Mahout基於內存的DataMode 推薦引擎Demo2

Mahout基於內存的DataMode 推薦引擎Demo2

//註釋的部分是基於文件也可以理解爲基於日誌文件的， 

//DataModel 可以有很多種，實現abstractDataMode的子類，原則上都可以作爲數據源，個人覺得，不管是那種DataMode各自有優缺點 

//應該視情況而定， 

package com.test.mahout; 

import java.util.List; 

import org.apache.mahout.cf.taste.impl.common.FastByIDMap; 
import org.apache.mahout.cf.taste.impl.model.GenericDataModel; 
import org.apache.mahout.cf.taste.impl.model.GenericUserPreferenceArray; 
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.PreferenceArray; 
import org.apache.mahout.cf.taste.neighborhood.UserNeighborhood; 
import org.apache.mahout.cf.taste.recommender.RecommendedItem; 
import org.apache.mahout.cf.taste.recommender.Recommender; 
import org.apache.mahout.cf.taste.similarity.UserSimilarity; 

class RecommenderIntro { 
    public static void main(String[] args) throws Exception { 
       
       
    /*    FastByIDMap<PreferenceArray> preferences =new FastByIDMap<PreferenceArray>(); 
     * 
            PreferenceArray prefsForUser1 = new GenericUserPreferenceArray(10); A 
           
            prefsForUser1.setUserID(0, 1L); 
            prefsForUser1.setItemID(0, 101L); B 
            prefsForUser1.setValue(0, 3.0f); B 
           
            prefsForUser1.setItemID(1, 102L); 
            prefsForUser1.setValue(1, 4.5f); 
           
            … (8 more) 
            preferences.put(1L, prefsForUser1); C 
            DataModel model = new GenericDataModel(preferences); D 
            //A Set up PreferenceArray for user 1 
            //B Add the first of 10 preferences 
            //C Attach user 1’s preference*/ 
       
        // laod Data File 
       
//        DataModel model = new FileDataModel(new File("C:\\mahout_data.txt")); 
       
      FastByIDMap<PreferenceArray> preferences =new FastByIDMap<PreferenceArray>(); 

/*FastByIDMap是經過優化了的key-value數據結夠，在這裏用來存儲基本的元數據，《mahout in action》書中對這個數據結構做了詳細的解釋*/ 
       
        PreferenceArray prefsForUser1=new GenericUserPreferenceArray(3);//注意這裏的數字 
      // 這裏是用來存儲一個用戶的元數據，這些元數據通常來自日誌文件，比如瀏覽歷史，等等，不同的業務場合，它的業務語義應該是不一樣 
        prefsForUser1.setUserID(0, 1);/*個人對這裏也感覺到非常模糊，其實這裏保存用戶ID完全用不着key-value結構，也用不着用一個long型的ID，這裏個人覺得是爲了以後擴展或者，保持統一的格式，注意用就是了*/ 


       
        prefsForUser1.setItemID(0, 101); 
        prefsForUser1.setValue(0, 5.0f); 
       
        prefsForUser1.setItemID(1, 102); 
        prefsForUser1.setValue(1, 3.0f); 
       
       
        prefsForUser1.setItemID(2, 103); 
        prefsForUser1.setValue(2, 2.5f); 
       
       
         preferences.put(1l, prefsForUser1);//在這裏添加數據 
       
       
         PreferenceArray prefsForUser2=new GenericUserPreferenceArray(4); 
         
         prefsForUser2.setUserID(0, 2); 
         
         prefsForUser2.setItemID(0, 101); 
         prefsForUser2.setValue(0, 2.0f); 
         
         prefsForUser2.setItemID(1, 102); 
         prefsForUser2.setValue(1,2.5f); 
         
         prefsForUser2.setItemID(2, 103); 
         prefsForUser2.setValue(2,5.0f); 
         

         prefsForUser2.setItemID(3,104); 
         prefsForUser2.setValue(3,2.0f); 
         
         
         preferences.put(2l, prefsForUser2); 
         
         
         PreferenceArray prefsForUser3=new GenericUserPreferenceArray(4); 
         
         
         prefsForUser3.setUserID(0, 3); 
         
         prefsForUser3.setItemID(0, 101); 
         prefsForUser3.setValue(0, 2.5f); 
         
         prefsForUser3.setItemID(1, 104); 
         prefsForUser3.setValue(1, 4.0f); 

        
         prefsForUser3.setItemID(2, 105); 
         prefsForUser3.setValue(2, 4.5f); 
         
         prefsForUser3.setItemID(3, 107); 
         prefsForUser3.setValue(3, 5.0f); 
         
         preferences.put(3l, prefsForUser3); 
         
         
         PreferenceArray prefsForUser4=new GenericUserPreferenceArray(4); 
         
         prefsForUser4.setUserID(0, 4); 
         
         prefsForUser4.setItemID(0,101); 
         prefsForUser4.setValue(0, 5.0f); 
         
         prefsForUser4.setItemID(1,103); 
         prefsForUser4.setValue(1, 3.0f); 
         
         prefsForUser4.setItemID(2,104); 
         prefsForUser4.setValue(2,4.5f); 
         
         
         prefsForUser4.setItemID(3, 106); 
         prefsForUser4.setValue(3, 4.0f); 
         
         preferences.put(4l, prefsForUser4); 
         
         
         
         PreferenceArray prefsForUser5=new GenericUserPreferenceArray(6); 
         
         prefsForUser5.setUserID(0, 5); 
         
         prefsForUser5.setItemID(0, 101); 
         prefsForUser5.setValue(0, 4.0f); 
         
         prefsForUser5.setItemID(1, 102); 
         prefsForUser5.setValue(1, 3.0f); 
         
         prefsForUser5.setItemID(2, 103); 
         prefsForUser5.setValue(2, 2.0f); 
         
         
         prefsForUser5.setItemID(3, 104); 
         prefsForUser5.setValue(3, 4.0f); 
         
         prefsForUser5.setItemID(4, 105); 
         prefsForUser5.setValue(4, 3.5f); 
         
         prefsForUser5.setItemID(5, 106); 
         prefsForUser5.setValue(5, 4.0f); 
         
         preferences.put(5l, prefsForUser5); 
         
        DataModel model=new  GenericDataModel(preferences) ;//DataModel的建立 
        UserSimilarity similarity = new PearsonCorrelationSimilarity(model); 
       
        UserNeighborhood neighborhood = new NearestNUserNeighborhood(2,similarity, model); 

        // Create the recommender engine 
        Recommender recommender = new GenericUserBasedRecommender(model,neighborhood, similarity); 

        // C For user 1, recommend 1 item 
        List<RecommendedItem> recommendations = recommender.recommend(1, 1); 
        for (RecommendedItem recommendation : recommendations) { 
            System.out.println(recommendation); 
        } 
       
       
       
    } 
}