mahout中a读e--学习笔记

the mahout achieves many ml algorithm ，such as recommend system， clustering ，classify and so on。。

primary recommend system。

data loaded :  datamodel  = newfilemodel(new file("    "));

                                               =mydatamodel.getdatamodel   // via data base    >sql

recommend method :  

(1)    user-based

1 2 3 4 5 6 7 8

DataModel  model = new  FileDataModel( new  File( "data/intro.csv" )); UserSimilarity similarity = new  PearsonCorrelationSimilarity(model); UserNeighborhood neighborhood = new  NearestNUserNeighborhood( 2 ,similarity,model); Recommender recommender=  new  GenericUserBasedRecommender(model,neighborhood,similarity); List<RecommendedItem> recommendations =recommender.recommend( 1 ,  1 ); for (RecommendedItem recommendation :recommendations){      System.out.println(recommendation);

(2)   item-based

1 2 3 4 5 6 7

DataModel  model = new  FileDataModel( new  File( "data/intro.csv" )); ItemSimilarity similarity = new  PearsonCorrelationSimilarity(model); Recommender recommender=  new  GenericItemBasedRecommender(model,similarity); List<RecommendedItem> recommendations =recommender.recommend( 1 ,  1 ); for (RecommendedItem recommendation :recommendations){      System.out.println(recommendation);

(3)  slope-based

1 2 3 4 5 6

DataModel  model = new  FileDataModel( new  File( "data/intro.csv" )); Recommender recommender=  new  SlopeOneRecommender(model); List<RecommendedItem> recommendations =recommender.recommend( 1 ,  1 ); for (RecommendedItem recommendation :recommendations){      System.out.println(recommendation);

(4) knn-based, svd-based ,tree-based

itemCF,itemKNN,SVD's Rrecision,Recall  are the best

frequently api:

• recommend(long userID, int howMany): 获得推荐结果，给userID推荐howMany个Item
• recommend(long userID, int howMany, IDRescorer rescorer): 获得推荐结果，给userID推荐howMany个Item，可以根据rescorer对结构重新排序。
• estimatePreference(long userID, long itemID): 当打分为空，估计用户对物品的打分
• setPreference(long userID, long itemID, float value): 赋值用户，物品，打分
• removePreference(long userID, long itemID): 删除用户对物品的打分
• getDataModel(): 提取推荐数据