def __init__(self, users, eps, MinPts, reduce_dimensions=False): if reduce_dimensions: autoencoder = AutoEncoder(users) users = autoencoder.reduce_dimensions( ) # num_dimensions should be bigger than 4. else it runs for 4. self.userLength = users.shape[0] self.eps = eps self.MinPts = MinPts self.DistanceMatrix = self.__calcDistanceMatrix(users)
def __init__(self, train, test, reduce_dimensions=False): self.train_data = train self.test_data = test labels = self.extractLabels(self.train_data) if reduce_dimensions: train_autoencoder = AutoEncoder(self.train_data) self.train_data = train_autoencoder.reduce_dimensions() # num_dimensions should be bigger than 4. else it runs for 4. test_autoencoder = AutoEncoder(self.test_data) self.test_data = test_autoencoder.reduce_dimensions() self.train_data["label"] = labels
def __init__(self, train, test, reduce_dimensions=False): self.train_data = train self.test_data = test self.train_data = self.train_data.sort_values( by='popularity', ascending=False ) # test data is also sorted, so in this way i can find them and discriminate them self.train_data = self.train_data[self.test_data.shape[ 0]:] # keep for train only the data that isnt contained in test self.train_data.reset_index(drop=True, inplace=True) labels = self.extractLabels() if reduce_dimensions: train_autoencoder = AutoEncoder(self.train_data) self.train_data = train_autoencoder.reduce_dimensions( ) # num_dimensions should be bigger than 4. else it runs for 4. test_autoencoder = AutoEncoder(self.test_data) self.test_data = test_autoencoder.reduce_dimensions() self.train_data["label"] = labels