends Mapper<Object,Text,IntWritable,Text> {
private final static IntWritable k = new IntWritable();
private final static Text v = new Text();
@Override
protected void map(Object key, Text value, Context context)
throws IOException, InterruptedException {
String[] tokens = Recommend.DELIMITER.split(value.toString());
int userID = Integer.parseInt(tokens[0]);
String itemID = tokens[1];
String pref = tokens[2];
k.set(userID);
v.set(itemID + ":" + pref);
context.write(k, v);
}
}
public static class myReducer extends Reducer<IntWritable, Text, IntWritable, Text>{
private final static Text v = new Text();
@Override
protected void reduce(IntWritable k2, Iterable<Text> v2s,Context context)throws IOException, InterruptedException {
StringBuilder sb = new StringBuilder();
for (Text v2 : v2s) {
sb.append("," + v2.toString());
}
v.set(sb.toString().replaceFirst(",", ""));
context.write(k2, v);
}
}
public static void runStep1(Map<String, String> path) throws IOException, URISyntaxException, ClassNotFoundException, InterruptedException {
String input = path.get("Step1Input");
String output = path.get("Step1Output");
//1.1读取hdfs文件
Configuration conf = new Configuration();
HDFSDao hdfsDao = new HDFSDao(input,conf,Step1.class.getSimpleName());
Job job = hdfsDao.conf();
//打包运行方法
job.setJarByClass(Step1.class);
//1.1读取文件 设置输入路径以及文件输入格式
FileInputFormat.addInputPath(job, new Path(input));
job.setInputFormatClass(TextInputFormat.class);
//指定自定义的Mapper类
job.setMapperClass(myMapper.class);
//指定Mapper输出的key value类型
job.setMapOutputKeyClass(IntWritable.class);
job.setMapOutputValueClass(Text.class);
//1.3分区
/* job.setPartitionerClass(HashPartitioner.class);
job.setNumReduceTasks(1);*/
//1.4排序 分组
//1.5归约
job.setCombinerClass(myReducer.class);
//2.1
//2.2指定自定义的Reduce类
job.setReducerClass(myReducer.class);
job.setOutputKeyClass(IntWritable.class);
job.setOutputValueClass(Text.class);
//2.3指定输出的路径
FileOutputFormat.setOutputPath(job, new Path(output));
job.setOutputFormatClass(TextOutputFormat.class);
hdfsDao.rmr(output);
job.waitForCompletion(true);
}
}
结果如下:
以第一条数据为例说明,用户1评价了三个商品分别是101,102,103,对应评分分别是:5,3,2.5。
3.4、对物品组合列表进行计数,建立物品的同现矩阵,Step2.java内容如下:
package recommand.hadoop.phl;
import java.io.IOException;
import java.net.URISyntaxException;
import java.util.Map;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
public class Step2 {
public static class myMapper extends Mapper<Object, Text, Text, IntWritable>{
private final static Text k = new Text();
private final static IntWritable v = new IntWritable(1);
@Override
protected void map(Object key, Text value, Context context)
throws IOException, InterruptedException {
//[1, 101:5, 102:3, 103:2.5]
String[] tokens = Recommend.DELIMITER.split(value.toString());
for (int i = 1; i < tokens.length; i++) {
String itemID = tokens[i].split(":")[0];//itemID 101
for (int j = 1; j < tokens.length; j++) {
String itemID2 = tokens[j].split(":")[0]; //101
k.set(itemID + ":" + itemID2);
context.write(k, v);
}
}
}
}
public static class myReducer extends Reducer<Text, IntWritable, Text, IntWritable>{
private final static IntWritable v = new IntWritable();
@Override
protected void reduce(Text k2, Iterable<IntWritable> v2s,
Context context) throws IOException, InterruptedException {
int count=0;
for (IntWritable v2 : v2s) {
count+=v2.get();
}
v.set(count);
context.write(k2, v);
}
}
public static void runStep2(Map<String, String> path) throws IOException, URISyntaxException, ClassNotFoundException, InterruptedException {
String input = path.get("Step2Input");
String output = path.get("Step2Output");
//1.1读取hdfs文件
Configuration conf = new Configuration();
HDFSDao hdfsDao = new HDFSDao(input,conf,Step2.class.getSimpleName());
Job job = hdfsDao.conf();
//打包运行方法
job.setJarByClass(Step2.class);
//1.1读取文件 设置输入路径以及文件输入格式
FileInputFormat.addInputPath(job, new Path(input));
job.setInputFormatClass(TextInputFormat.class);
//指定自定义的Mapper类
job.setMapperClass(myMapper.class);
//指定Mapper输出的key value类型
job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(IntWritable.class);
//1.3分区
/* job.setPartitionerClass(HashPartitioner.class);
job.setNumReduceTasks(1);*/
//1.4排序 分组
//1.5归约
job.setCombinerClass(myReducer.class);
//2.1
//2.2指定自定义的Reduce类
job.setReducerClass(myReducer.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(IntWritable.class);
//2.3指定输出的路径
FileOutputFormat.setOutputPath(job, new Path(output));
job.setOutputFormatClass(TextOutputFormat.class);
//检查是否已有相同输出,有则删除
hdfsDao.rmr(output);
job.waitForCompletion(true);
}
}
运行结果如下:
101:101 5
101:102 3
101:103 4
101:104 4
101:105 2
101:106 2
101:107 1
102:101 3
102:102 3
102:103 3
102:104 2
102:105 1
102:106 1
103:101 4
103:102 3
103:103 4
103:104 3
103:105 1
103:106 2
104:101 4
104:102 2
104:103 3
104:104 4
104:105 2
104:106 2
104:107 1
105:101 2
105:102 1
105:103 1
105:104 2
105:105 2
105:106 1
105:107 1
106:101 2
106:102 1
106:103 2
106:104 2
106:105 1
106:106 2
107:101 1
107:104 1
107:105 1
107:107 1
我们构建的同现矩阵是以两个商品为一组,以第二组为例说明,101:102 3意思是既对101商品评分又对102商品评分的有3个用户。为了方便阅读我博客的人,我把上述结果转化为数学中的矩阵,如下所示:
[101] [102] [103] [104] [105] [106] [107]
[101] 5 3 4 4 2 2 1
[102] 3 3 3 2 1 1 0
[103] 4 3 4 3 1 2 0
[104] 4 2 3 4 2 2 1
[105] 2 1 1 2 2 1 1
[106] 2 1 2 2 1 2 0
[107] 1 0 0 1 1 0 1
3.5、合并同现矩阵和评分矩阵,Step3.java内容如下:
package recommand.hadoop.phl;
import java.io.IOException;
import java.net.URISyntaxException;
import java.util.Map;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
public class Step3 {
public static class myMapper1 extends Mapper<LongWritable, Text, IntWritable, Text>
{
private final static IntWritable k = new IntWritable();
private final static Text v = new Text();
@Override
protected void map(LongWritable key, Text value, Context context)
throws IOException, InterruptedException {
//101:101 5
String[] tokens = Recommend.DELIMITER.split(value.toString());
for (int i = 1; i < tokens.length; i++) {
String[] vector = tokens[i].split(":");
int itemID = Integer.parseInt(vector[0]);
String pref = vector[1];
k.set(itemID);
v.set(tokens[0] + ":" + pref);
context.write(k, v);
}
}
}
public static class myMapper2 extends Mapper<LongWritable, Text, Text,IntWritable>
{
private final static Text k = new Text();
private final static IntWritable v = new IntWritable();
@Override
protected void map(LongWritable key, Text value, Context context)
throws IOException, InterruptedException {
String[] tokens = Recommend.DELIMITER.split(value.toString());
k.set(tokens[0]);
v.set(Integer.parseInt(tokens[1]));
context.write(k, v);
}
}
public static void runStep3_1(Map<String, String> path) throws IOException, URISyntaxException, ClassNotFoundException, InterruptedException{
String input = path.get("Step3Input1");
String output = path.get("Step3Output1");
//1.1读取hdfs文件
Configuration conf = new Configuration();
HDFSDao hdfsDao = new HDFSDao(input,conf,Step3.class.getSimpleName());
Job job = hdfsDao.conf();
//打包运行方法
job.setJarByClass(Step2.class);
//1.1读取文件 设置输入路径以及文件输入格式
FileInputFormat.addInputPath(job, new Path(input));
job.setInputFormatClass(TextInputFormat.class);
//指定自定义的Mapper类
job.setMapperClass(myMapper1.class);
//指定Mapper输出的key value类型
job.setMapOutputKeyClass(IntWritable.class);
job.setMapOutputValueClass(Text.class);
//2.3指定输出的路径
FileOutputFormat.setOutputPath(job, new Path(output));
job.setOutputFormatClass(TextOutputFormat.class);
//检查是否已有相同输出,有则删除
hdfsDao.rmr(output);
job.waitForCompletion(true);
}
public static void runStep3_2(Map<String, String> path) throws IOException, URISyntaxException, ClassNotFoundException, InterruptedException{
String input = path.get("Step3Input2");
String output = path.get("Step3Output2");
//1.1读取hdfs文件
Configuration conf = new Configuration();
HDFSDao hdfsDao = new HDFSDao(input,conf,Step3.class.getSimpleName());
Job job = hdfsDao.conf();
//打包运行方法
job.setJarByClass(Step2.class);
//1.1读取文件 设置输入路径以及文件输入格式
FileInputFormat.addInputPath(job, new Path(input));
job.setInputFormatClass(TextInputFormat.class);
//指定自定义的Mapper类
job.setMapperClass(myMapper2.class);
//指定Mapper输出的key value类型
job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(IntWritable.class);
//2.3指定输出的路径
FileOutputFormat.setOutputPath(job, new Path(output));
job.setOutputFormatClass(TextOutputFormat.class);
//检查是否已有相同输出,有则删除
hdfsDao.rmr(output);
job.waitForCompletion(true);
}
}
同现矩阵运行结果和3.4步给出的结果一样,用户评分矩阵如下:
以第一条为例说明,101商品用户1给评价了5分。转化成数据中矩阵如下:
[user1] [user2] [user3] [user4] [user5]
[101] 5 2 2 5 4
[102] 3 2.5 0 0 3
[103] 2.5 5 0 3 2
[104] 0 2 4 4 4
[105] 0 0 4.5 0 3.5
[106] 0 0 0 4 4
[107] 0 0 5 0 0
3.6、为了表示方便,我引入了一个类,Cooccurrence.java内容如下:
package recommand.hadoop.phl;
public class Cooccurrence {
private int itemID1;
private int itemID2;
private int num;
public Cooccurrence(int itemID1, int itemID2, int num) {
super();
this.itemID1 = itemID1;
this.itemID2 = itemID2;
this.num = num;
}
public int getItemID1() {
return itemID1;
}
public void setItemID1(int itemID1) {
this.itemID1 = itemID1;
}
public int getItemID2() {
return itemID2;
}
public void setItemID2(int itemID2) {
this.itemID2 = itemID2;
}
public int getNum() {
return num;
}
public void setNum(int num) {
this.num = num;
}
}
由于我们同现矩阵是以两个商品分组的,所以itemID1,itemID2,代表既评价itemID1又评价itemID2,num代表评价的用户数。
3.7、计算推荐列表,Step4.java内容如下:
package recommand.hadoop.phl;
import java.io.IOException;
import java.net.URISyntaxException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
import recommand.hadoop.phl.Step1.myReducer;
public class Step4 {
public static class myMapper extends Mapper<LongWritable, Text, IntWritable, Text>{
private final static IntWritable k = new IntWritable();
private final static Text v = new Text();
private final static Map<Integer, List> cooccurrenceMatrix = new HashMap<Integer, List>();
@Override
protected void map(LongWritable key, Text values, Mapper<LongWritable, Text, IntWritable, Text>.Context context)
throws IOException, InterruptedException {
String[] tokens = Recommend.DELIMITER.split(values.toString());
//同现矩阵时候,v1存的是101 101,v2存的是5。用户评分矩阵,v1存的是101 v2存的是1 5.0
String[] v1 = tokens[0].split(":");
String[] v2 = tokens[1].split(":");
if (v1.length > 1) {// cooccurrence
int itemID1 = Integer.parseInt(v1[0]);
int itemID2 = Integer.parseInt(v1[1]);
int num = Integer.parseInt(tokens[1]);
List<Cooccurrence> list = null;
if (!cooccurrenceMatrix.containsKey(itemID1)) {
list = new ArrayList<Cooccurrence>();
} else {
list = cooccurrenceMatrix.get(itemID1);
}
list.add(new Cooccurrence(itemID1, itemID2, num));
System.out.println("<itemID1,{itemID1, itemID2, num}: "+itemID1+"{"+itemID1+" , "+itemID2+" , "+num+"}");
cooccurrenceMatrix.put(itemID1, list);
}
if (v2.length > 1) {// userVector
int itemID = Integer.parseInt(tokens[0]);
int userID = Integer.parseInt(v2[0]);
double pref = Double.parseDouble(v2[1]);
k.set(userID);
Iterator<Cooccurrence> iterator = cooccurrenceMatrix.get(itemID).iterator();
while(iterator.hasNext()){
Cooccurrence co = iterator.next();
v.set(co.getItemID2() + "," + pref * co.getNum());
System.out.println("ItemID2,pref * co.getNum()"+userID+":{"+co.getItemID2()+ "," + pref * co.getNum()+"}");
context.write(k, v);
}
/*for (Cooccurrence co : cooccurrenceMatrix.get(itemID)) {
v.set(co.getItemID2() + "," + pref * co.getNum());
context.write(k, v);
}*/
}
}
}
public static class myReducer extends Reducer<IntWritable, Text, IntWritable, Text>{
private final static Text v = new Text();
@Override
protected void reduce(IntWritable key, Iterable<Text> values,
Context context)throws IOException, InterruptedException {
Map<String, Double> result = new HashMap<String, Double>();
for(Text value:values){
String[] str = value.toString().split(",");
if (result.containsKey(str[0])) {
result.put(str[0], result.get(str[0]) + Double.parseDouble(str[1]));
}else{
result.put(str[0], Double.parseDouble(str[1]));
}
}
Iterator iter = result.keySet().iterator();
while (iter.hasNext()) {
String itemID = iter.next().toString();
System.out.println("itemID:"+itemID);
double score = result.get(itemID);
v.set(itemID + "," + score);
context.write(key, v);
}
}
}
public static void runStep4(Map<String, String> path) throws IOException, URISyntaxException, ClassNotFoundException, InterruptedException {
String input1 = path.get("Step4Input1");
String input2 = path.get("Step4Input2");
String output = path.get("Step4Output");
//1.1读取hdfs文件
Configuration conf = new Configuration();
HDFSDao hdfsDao = new HDFSDao(conf,Step1.class.getSimpleName());
Job job = hdfsDao.conf();
//打包运行方法
job.setJarByClass(Step1.class);
//1.1读取文件 设置输入路径以及文件输入格式
FileInputFormat.setInputPaths(job, new Path(input1), new Path(input2));
job.setInputFormatClass(TextInputFormat.class);
//指定自定义的Mapper类
job.setMapperClass(myMapper.class);
//指定Mapper输出的key value类型
job.setMapOutputKeyClass(IntWritable.class);
job.setMapOutputValueClass(Text.class);
//2.2指定自定义的Reduce类
job.setReducerClass(myReducer.class);
job.setOutputKeyClass(IntWritable.class);
job.setOutputValueClass(Text.class);
//2.3指定输出的路径
FileOutputFormat.setOutputPath(job, new Path(output));
job.setOutputFormatClass(TextOutputFormat.class);
hdfsDao.rmr(output);
job.waitForCompletion(true);
}
}
计算结果如下所示:
1 107,5.0
1 106,18.0
1 105,15.5
1 104,33.5
1 103,39.0
1 102,31.5
1 101,44.0
2 107,4.0
2 106,20.5
2 105,15.5
2 104,36.0
2 103,41.5
2 102,32.5
2 101,45.5
3 107,15.5
3 106,16.5
3 105,26.0
3 104,38.0
3 103,24.5
3 102,18.5
3 101,40.0
4 107,9.5
4 106,33.0
4 105,26.0
4 104,55.0
4 103,53.5
4 102,37.0
4 101,63.0
5 107,11.5
5 106,34.5
5 105,32.0
5 104,59.0
5 103,56.5
5 102,42.5
5 101,68.0
以用户1为了说明,
1 107,5.0
1 106,18.0
1 105,15.5
1 104,33.5
1 103,39.0
1 102,31.5
1 101,44.0
通过计算,得到用户1多这七中商品的预测评分,其中101,102,103用户评价过,104,105,106,107还没有评价过,通过得分,我们发现这四个商品中评分最高的是104,其次是106…,可以把104商品推荐给用户1。
以上是关于基于项目的协同过过滤推荐的一个入门的介绍,后续文章会对代码做一个调优。