用Hadoop流实现mapreduce版推荐系统基于物品的协同过滤算法

       以个性化新闻推荐为例，整个过程分成两个mapreduce阶段，由于hadoop流不支持多个mapreduce过程的自动化，所以所有mapreduce过程命令必须人工一个一个的执行。

1、首先需要将原始数据处理成如下形式的两个文件

文件一：Item_user_score.txt

格式：物品—用户—分数

如下图中第一行，物品100655565被用户1634974浏览过，则将分数记为1

文件二：Item_Item_number.txt

格式：物品—物品—相似度

如下图中第二行，物品100654360与物品100650498同时被两个用户浏览过

2、矩阵乘法

其实文件一和文件二分别保存着一个矩阵，第二步就是要做矩阵乘法。

step1mapper.py

#!/usr/bin/env python"""A more advanced Reducer, using Python iterators and generators."""from itertools import groupbyfrom operator import itemgetterimport sysdef read_mapper_output(file, separator='\t'):for line in file:yield line.strip().split(separator,1)def main(separator='\t'):# input comes from STDIN (standard input)datas = read_mapper_output(sys.stdin, separator=separator)# groupby groups multiple word-count pairs by word,# and creates an iterator that returns consecutive keys and their group:#   current_word - string containing a word (the key)#   group - iterator yielding all ["<current_word>", "<count>"] items#print datafor data in datas:print "%s%s%s" % (data[0], separator, data[1])if __name__ == "__main__":main()

step1mapper.py的主要目的是将两个文件中的键值对取出，然后相同的key输入到同一个reducer。

step1reducer.py

#!/usr/bin/env python"""A more advanced Reducer, using Python iterators and generators."""from itertools import groupbyfrom operator import itemgetterimport sysdef read_mapper_output(file, separator='\t'):for line in file:yield line.split()def main(separator='\t'):# input comes from STDIN (standard input)data = read_mapper_output(sys.stdin, separator=separator)# groupby groups multiple word-count pairs by word,# and creates an iterator that returns consecutive keys and their group:#   current_word - string containing a word (the key)#   group - iterator yielding all ["<current_word>", "<count>"] items#print datadatanew={}for key in data:if key[0] not in datanew.keys():datanew[key[0]]=[]datanew[key[0]].append([key[1],key[2]])for line in datanew.keys():user_score=[]item_number=[]for temp in datanew[line]:if int(temp[0])/1000000 < 99:user_score.append(temp)else :item_number.append(temp)for attr1 in user_score:for attr2 in item_number:print "%s%s%s" % (attr1[0], separator, attr2[0]+separator+str(float(attr1[1])*int(attr2[1])))if __name__ == "__main__":main()

step1reducer.py的输出形式为：1634974 10065436034.0

可以理解为用户1634974对新闻100654360的喜爱程度为34.

执行命令为：

  ~/hadoop-2.3.0/bin/hadoop jar hadoop-streaming-2.3.0.jar -mapper step1mapper.py -reducer step1reducer.py -input /input/test.txt -output /step1out.txt -file step1mapper.py -file step1reducer.py  

3、结果融合

将上一个过程reducer的输出作为该过程mapper阶段的输入。

step2mapper.py

#!/usr/bin/env python"""A more advanced Reducer, using Python iterators and generators."""from itertools import groupbyfrom operator import itemgetterimport sysdef read_mapper_output(file, separator='\t'):for line in file:yield line.strip().split(separator,1)def main(separator='\t'):# input comes from STDIN (standard input)datas = read_mapper_output(sys.stdin, separator=separator)# groupby groups multiple word-count pairs by word,# and creates an iterator that returns consecutive keys and their group:#   current_word - string containing a word (the key)#   group - iterator yielding all ["<current_word>", "<count>"] items#print datafor data in datas:print "%s%s%s" % (data[0], separator, data[1])if __name__ == "__main__":main()

step2reducer.py

推荐最可能的5个

#!/usr/bin/env python"""A more advanced Reducer, using Python iterators and generators."""import operatorfrom itertools import groupbyfrom operator import itemgetterimport sysdef read_mapper_output(file, separator='\t'):for line in file:yield line.split()def main(separator='\t'):# input comes from STDIN (standard input)data = read_mapper_output(sys.stdin, separator=separator)# groupby groups multiple word-count pairs by word,# and creates an iterator that returns consecutive keys and their group:#   current_word - string containing a word (the key)#   group - iterator yielding all ["<current_word>", "<count>"] items#print datadatanew={}for key in data:#print "\n--------------------------"if key[0] not in datanew.keys():datanew[key[0]]=[]datanew[key[0]].append([key[1],key[2]])for key in datanew.keys():result={}for data in datanew[key]:if data[0] not in result.keys():result[data[0]]=0.0result[data[0]] += float(data[1])for rec_item in sorted(result.iteritems(),key=operator.itemgetter(1),reverse=True)[:5]:print "%s\t%s\t%s" % (key, rec_item[0], rec_item[1])if __name__ == "__main__":main()

执行命令为：

  ~/hadoop-2.3.0/bin/hadoop jar hadoop-streaming-2.3.0.jar -mapper step2mapper.py -reducer step2reducer.py -input/output/step1out.txt -output /step2out.txt -file step2mapper.py -file step2reducer.py  

输出形式为：

PS：

其实两个mapreduce阶段的mapper都是冗余的，完全可以在执行命令的时候加入参数：

-D stream.map.input.field.separator='\t'   //输入文件中的数据每行都以tab作为分隔符

-D stream.num.map.input.key.fields=1       //每行第一个tab之前的数据作为key，其余的作为value

这样两个mapreduce过程就可以合并成一个mapreduce过程执行。