Mahout推荐引擎测评及源码分析

评估一个推荐程序：

利用平均值来评分：AverageAbsoluteDifferenceRecommenderEvaluator

利用均方根来评分：RMSRecommenderEvaluator

public void recommenderEvaluator() throws Exception{
		DataModel dataModel = new FileDataModel(new File("F:\\tdunning-MiA-5b8956f\\intro.csv"));
		//选择评测模型（均方根评测）
		RecommenderEvaluator recommenderEvaluator = new RMSRecommenderEvaluator();
		//recommenderBuilder对象主要是为了根据传入的dataModel模型建立推荐引擎模型
		RecommenderBuilder recommenderBuilder = new RecommenderBuilder() {
			@Override
			public Recommender buildRecommender(DataModel dataModel) throws TasteException {
				UserSimilarity similarity = new LogLikelihoodSimilarity(dataModel);
				UserNeighborhood neigborhood = new NearestNUserNeighborhood(2, similarity, dataModel);
				return new GenericUserBasedRecommender(dataModel, neigborhood, similarity);
			}
		};
		
		//根据上面建立的recommenderBuilder对象和dataModel对象，取数据的70%做测试，预测剩下的30%，查看评测效果
		double evaluate = recommenderEvaluator.evaluate(recommenderBuilder, null, dataModel, 0.7, 1.0);
		System.out.println(evaluate);
	}

测评函数evaluate源码分析：