基于mahout的推荐系统 下
3. 代码实现
3.1 数据准备
创建商品表
CREATE TABLE `tb_item` (
`pid` bigint(11) NOT NULL AUTO_INCREMENT,
`name` varchar(2000) CHARACTER SET latin1 DEFAULT NULL,
`types` varchar(2000) CHARACTER SET latin1 DEFAULT NULL,
PRIMARY KEY (`pid`)
) ENGINE=InnoDB AUTO_INCREMENT=65134 DEFAULT CHARSET=utf8;创建用户偏好表
CREATE TABLE `user_pianhao_data1` (
`id` bigint(11) NOT NULL AUTO_INCREMENT,
`uid` bigint(11) DEFAULT NULL,
`pid` bigint(11) DEFAULT NULL,
`val` bigint(11) DEFAULT NULL,
`time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
PRIMARY KEY (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=1001 DEFAULT CHARSET=utf8;
表中测试数据较多，可在网盘下载数据进行初始化链接：https://pan.baidu.com/s/1VSgD2uiJ69zDsw-KCtfEUQ
3.2 搭建springboot工程 
3.3 编写application.properties配置文件
#DB Configuration:
spring.datasource.driverClassName=com.mysql.jdbc.Driver
spring.datasource.url=jdbc:mysql://127.0.0.1:3306/recommend?useUnicode=true&characterEncoding=utf8
spring.datasource.username=root
spring.datasource.password=123456
#spring集成Mybatis环境
#pojo别名扫描包
mybatis.type-aliases-package=cn.itcast.domain
#加载Mybatis映射文件
mybatis.mapper-locations=classpath:mapper/*Mapper.xml3.4 编写domain实体类
package cn.itcast.domain;
public class Item {
private Long pid;
private String name;
private String types;
public Long getPid() {
return pid;
}
public void setPid(Long pid) {
this.pid = pid;
}
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
public String getTypes() {
return types;
}
public void setTypes(String types) {
this.types = types;
}
@Override
public String toString() {
return "Item{" +
"pid=" + pid +
", name='" + name + '\'' +
", types='" + types + '\'' +
'}';
}
}3.5 编写Dao层的mapper接口和xml映射文件
package cn.itcast.dao;
import cn.itcast.domain.Item;
import org.apache.ibatis.annotations.Mapper;
import org.apache.ibatis.annotations.Param;
import java.util.List;
@Mapper
public interface ItemMapper {
public List- findAllByIds(@Param("Ids") List
Ids);
}
3.6 编写MyConfig配置类和Service层
MyConfig配置类：用于设置基于数据库的DataModel模型。
package cn.itcast.myconfig;
import com.mysql.jdbc.jdbc2.optional.MysqlDataSource;
import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel;
import org.apache.mahout.cf.taste.model.DataModel;
import org.apache.mahout.cf.taste.model.JDBCDataModel;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
@Configuration
public class MyConfig {
@Bean
public DataModel getMySQLDataModel(){
MysqlDataSource dataSource=new MysqlDataSource();
dataSource.setServerName("localhost");
dataSource.setUser("root");
dataSource.setPassword("123456");
dataSource.setDatabaseName("recommend");//数据库名字
//参数1：mysql数据源信息，参数2：表名，参数3：用户列字段，参数4：商品列字段，参数5：偏好值字段，参数6：时间戳
JDBCDataModel dataModel=new MySQLJDBCDataModel(dataSource,"user_pianhao_data1","uid","pid","val", "time");
/**
* DataModel可基于数据也可基于文件
* 文件汇总数据格式
* 用户id::商品id::偏好分值::时间戳
* 1::122::5::838985046
* 1::185::5::838983525
* 1::231::5::838983392
* .........
*/
//File file = new File("E:\\initData.dat");
//try {
// DataModel dataModel = new GroupLensDataModel(file);
//} catch (IOException e) {
// e.printStackTrace();
//}
return dataModel;
}
}Service接口：
package cn.itcast.service;
import cn.itcast.domain.Item;
import java.util.List;
public interface RecommendService {
//基于用户的商品推荐
List- getRecommendItemsByUser(Long userId, int howMany);
//基于内容的商品推荐
List
- getRecommendItemsByItem(Long userId, Long itemId, int howMany);
}
Service实现类：
package cn.itcast.service.impl;
import cn.itcast.dao.ItemMapper;
import cn.itcast.domain.Item;
import cn.itcast.service.RecommendService;
import com.mysql.jdbc.jdbc2.optional.MysqlDataSource;
import org.apache.mahout.cf.taste.common.TasteException;
import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel;
import org.apache.mahout.cf.taste.impl.neighborhood.NearestNUserNeighborhood;
import org.apache.mahout.cf.taste.impl.recommender.GenericItemBasedRecommender;
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.model.JDBCDataModel;
import org.apache.mahout.cf.taste.neighborhood.UserNeighborhood;
import org.apache.mahout.cf.taste.recommender.RecommendedItem;
import org.apache.mahout.cf.taste.recommender.Recommender;
import org.apache.mahout.cf.taste.similarity.ItemSimilarity;
import org.apache.mahout.cf.taste.similarity.UserSimilarity;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;
import java.util.ArrayList;
import java.util.List;
@Service
public class RecommendServiceImpl implements RecommendService {
@Autowired
private ItemMapper itemMapper;
@Autowired
private DataModel dataModel;
@Override
public List- getRecommendItemsByUser(Long userId, int howMany) {
List
- list = null;
try {
//计算相似度，相似度算法有很多种，采用基于皮尔逊相关性的相似度
UserSimilarity similarity = new PearsonCorrelationSimilarity(dataModel);
//计算最近邻域，邻居有两种算法，基于固定数量的邻居和基于相似度的邻居，这里使用基于固定数量的邻居
UserNeighborhood userNeighborhood = new NearestNUserNeighborhood(100, similarity, dataModel);
//构建推荐器，基于用户的协同过滤推荐
Recommender recommender = new GenericUserBasedRecommender(dataModel, userNeighborhood, similarity);
long start = System.currentTimeMillis();
//推荐商品
List
recommendedItemList = recommender.recommend(userId, howMany);
List itemIds = new ArrayList();
for (RecommendedItem recommendedItem : recommendedItemList) {
System.out.println(recommendedItem);
itemIds.add(recommendedItem.getItemID());
}
System.out.println("推荐出来的商品id集合"+itemIds);
//根据商品id查询商品
if(itemIds!=null &&itemIds.size()>0) {
list = itemMapper.findAllByIds(itemIds);
}else{
list = new ArrayList<>();
}
System.out.println("推荐数量:"+list.size() +"耗时："+(System.currentTimeMillis()-start));
} catch (TasteException e) {
e.printStackTrace();
}
return list;
}
@Override
public List- getRecommendItemsByItem(Long userId, Long itemId, int howMany) {
List
- list = null;
try {
//计算相似度，相似度算法有很多种，采用基于皮尔逊相关性的相似度
ItemSimilarity itemSimilarity = new PearsonCorrelationSimilarity(dataModel);
//4)构建推荐器，使用基于物品的协同过滤推荐
GenericItemBasedRecommender recommender = new GenericItemBasedRecommender(dataModel, itemSimilarity);
long start = System.currentTimeMillis();
// 物品推荐相似度，计算两个物品同时出现的次数，次数越多任务的相似度越高。
List
recommendedItemList = recommender.recommendedBecause(userId, itemId, howMany);
//打印推荐的结果
List itemIds = new ArrayList();
for (RecommendedItem recommendedItem : recommendedItemList) {
System.out.println(recommendedItem);
itemIds.add(recommendedItem.getItemID());
}
System.out.println("推荐出来的商品id集合"+itemIds);
//根据商品id查询商品
if(itemIds!=null &&itemIds.size()>0) {
list = itemMapper.findAllByIds(itemIds);
}else{
list = new ArrayList<>();
}
System.out.println("推荐数量:"+list.size() +"耗时："+(System.currentTimeMillis()-start));
} catch (TasteException e) {
e.printStackTrace();
}
return list;
}
} 3.7 编写controller
package cn.itcast.controller;
import cn.itcast.domain.Item;
import cn.itcast.service.RecommendService;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RestController;
import java.util.List;
@RestController
public class RecommendController {
@Autowired
private RecommendService recommendService;
/**
* 基于用户的推荐
* @param userId 用户id
* @param num 推荐数量
* @return
*/
@RequestMapping("recommendByUser")
public List- getRecommendItemsByUser(Long userId, int num){
List
- items= recommendService.getRecommendItemsByUser(userId,num);
return items;
}
/**
* 基于内容的推荐
* @param userId 用户id
* @param itemId 商品id
* @param num 推荐数量
* @return
*/
@RequestMapping("recommendByItem")
public List
- getRecommendItemsByItem(Long userId,Long itemId, int num){
List
- items= recommendService.getRecommendItemsByItem(userId,itemId, num);
return items;
}
}
3.8 测试
打开浏览器访问http://localhost:8080/recommendByUser?userId=5&num=10, 给id为5的用户推荐10个商品,输出:
[{"pid":50,"name":"Usual Suspects, The (1995)","types":"Crime|Mystery|Thriller"},{"pid":260,"name":"Star Wars: Episode IV - A New Hope (a.k.a. Star Wars) (1977)","types":"Action|Adventure|Sci-Fi"},{"pid":590,"name":"Dances with Wolves (1990)","types":"Adventure|Drama|Western"},{"pid":1732,"name":"Big Lebowski, The (1998)","types":"Comedy|Crime|Mystery|Thriller"},{"pid":2335,"name":"Waterboy, The (1998)","types":"Comedy"},{"pid":2478,"name":"Three Amigos (1986)","types":"Comedy|Western"},{"pid":4027,"name":"O Brother, Where Art Thou? (2000)","types":"Adventure|Comedy|Crime"},{"pid":4226,"name":"Memento (2000)","types":"Crime|Drama|Mystery|Thriller"},{"pid":5481,"name":"Austin Powers in Goldmember (2002)","types":"Comedy"},{"pid":5502,"name":"Signs (2002)","types":"Sci-Fi|Thriller"}]访问http://localhost:8080/recommendByItem?userId=5&itemId=231&num=10, 给id为5的用户推荐10个与231相似的商品,输出:
[{"pid":253,"name":"Interview with the Vampire: The Vampire Chronicles (1994)","types":"Drama|Horror"},{"pid":592,"name":"Batman (1989)","types":"Action|Crime|Sci-Fi|Thriller"}]4.总结
大家可以看得出来，用mahout这个算法库来做推荐系统写起来很简单，大致使用过程有：
步骤1：创建DataModel模型，可以基于文件File的DataModel，也可基于数据库的JDBCDataModel，如果数据库中表数据比较多，推荐耗时非常非常的慢，一般来说数据量都比较大可以基于文件DataModel模型来推荐，也可以将文件上传到hadoop，使用hadoop进行mapreduce计算，提高运算性能。
步骤2：采用欧几里得、皮尔逊等算法计算相似度。
步骤3：构建推荐器，基于用户或基于内容进行推荐。
步骤4：将推荐出来的商品id补全其他数据返回给用户展示。