/** 基本算法 */
public static void demo1(DataModel m) throws Exception {
System.out.println("==============用户,打分,平均差值(0)==============");
// 推荐结果
UserSimilarity s = MySimilarity.getPearsonCorrelation(m);
UserNeighborhood n = MyNeighborhood.getNearestN(m, s, 2);
RecommenderBuilder rb = MyRecommender.userBuilder(s, n);
MyRecommender.showItems(rb.buildRecommender(m).recommend(26, 1));
RecommenderEvaluator e = new AverageAbsoluteDifferenceRecommenderEvaluator();
double score = e.evaluate(rb, null, m, 0.9, 1.0); // testing data, all
// data
System.out.println("算法得分: " + score);
System.out.println();
}
public static void prefEvaluate(DataModel m) throws Exception {
System.out.println("==============用户,打分,评估(1)==============");
// 推荐结果
UserSimilarity s = MySimilarity.getPearsonCorrelation(m);
UserNeighborhood n = MyNeighborhood.getNearestN(m, s, 2);
// 算法评估
RecommenderBuilder rb = MyRecommender.userBuilder(s, n);
RecommenderIRStatsEvaluator e = new GenericRecommenderIRStatsEvaluator();
IRStatistics stats =
e.evaluate(rb, null, m, null, 1, GenericRecommenderIRStatsEvaluator.CHOOSE_THRESHOLD, 1);
System.out.println("查准率 : " + stats.getPrecision());
System.out.println("全查率 : " + stats.getRecall());
System.out.println();
}
public static void demo2(DataModel m) throws Exception {
System.out.println("==============用户,打分,均方根(0)==============");
// 推荐结果
UserSimilarity s = MySimilarity.getPearsonCorrelation(m);
UserNeighborhood n = MyNeighborhood.getNearestN(m, s, 2);
RecommenderBuilder rb = MyRecommender.userBuilder(s, n);
MyRecommender.showItems(rb.buildRecommender(m).recommend(26, 1));
// RandomUtils.useTestSeed();
RecommenderEvaluator e = new RMSRecommenderEvaluator();
double score = e.evaluate(rb, null, m, 0.7, 0.3); // testing data, all
// data
System.out.println("算法得分: " + score);
System.out.println();
}
public static void demo3(DataModel m) throws Exception {
System.out.println("==============用户,无打分,平均差值(0)==============");
// 推荐结果
UserSimilarity s = new LogLikelihoodSimilarity(m);
UserNeighborhood n = MyNeighborhood.getNearestN(m, s, 2);
RecommenderBuilder rb = MyRecommender.userBuilder(s, n);
MyRecommender.showItems(rb.buildRecommender(m).recommend(1, 1));
// 算法得分
// RandomUtils.useTestSeed();
RecommenderEvaluator e = new AverageAbsoluteDifferenceRecommenderEvaluator();
DataModelBuilder db = MyDataModel.createNoPrefDataModelBuilder();
double score = e.evaluate(rb, db, m, 0.7, 1.0); // testing data, all data
System.out.println("算法得分: " + score);
System.out.println();
}
public static void noPrefEvaluate(DataModel m) throws Exception {
System.out.println("==============用户,无打分,评估(1)==============");
// 推荐结果
UserSimilarity s = new LogLikelihoodSimilarity(m);
UserNeighborhood n = MyNeighborhood.getNearestN(m, s, 2);
// 算法评估
RecommenderBuilder rb = MyRecommender.userBuilder(s, n);
RecommenderIRStatsEvaluator e = new GenericRecommenderIRStatsEvaluator();
DataModelBuilder db = MyDataModel.createNoPrefDataModelBuilder();
IRStatistics stats =
e.evaluate(rb, db, m, null, 1, GenericRecommenderIRStatsEvaluator.CHOOSE_THRESHOLD, 1.0);
System.out.println("查准率 : " + stats.getPrecision());
System.out.println("查全率 : " + stats.getRecall());
System.out.println();
}