package utils;

import org.apache.mahout.cf.taste.common.TasteException;

import org.apache.mahout.cf.taste.eval.IRStatistics;

import org.apache.mahout.cf.taste.eval.RecommenderBuilder;

import org.apache.mahout.cf.taste.eval.RecommenderIRStatsEvaluator;

import org.apache.mahout.cf.taste.impl.eval.GenericRecommenderIRStatsEvaluator;

import org.apache.mahout.cf.taste.impl.model.jdbc.ConnectionPoolDataSource;

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.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.Recommender;

import org.apache.mahout.cf.taste.similarity.UserSimilarity;

import com.mysql.jdbc.jdbc2.optional.MysqlDataSource;

public class RecommendUtils {

public static void evaluate() throws Exception{

ConnectionPoolDataSource connectionPool = new ConnectionPoolDataSource(getRecommendDateSource());

JDBCDataModel model = new MySQLJDBCDataModel(connectionPool, "userperference", "student_Id", "competition_Id", "perferences", "time");

//创建评估器

RecommenderIRStatsEvaluator evaluator = new GenericRecommenderIRStatsEvaluator();

//构建推荐引擎

RecommenderBuilder recommenderBuilder = new RecommenderBuilder() {

public Recommender buildRecommender(DataModel dataModel)

throws TasteException {

// TODO Auto-generated method stub

UserSimilarity similarity = new PearsonCorrelationSimilarity(

dataModel);

UserNeighborhood neighborhood = new NearestNUserNeighborhood(2,

similarity, dataModel);

return new GenericUserBasedRecommender(dataModel, neighborhood,

similarity);

}

};

int at = 1;// 评估推荐两个ITEM时的查准率和查全率

//开始评估查准率和查全率

IRStatistics statistics = evaluator.evaluate(recommenderBuilder, null,model, null, at,GenericRecommenderIRStatsEvaluator.CHOOSE_THRESHOLD,

1.0);

System.out.println("precision at " + at + " 查准率:"+ statistics.getPrecision());

System.out.println("recall at " + at + " 查全率:" + statistics.getRecall());

}

public static MysqlDataSource getRecommendDateSource(){

MysqlDataSource dataSource = new MysqlDataSource();

dataSource.setServerName("localhost");

dataSource.setUser("root");

dataSource.setPassword("mysql");

dataSource.setDatabaseName("dcms");

return dataSource;

}

}