设计好了一个推荐程序，如何来完成评估呢？
一般是使用一个真实数据的样例作为测试数据来仿真，来看估计值和实际值的差别，0.0意味着完美的估计，就是没有差别。
一是使用平均差值（直观，易于理解），一是使用均方根。
针对mahout推荐1的推荐程序进行评估：
package mahout;
import java.io.File;
import org.apache.mahout.cf.taste.common.TasteException;
import org.apache.mahout.cf.taste.eval.RecommenderBuilder;
import org.apache.mahout.cf.taste.eval.RecommenderEvaluator;
import org.apache.mahout.cf.taste.impl.eval.AverageAbsoluteDifferenceRecommenderEvaluator;
import org.apache.mahout.cf.taste.impl.model.file.FileDataModel;
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.neighborhood.UserNeighborhood;
import org.apache.mahout.cf.taste.recommender.Recommender;
import org.apache.mahout.cf.taste.similarity.UserSimilarity;
import org.apache.mahout.common.RandomUtils;
public class TestRecommenderEvaluator {
public static void main(String[] args) throws Exception {
//强制每次生成相同的随机值，生成可重复的结果
RandomUtils.useTestSeed();
//数据装填
DataModel model = new FileDataModel(new File("data/intro.csv"));
//推荐评估，使用平均值
RecommenderEvaluator evaluator = new AverageAbsoluteDifferenceRecommenderEvaluator();
//推荐评估，使用均方差
//RecommenderEvaluator evaluator = new RMSRecommenderEvaluator();
//用于生成推荐引擎的构建器，与上一例子实现相同
RecommenderBuilder builder = new RecommenderBuilder() {
public Recommender buildRecommender(DataModel model) throws TasteException {
// TODO Auto-generated method stub
//用户相似度，多种方法
UserSimilarity similarity = new PearsonCorrelationSimilarity(model);
//用户邻居
UserNeighborhood neighborhood = new NearestNUserNeighborhood(2, similarity, model);
//一个推荐器
return new GenericUserBasedRecommender(model, neighborhood, similarity);
}
};
//推荐程序评估值（平均差值）训练90%的数据，测试数据10%，《mahout in Action》使用的是0.7,但是出现结果为NaN
double score = evaluator.evaluate(builder, null, model, 0.9, 1.0);
System.out.println(score);
}
}
结果：
14/08/04 09:16:21 INFO file.FileDataModel: Creating FileDataModel for file data\intro.csv 14/08/04 09:16:21 INFO file.FileDataModel: Reading file info... 14/08/04 09:16:21 INFO file.FileDataModel: Read lines: 21 14/08/04 09:16:21 INFO model.GenericDataModel: Processed 5 users 14/08/04 09:16:21 INFO eval.AbstractDifferenceRecommenderEvaluator: Beginning evaluation using 0.9 of FileDataModel[dataFile:D:\workspace\zoodemo\data\intro.csv] 14/08/04 09:16:21 INFO model.GenericDataModel: Processed 5 users 14/08/04 09:16:21 INFO eval.AbstractDifferenceRecommenderEvaluator: Beginning evaluation of 3 users 14/08/04 09:16:21 INFO eval.AbstractDifferenceRecommenderEvaluator: Starting timing of 3 tasks in 4 threads 14/08/04 09:16:21 INFO eval.StatsCallable: Average time per recommendation: 0ms 14/08/04 09:16:21 INFO eval.StatsCallable: Approximate memory used: 11MB / 62MB 14/08/04 09:16:21 INFO eval.StatsCallable: Unable to recommend in 2 cases 14/08/04 09:16:21 INFO eval.AbstractDifferenceRecommenderEvaluator: Evaluation result: 1.0 1.0
上面有一个：Unable to recommend in 2 cases，就是说有两个测试数据无法完成推荐。
2、如果将训练数据改为0.7或者0.8结果会是一个NaN
14/08/04 09:20:22 INFO file.FileDataModel: Creating FileDataModel for file data\intro.csv 14/08/04 09:20:22 INFO file.FileDataModel: Reading file info... 14/08/04 09:20:22 INFO file.FileDataModel: Read lines: 21 14/08/04 09:20:22 INFO model.GenericDataModel: Processed 5 users 14/08/04 09:20:22 INFO eval.AbstractDifferenceRecommenderEvaluator: Beginning evaluation using 0.7 of FileDataModel[dataFile:D:\workspace\zoodemo\data\intro.csv] 14/08/04 09:20:22 INFO model.GenericDataModel: Processed 5 users 14/08/04 09:20:22 INFO eval.AbstractDifferenceRecommenderEvaluator: Beginning evaluation of 4 users 14/08/04 09:20:22 INFO eval.AbstractDifferenceRecommenderEvaluator: Starting timing of 4 tasks in 4 threads 14/08/04 09:20:22 INFO eval.StatsCallable: Average time per recommendation: 0ms 14/08/04 09:20:22 INFO eval.StatsCallable: Approximate memory used: 11MB / 62MB 14/08/04 09:20:22 INFO eval.StatsCallable: Unable to recommend in 9 cases 14/08/04 09:20:22 INFO eval.AbstractDifferenceRecommenderEvaluator: Evaluation result: NaN NaN
Unable to recommend in 9 cases ,9个不能完成，太多了。
不过《Mahout in Action》使用的是0.7，出现了结果为1.0的情况，不知是怎么弄的。
3、当然我们也可以换成使用均方根的测试：结果如下（打开注释就可以了）
14/08/04 09:22:53 INFO file.FileDataModel: Creating FileDataModel for file data\intro.csv 14/08/04 09:22:53 INFO file.FileDataModel: Reading file info... 14/08/04 09:22:53 INFO file.FileDataModel: Read lines: 21 14/08/04 09:22:53 INFO model.GenericDataModel: Processed 5 users 14/08/04 09:22:53 INFO eval.AbstractDifferenceRecommenderEvaluator: Beginning evaluation using 0.9 of FileDataModel[dataFile:D:\workspace\zoodemo\data\intro.csv] 14/08/04 09:22:53 INFO model.GenericDataModel: Processed 5 users 14/08/04 09:22:53 INFO eval.AbstractDifferenceRecommenderEvaluator: Beginning evaluation of 3 users 14/08/04 09:22:53 INFO eval.AbstractDifferenceRecommenderEvaluator: Starting timing of 3 tasks in 4 threads 14/08/04 09:22:53 INFO eval.StatsCallable: Average time per recommendation: 0ms 14/08/04 09:22:53 INFO eval.StatsCallable: Approximate memory used: 11MB / 62MB 14/08/04 09:22:53 INFO eval.StatsCallable: Unable to recommend in 2 cases 14/08/04 09:22:53 INFO eval.AbstractDifferenceRecommenderEvaluator: Evaluation result: 1.0 1.0
原文：http://www.cnblogs.com/jsunday/p/3889383.html