Source code for cornac.models.pmf.recom_pmf

# -*- coding: utf-8 -*-

"""
@author: Aghiles Salah
"""

import numpy as np
import scipy.sparse as sp
import pmf
from ..recommender import Recommender
from ...utils.util_functions import sigmoid
from ...utils.util_functions import which_
from ...utils.util_functions import map_to
from ...utils.util_functions import clipping


[docs]class PMF(Recommender): """Probabilistic Matrix Factorization. Parameters ---------- k: int, optional, default: 5 The dimension of the latent factors. max_iter: int, optional, default: 100 Maximum number of iterations or the number of epochs for SGD. learning_rate: float, optional, default: 0.001 The learning rate for SGD_RMSProp. gamma: float, optional, default: 0.9 The weight for previous/current gradient in RMSProp. lamda: float, optional, default: 0.001 The regularization parameter. name: string, optional, default: 'PMF' The name of the recommender model. variant: {"linear","non_linear"}, optional, default: 'non_linear' Pmf variant. If 'non_linear', the Gaussian mean is the output of a Sigmoid function.\ If 'linear' the Gaussian mean is the output of the identity function. trainable: boolean, optional, default: True When False, the model is not trained and Cornac assumes that the model already \ pre-trained (U and V are not None). rating_range: 1d array, optional, default: [None,None] The minimum and maximum rating values, e.g., [1,5]. init_params: dictionary, optional, default: {'U':None,'V':None} List of initial parameters, e.g., init_params = {'U':U, 'V':V}. \ U: a csc_matrix of shape (n_users,k), containing the user latent factors. \ V: a csc_matrix of shape (n_items,k), containing the item latent factors. References ---------- * Mnih, Andriy, and Ruslan R. Salakhutdinov. Probabilistic matrix factorization. \ In NIPS, pp. 1257-1264. 2008. """ def __init__(self, k=5, max_iter=100, learning_rate = 0.001,gamma = 0.9, lamda = 0.001, name = "pmf", variant ='non_linear', trainable = True, rating_range = [None,None] ,init_params = {'U':None,'V':None}): Recommender.__init__(self,name=name, trainable = trainable) self.k = k self.init_params = init_params self.max_iter = max_iter self.learning_rate = learning_rate self.gamma = gamma self.lamda = lamda self.variant = variant self.ll = np.full(max_iter, 0) self.eps = 0.000000001 self.U = init_params['U'] #matrix of user factors self.V = init_params['V'] #matrix of item factors self.min_rating = rating_range[0] self.max_rating = rating_range[1] #fit the recommender model to the traning data
[docs] def fit(self,X): """Fit the model to observations. Parameters ---------- X: scipy sparse matrix, required the user-item preference matrix (traning data), in a scipy sparse format\ (e.g., csc_matrix). """ if self.min_rating is None: self.min_rating = np.min(X.data) if self.max_rating is None: self.max_rating = np.max(X.data) if self.trainable: #converting data to the triplet format (needed for cython function pmf) (rid,cid,val)=sp.find(X) val = np.array(val,dtype='float32') if self.variant == 'non_linear': #need to map the ratings to [0,1] if[self.min_rating,self.max_rating] != [0,1]: val = map_to(val,0.,1.,self.min_rating,self.max_rating) rid = np.array(rid,dtype='int32') cid = np.array(cid,dtype='int32') tX = np.concatenate((np.concatenate(([rid], [cid]), axis=0).T,val.reshape((len(val),1))),axis = 1) del rid, cid, val print('Learning...') if self.variant == 'linear': res = pmf.pmf_linear(tX,k = self.k,n_X= X.shape[0], d_X = X.shape[1], n_epochs = self.max_iter,lamda = self.lamda, learning_rate= self.learning_rate,gamma = self.gamma, init_params = self.init_params) elif self.variant == 'non_linear': res = pmf.pmf_non_linear(tX,k = self.k,n_X= X.shape[0], d_X = X.shape[1], n_epochs = self.max_iter,lamda = self.lamda, learning_rate= self.learning_rate,gamma = self.gamma, init_params = self.init_params) else: raise ValueError('variant must be one of {"linear","non_linear"}') self.U = sp.csc_matrix(res['U']) self.V = sp.csc_matrix(res['V']) print('Learning completed') else: print('%s is trained already (trainable = False)' % (self.name))
[docs] def score(self, user_index, item_indexes = None): """Predict the scores/ratings of a user for a list of items. Parameters ---------- user_index: int, required The index of the user for whom to perform score predictions. item_indexes: 1d array, optional, default: None A list of item indexes for which to predict the rating score.\ When "None", score prediction is performed for all test items of the given user. Returns ------- Numpy 1d array Array containing the predicted values for the items of interest """ if item_indexes is None: user_pred = self.V.todense()*self.U[user_index,:].T.todense() else: user_pred = self.V[item_indexes,:].todense()*self.U[user_index,:].T.todense() user_pred = np.array(user_pred,dtype='float64').flatten() if self.variant == "non_linear": user_pred = sigmoid(user_pred) user_pred = map_to(user_pred,self.min_rating,self.max_rating,0.,1.) else: #perform clipping to enforce the predictions to lie in the same range as the original ratings user_pred = clipping(user_pred,self.min_rating,self.max_rating) return user_pred
[docs] def rank(self, user_index, known_items = None): """Rank all test items for a given user. Parameters ---------- user_index: int, required The index of the user for whom to perform item raking. known_items: 1d array, optional, default: None A list of item indices already known by the user Returns ------- Numpy 1d array Array of item indices sorted (in decreasing order) relative to some user preference scores. """ u_pref_score = np.array(self.score(user_index)) if known_items is not None: u_pref_score[known_items] = None rank_item_list = (-u_pref_score).argsort() # ordering the items (in decreasing order) according to the preference score return rank_item_list