MusciBase

exp3/src/MusicBase/MatSim.py

@@ -0,0 +1,65 @@
+import numpy as np
+
+MusicNum = 21602
+MIN_NUM = -30000
+data_path = '../../data/DoubanMusic.txt'
+output_path = '../../output/'
+
+
+def JaccobiSim(list1:list, list2:list):
+ intersection = len(set(list1).intersection(set(list2)))
+ union = len(list1) + len(list2) - intersection
+ return intersection/union
+ # intersection = 0
+ # i = 0
+ # j = 0
+ # while i < len(list1) and j < len(list2):
+ # if list1[i] == list2[j]:
+ # intersection += 1
+ # i += 1
+ # j += 1
+ # elif list1[i] < list2[j]:
+ # i += 1
+ # else:
+ # j += 1
+ # union = len(list1) + len(list2) - intersection
+ # return intersection/union
+
+def MatSimGen():
+ MapMusicID = {}
+ MatSim = np.zeros((MusicNum, MusicNum))
+
+ f = open(data_path, 'r')
+ lines = f.readlines()
+ f.close()
+
+ for line in lines:
+ temp = line.split()
+ UserID = int(temp[0])
+ for pair in temp[1:]:
+ MusicID = int(pair.split(',')[0])
+ if MusicID not in MapMusicID:
+ MapMusicID[MusicID] = [UserID]
+ else:
+ MapMusicID[MusicID].append(UserID)
+
+ num = 0
+ for Music1 in MapMusicID:
+ num += 1
+ if num %10 == 0:
+ print(num)
+ for Music2 in MapMusicID:
+ MatSim[Music1][Music2] = JaccobiSim(MapMusicID[Music1], MapMusicID[Music2])
+ np.save('MatJaccobiSim.npy', MatSim)
+
+def MatDiagMinimize():
+ MatSim = np.load('MatJaccobiSim.npy')
+ for i in range(MusicNum):
+ MatSim[i][i] = MIN_NUM
+ np.save('MatJaccobiSim_DiagMinimized.npy', MatSim)
+
+if __name__ == '__main__':
+ MatDiagMinimize()
+ # MatSimGen()
+
+

exp3/src/MusicBase/main.py

@@ -0,0 +1,57 @@
+import numpy as np
+
+MatSimPath = 'MatJaccobiSim_DiagMinimized.npy'
+data_path = '../../data/DoubanMusic.txt'
+result_path = 'MusicBase.txt'
+MusicNum = 21602
+
+def main(CarryOn = False):
+ MatSim = np.load(MatSimPath)
+ UserMap = {}
+
+ f = open(data_path, 'r')
+ lines = f.readlines()
+ f.close()
+
+ for line in lines:
+ temp = line.split()
+ UserID = int(temp[0])
+ for pair in temp[1:]:
+ MusicID = int(pair.split(',')[0])
+ if UserID not in UserMap:
+ UserMap[UserID] = [MusicID]
+ else:
+ UserMap[UserID].append(MusicID)
+
+ if CarryOn:
+ f = open(result_path, 'r')
+ count = len(f.readlines())
+ f.close()
+ f = open(result_path, 'a')
+ else:
+ f = open(result_path, 'w')
+ count = 0
+
+ for UserID in UserMap:
+ if UserID < count:
+ continue
+ predict_list = []
+ for MusicID in range(MusicNum):
+ predict = 0
+ for UserMusic in UserMap[UserID]:
+ predict += MatSim[UserMusic][MusicID]
+ predict_list.append((predict,MusicID))
+ predict_list.sort(reverse=True)
+ f.write(str(UserID))
+ f.write('\t')
+ f.write(str(predict_list[0][1]))
+ for i in range(1,100):
+ f.write(',')
+ f.write(str(predict_list[i][1]))
+ f.write('\n')
+ count += 1
+ if count % 1 == 0:
+ print(count)
+
+if __name__ == '__main__':
+ main(CarryOn=True)