import numpy as np
from LoadData import load_rating_data, spilt_rating_dat
from sklearn.model_selection import train_test_split
from ProbabilisticMatrixFactorization import PMF
import glob
from os.path import exists, basename
import sys
if __name__ == "__main__":
for file_path in sorted(glob.glob("./monthly-data-bkp/monthly-data/"+sys.argv[1]+"*")):
print(file_path)
if( not exists("./monthly-pmf/"+basename(file_path)+"_u")):
print("Processing ...:"+basename(file_path));
pmf = PMF()
pmf.set_params({"num_feat": 30, "epsilon": 1, "_lambda": 0.1, "momentum": 0.8, "maxepoch": 30, "num_batches": 100,
"batch_size": 1000})
ratings = load_rating_data(file_path)
print(len(np.unique(ratings[:, 0])), len(np.unique(ratings[:, 1])), pmf.num_feat)
train, test = train_test_split(ratings, test_size=0.0) # spilt_rating_dat(ratings)
pmf.fit(train, test, basename(file_path))
# Check performance by plotting train and test errors
#plt.plot(range(pmf.maxepoch), pmf.rmse_train, marker='o', label='Training Data')
#plt.plot(range(pmf.maxepoch), pmf.rmse_test, marker='v', label='Test Data')
#plt.title('The MovieLens Dataset Learning Curve')
#plt.xlabel('Number of Epochs')
#plt.ylabel('RMSE')
#plt.legend()
#plt.grid()
import matplotlib.pyplot as plt
import numpy as np
from LoadData import load_rating_data, spilt_rating_dat
from sklearn.model_selection import train_test_split
from ProbabilisticMatrixFactorization import PMF
if __name__ == "__main__":
file_path = "data/ml-100k/u.data"
pmf = PMF()
pmf.set_params({
"num_feat": 10,
"epsilon": 1,
"_lambda": 0.1,
"momentum": 0.8,
"maxepoch": 10,
"num_batches": 100,
"batch_size": 1000
})
ratings = load_rating_data(file_path)
print(len(np.unique(ratings[:, 0])), len(np.unique(ratings[:, 1])),
pmf.num_feat)
train, test = train_test_split(ratings,
test_size=0.2) # spilt_rating_dat(ratings)
pmf.fit(train, test)
# Check performance by plotting train and test errors
plt.plot(range(pmf.maxepoch),
pmf.rmse_train,
marker='o',
label='Training Data')
plt.plot(range(pmf.maxepoch), pmf.rmse_test, marker='v', label='Test Data')
import matplotlib.pyplot as plt
import numpy as np
from LoadData import load_rating_data, spilt_rating_dat
from sklearn.model_selection import train_test_split
from ProbabilisticMatrixFactorization import PMF
if __name__ == "__main__":
file_path = "data/ml-100k/u.data"
pmf = PMF() # 建立模型并进行初始化
pmf.set_params({
"num_feat": 10,
"alpha": 0.01,
"lambda_u": 0.1,
"lambda_v": 0.1,
"max_epoch": 60,
"num_batches": 100,
"batch_size": 500
})
ratings = load_rating_data(file_path) # 获得数据集
train, test = train_test_split(ratings, test_size=0.2) # 将数据集拆分为训练集和测试集
pmf.train(train, test) # 训练
rlt = pmf.predict(3)
idx = np.argsort(rlt)
idx = idx[::-1]
size = rlt.shape[0]
print("对用户3推荐的电影为：")
for i in range(size):
print("电影：{}，推荐值为：{:.2f}".format(idx[i], rlt[idx[i]]))
gmm = GaussianMixture(n_components=args.K)
gmm.fit(obs)
gmm_z = gmm.predict_proba(obs)
fout.create_dataset("gmm/z", data=gmm_z)
fout.create_dataset("gmm/components", data=gmm.means_)
fout.create_dataset("gmm/covariances", data=gmm.covariances_)
fout.create_dataset("gmm/weights", data=gmm.weights_)
# Probablistic Matrix Factorization
from ProbabilisticMatrixFactorization import PMF
# reshape data for input
if args.PMF_bool:
N = obs.shape[0] * obs.shape[1]
x = np.arange(0, obs.shape[0], 1)
y = np.arange(0, obs.shape[1], 1)
xv, yv = np.meshgrid(x, y)
l3 = np.reshape(obs, (N, 1))
l2 = np.reshape(yv.T, (N, 1))
l1 = np.reshape(xv.T, (N, 1))
Y = np.concatenate((l1, l2, l3), axis=1)
pmf = PMF()
pmf.set_params({"num_feat": args.K})
pmf.fit(Y, Y)
fout.create_dataset("pmf/z", data=pmf.w_User)
fout.create_dataset("pmf/components", data=pmf.w_Item.T)
fout.close()
import matplotlib.pyplot as plt
import numpy as np
from LoadData import load_rating_data, spilt_rating_dat
from sklearn.model_selection import train_test_split
from ProbabilisticMatrixFactorization import PMF
if __name__ == "__main__":
file_path = "data/ml-100k/u.data"
pmf = PMF()
pmf.set_params({"num_feat": 10, "epsilon": 1, "_lambda": 0.1, "momentum": 0.8, "maxepoch": 10, "num_batches": 100,
"batch_size": 1000})
ratings = load_rating_data(file_path)
print(len(np.unique(ratings[:, 0])), len(np.unique(ratings[:, 1])), pmf.num_feat)
train, test = train_test_split(ratings, test_size=0.2) # spilt_rating_dat(ratings)
pmf.fit(train, test)
# Check performance by plotting train and test errors
plt.plot(range(pmf.maxepoch), pmf.rmse_train, marker='o', label='Training Data')
plt.plot(range(pmf.maxepoch), pmf.rmse_test, marker='v', label='Test Data')
plt.title('The MovieLens Dataset Learning Curve')
plt.xlabel('Number of Epochs')
plt.ylabel('RMSE')
plt.legend()
plt.grid()
plt.show()
print("precision_acc,recall_acc:" + str(pmf.topK(test)))