spark mllib 协同过滤推荐算法（ALS) python 实现 完整实例程序

# -*- coding:UTF-8 -*- import sys def PrepareMData(sc,path): print("========读取电影信息,from "+path+"/data/u.item") itemRDD=sc.textFile(path+"/data/u.item") #itemRDD.cache() movieTitle = itemRDD.map(lambda line: line.split("|")) \ .map(lambda a: (float(a[0]), a[1])) \ .collectAsMap() return movieTitle def loadModel(sc,path): #from pyspark.mllib.recommendation import ALS from pyspark.mllib.recommendation import MatrixFactorizationModel model=None try: model=MatrixFactorizationModel.load(sc,path+"/ALSmodel") model.userFeatures().cache() model.productFeatures().cache() print("=== 载入ALSmode，from "+path) except Exception as e: print "无法载入ALSmodel"+e.message return model def Recommend(model,movieTitle): if sys.argv[1]=="--U": RecommendMovies(model,movieTitle,int(sys.argv[2])) elif sys.argv[1]=="--M": RecommendUsers(model,movieTitle,int(sys.argv[2])) def RecommendMovies(model,movieTitle,inputUserID): recMovie=model.recommendProducts(inputUserID,10) print ("针对用户"+str(inputUserID)+" 推荐电影：") for rmd in recMovie: print("针对用户ID:{0},推荐电影ID:{1},电影名:{2},推荐评分:{3}").\ format (rmd[0],rmd[1],movieTitle[rmd[1]],rmd[2]) def RecommendUsers(model,movieTitle,inputMovieID): recUser=model.recommendUsers(inputMovieID,10) print ("针对电影id：{0} 电影名：{1},推荐下列用户ID：").\ format(inputMovieID,movieTitle[inputMovieID]) for rmd in recUser: print("用户ID:{0},推荐评分:{1}").\ format (rmd[0],rmd[2]) if __name__ == "__main__": from ALS.Com import * if len(sys.argv) != 3: print("请输入2个参数") exit(-1) sc = CreateSparkContext() print ("======== 推荐程序，数据准备 ========") path = SetPath(sc) movieTitle = PrepareMData(sc,path) print ("======== 推荐程序，载入模型 ========") model=loadModel(sc,path) #model.userFeatures().cache() #model.productFeatures().cache() if(model!=None): print ("====== 进行推荐 ==========") Recommend(model,movieTitle)