''' 基于物品的协同推荐 矩阵数据 说明: 1.修正的余弦相似度是一种基于模型的协同过滤算法。我们前面提过,这种算法的优势之 一是扩展性好,对于大数据量而言,运算速度快、占用内存少。 2.用户的评价标准是不同的,比如喜欢一个歌手时有些人会打4分,有些打5分;不喜欢时 有人会打3分,有些则会只给1分。修正的余弦相似度计算时会将用户对物品的评分减去 用户所有评分的均值,从而解决这个问题。 ''' import pandas as pd from io import StringIO #数据类型一:csv矩阵(用户-商品)(适用于小数据量) csv_txt = '''"user","Blues Traveler","Broken Bells","Deadmau5","Norah Jones","Phoenix","Slightly Stoopid","The Strokes","Vampire Weekend" "Angelica",3.5,2.0,,4.5,5.0,1.5,2.5,2.0 "Bill",2.0,3.5,4.0,,2.0,3.5,,3.0 "Chan",5.0,1.0,1.0,3.0,5,1.0,, "Dan",3.0,4.0,4.5,,3.0,4.5,4.0,2.0 "Hailey",,4.0,1.0,4.0,,,4.0,1.0 "Jordyn",,4.5,4.0,5.0,5.0,4.5,4.0,4.0 "Sam",5.0,2.0,,3.0,5.0,4.0,5.0, "Veronica",3.0,,,5.0,4.0,2.5,3.0,''' #数据类型一:csv矩阵(用户-商品)(适用于小数据量) csv_txt2 = '''"user","Kacey Musgraves","Imagine Dragons","Daft Punk","Lorde","Fall Out Boy" "David",,3,5,4,1 "Matt",,3,4,4,1 "Ben",4,3,,3,1 "Chris",4,4,4,3,1 "Tori",5,4,5,,3''' #数据类型一:csv矩阵(用户-商品)(适用于小数据量) #根据《data minning guide》第85页的users2数据 csv_txt3 = '''"user","Taylor Swift","PSY","Whitney Houston" "Amy",4,3,4 "Ben",5,2, "Clara",,3.5,4 "Daisy",5,,3''' df = None #方式一:加载csv数据 def load_csv_txt(): global df, csv_txt, csv_txt2, csv_txt3 df = pd.read_csv(StringIO(csv_txt3), header=0, index_col="user") #测试:读取数据 load_csv_txt() #======================================= # 注意:不需要build_xy #======================================= # 计算两个物品相似度 def computeSimilarity(goods1, goods2): '''根据《data minning guide》第71页的公式s(i,j)''' # 每行的用户评分都减去了该用户的平均评分 df2 = df[[goods1, goods2]].sub(df.mean(axis=1), axis=0).dropna(axis=0) #黑科技 # 返回修正的余弦相似度 return sum(df2[goods1] * df2[goods2]) / (sum(df2[goods1]**2) * sum(df2[goods2]**2))**0.5 # csv_txt #print('\n测试:计算Blues Traveler与Broken Bells的相似度') #print(computeSimilarity("Blues Traveler","Broken Bells")) # csv_txt2 #print('\n测试:计算Kacey Musgraves与Imagine Dragons的相似度') #print(computeSimilarity("Kacey Musgraves","Imagine Dragons")) # 计算给定用户对物品的可能评分 def p(user, goods): '''根据《data minning guide》第75页的公式p(u,i)''' assert pd.isnull(df.ix[user, goods]) # 必须用户对给定物品尚未评分 s1 = df.ix[user, df.ix[user].notnull()] #用户对已打分物品的打分数据 s2 = s1.index.to_series().apply(lambda x:computeSimilarity(x, goods)) #打分物品分别与给定物品的相似度 return sum(s1 * s2) / sum(abs(s2)) # csv_txt2 #print('\n测试:计算David对Kacey Musgraves的可能打分') #print(p("David","Kacey Musgraves")) #为了让公式的计算效果更佳,对物品的评价分值最好介于-1和1之间 def rate2newrate(rate): '''根据《data minning guide》第76页的公式NR(u,N)''' ma, mi = df.max().max(), df.min().min() return (2*(rate - mi) - (ma - mi))/(ma - mi) #已知rate2newrate求newrate2rate def newrate2rate(new_rate): '''根据《data minning guide》第76页的公式R(u,N)''' ma, mi = df.max().max(), df.min().min() return (0.5 * (new_rate + 1) * (ma - mi)) + mi print('\n测试:计算3的new_rate值') print(rate2newrate(3)) print('\n测试:计算0.5的rate值') print(newrate2rate(0.5)) # 计算给定用户对物品的可能评分(对评分进行了修正/还原) def p2(user, goods): '''根据《data minning guide》第75页的公式p(u,i)''' assert pd.isnull(df.ix[user, goods]) # 必须用户对给定物品尚未评分 s1 = df.ix[user, df.ix[user].notnull()] #用户对已打分物品的打分数据 s1 = s1.apply(lambda x:rate2newrate(x)) #修正 s2 = s1.index.to_series().apply(lambda x:computeSimilarity(x, goods)) #已打分物品分别与给定物品的相似度 return newrate2rate(sum(s1 * s2) / sum(abs(s2)))#还原 # csv_txt2 #print('\n测试:计算David对Kacey Musgraves的可能打分(修正)') #print(p2("David","Kacey Musgraves")) #================================== # 下面是Slope One算法 # # 两个步骤: # 1. 计算差值 # 2. 预测用户对尚未评分物品的评分 #================================== # 1.计算两物品之间的差异 def dev(goods1, goods2): '''根据《data minning guide》第80页的公式dev(i,j)''' s = (df[goods1] - df[goods2]).dropna() d = sum(s) / s.size return d, s.size #返回差异值,及权值(同时对两个物品打分的人数) # csv_txt2 #print('\n测试:计算Kacey Musgraves与Imagine Dragons的分数差异') #print(dev("Kacey Musgraves","Imagine Dragons")) #计算所有两两物品之间的评分差异,得到方阵pd.DataFrame(行对列) def get_dev_table(): '''根据《data minning guide》第87页的表''' goods_names = df.columns.tolist() df2 = pd.DataFrame(.0, index=goods_names, columns=goods_names) #零方阵 for i,goods1 in enumerate(goods_names): for goods2 in goods_names[i+1:]: d, _ = dev(goods1, goods2) # 注意:只取了物品差异值 df2.ix[goods1, goods2] = d df2.ix[goods2, goods1] = -d # 对称的位置取反 return df2 print('\n测试:计算所有两两物品之间的评分差异表') print(get_dev_table()) #预测某用户对给定物品的评分 # 加权Slope One算法 def slopeone(user, goods): '''根据《data minning guide》第82页的公式p(u,j)''' s1 = df.ix[user].dropna() #用户对已打分物品的打分数据 s2 = s1.index.to_series().apply(lambda x:dev(goods, x)) #待打分物品与已打分物品的差异值及权值 s3 = s2.apply(lambda x:x[0]) #差异值 s4 = s2.apply(lambda x:x[1]) #权值 #print(s1, s3, s4) return sum((s1 + s3) * s4)/sum(s4) print('\n测试:预测用户Ben对物品Whitney Houston的评分') print(slopeone('Ben', 'Whitney Houston')) # 3.375