# movies[j] = [np.random.rand(embedding_dim) for k in range(40)] # samples = [] # for i in range(10): # for j in range(15): # samples.append((i, j, float(np.random.randint(0, 5)))) # save(pst, samples, users, movies) samples, users, movies = load(pst) cf = CFUtil(samples) fullSimilarity = cf.simUser() adp = Adapt(samples, users, movies, fullSimilarity, userMaxLen, movieMaxLen, neiMaxLen, sim_thresh, embedding_dim) User_train_test, Movie_train_test, Neigh_train_test, Y_train_test, sflSmps = adp.kerasInput( ) X_train_test = { 'user': User_train_test, 'movie': Movie_train_test, 'nei': Neigh_train_test } att = NeuralModel(attParamDic, epoch, l2) history, tesLoss, predicts = att.build(X_train_test, Y_train_test, usingNeiModel) comparison = [] for i in range(len(predicts)): smp = sflSmps[i] comparison.append([smp[0], smp[1], smp[2], predicts[i][0]]) pst.recordResult(history, tesLoss, comparison, fileModifier='l2_e-2')
'user': [uhid, userMaxLen, embedding_dim], 'movie': [mhid, movieMaxLen, embedding_dim], 'nei': [nhid, neiMaxLen, embedding_dim] } path = os.path.abspath('.') pst = PostProcess(path) # execute cf = CFUtil(samples) fullSimilarity = cf.simUser() print("Adapt data...") adp = Adapt(samples, users, movies, fullSimilarity, userMaxLen, movieMaxLen, neiMaxLen, sim_thresh, embedding_dim) User_train_test, Movie_train_test, Neigh_train_test, Y_train_test, sflSmps = adp.kerasInput( ) X_train_test = { 'user': User_train_test, 'movie': Movie_train_test, 'nei': Neigh_train_test } att = NeuralModel(attParamDic, epoch, l2) history, tesLoss, predicts = att.build(X_train_test, Y_train_test, usingNeiModel) comparison = [] for i in range(len(predicts)): smp = sflSmps[i] comparison.append([smp[0], smp[1], smp[2], predicts[i][0]]) pst.recordResult(history, tesLoss, comparison, fileModifier=jobName)
# users = {} # for i in range(10): # users[i] = [np.random.rand(embedding_dim) for k in range(20)] # movies = {} # for j in range(15): # movies[j] = [np.random.rand(embedding_dim) for k in range(40)] # samples = [] # for i in range(10): # for j in range(15): # samples.append((i, j, float(np.random.randint(0, 5)))) # save(pst, samples, users, movies) samples, users, movies = load(pst) cf = CFUtil(samples) fullSimilarity = cf.simUser() adp = Adapt(samples, users, movies, fullSimilarity, userMaxLen, movieMaxLen, neiMaxLen, sim_thresh, embedding_dim) User_train_test, Movie_train_test, Neigh_train_test, Y_train_test = adp.kerasInput( ) att = NeuralModel(userParams, movieParams, neiParams, epoch, l2) model, history, tesLoss = att.build(User_train_test[0], Movie_train_test[0], Neigh_train_test[0], Y_train_test[0], User_train_test[1], Movie_train_test[1], Neigh_train_test[1], Y_train_test[1]) pst.recordResult(model, history, tesLoss)