mahout基本结构

package: org.apache.mahout.cf.taste

• common: 公共类包括，异常，数据刷新接口，权重常量
• eval: 定义构造器接口，类似于工厂模式
• model: 定义数据模型接口
• neighborhood: 定义近邻算法的接口
• recommender: 定义推荐算法的接口
• similarity: 定义相似度算法的接口
• transforms: 定义数据转换的接口
• hadoop: 基于hadoop的分步式算法的实现类
• impl: 单机内存算法实现类

common

Refreshable接口：

 public interface Refreshable { 
  void refresh(Collection<Refreshable> alreadyRefreshed); 
 }

数据模型

DataModel类结构图

相似度算法

• 基于用户的相似度

public interface UserSimilarity extends Refreshable { 
  double userSimilarity(long userID1, long userID2) throws TasteException; 
  void setPreferenceInferrer(PreferenceInferrer inferrer); 
 }

用户相似度算法结构图

• 基于物品的相似度

public interface ItemSimilarity extends Refreshable { 
  double itemSimilarity(long itemID1, long itemID2) throws TasteException; 
  double[] itemSimilarities(long itemID1, long[] itemID2s) throws TasteException; 
  long[] allSimilarItemIDs(long itemID) throws TasteException;
}

物品相似度算法结构图

近邻算法

public interface UserNeighborhood extends Refreshable {  
  long[] getUserNeighborhood(long userID) throws TasteException;
}

近邻算法结构图

推荐算法

public interface Recommender extends Refreshable { 
  List<RecommendedItem> recommend(long userID, int howMany) throws TasteException; 
  List<RecommendedItem> recommend(long userID, int howMany, boolean includeKnownItems) throws TasteException; 
  List<RecommendedItem> recommend(long userID, int howMany, IDRescorer rescorer) throws TasteException; 
  List<RecommendedItem> recommend(long userID, int howMany, IDRescorer rescorer, boolean includeKnownItems)   throws TasteException;
  float estimatePreference(long userID, long itemID) throws TasteException; 
  void removePreference(long userID, long itemID) throws TasteException;
  DataModel getDataModel();
}

推荐算法结构图

参考

http://blog.fens.me/mahout-recommend-engine/