Deep Learning Architecture for Collaborative Filtering Recommender Systems

Abstract

This paper provides an innovative deep learning architecture to improve collaborative filtering results in recommender systems. It exploits the potential of the reliability concept to raise predictions and recommendations quality by incorporating prediction errors (reliabilities) in the deep learning layers. The underlying idea is to recommend highly predicted items that also have been found as reliable ones. We use the deep learning architecture to extract the existing non-linear relations between predictions, reliabilities, and accurate recommendations. The proposed architecture consists of three related stages, providing three stacked abstraction levels: (a) real prediction errors, (b) predicted errors (reliabilities), and (c) predicted ratings (predictions). In turn, each abstraction level requires a learning process: (a) Matrix Factorization from ratings, (b) Multilayer Neural Network fed with real prediction errors and hidden factors, and (c) Multilayer Neural Network fed with reliabilities and hidden factors. A complete set of experiments has been run involving three representative and open datasets and a state-of-the-art baseline. The results show strong prediction improvements and also important recommendation improvements, particularly for the recall quality measure.

Application of sample balance-based multi-perspective feature ensemble learning for prediction of user purchasing behaviors on mobile wireless network platforms

“…According to algorithm principles, CF algorithm can be mainly divided into two types: memory-based CF and model-based CF [3,5], where the former seeks for neighbors having high similarity to user or commodity mainly through user's historical information and predicts user purchasing behaviors according to their comprehensive commodity evaluation. Memory-based CF has data sparsity problem with weak extensibility [13]. Different from memory-based CF, model-based CF trains model mainly through user ratings for commodities and then predicts user purchasing probability [14].…”

Application of Sample Balance-Based Multi-Perspective Feature Ensemble Learning for Prediction of User Purchasing Behaviors on Mobile Wireless Network Platforms

“…According to algorithm principles, CF algorithm can be mainly divided into two types: memory-based CF and model-based CF [1,2] , where the former seeks for neighbors having high similarity to user or commodity mainly through user's historical information and predicts user purchasing behaviors according to their comprehensive commodity evaluation. Memory-based CF has data sparsity problem with weak extensibility [9] . Different from memory-based CF, model-based CF trains model mainly through user ratings for commodities and then predicts user purchasing probability [10] .…”

“…8. N L l    (9) Slicing input data is the key to realizing multi-granularity deep forest. Assume "sliding window" size as   (10) Slice size is lL WW  , and slice number of the whole dataset is: (11) After multi-granularity scanning, the slices of all input data are input into the random forest If there are RF n random forests, the quantity of class vectors generated through "sliding window" sampling is: (13) The final output data size after multi-granularity scanning is: (14) Fig .…”

“…features (1), (2), (3), (5) and (10). In addition, time-related features (4), (7) and (9) are also about user purchasing behaviors 3 days before the prediction date, indicating that the closer the user's historical purchasing behavior to the prediction date, the greater the influence on prediction of user purchasing behavior. The proportional features (5) and (7) of quantity of one commodity purchased by one user to total quantity of commodity purchased, and historical purchase life cycle (6) of one user for one commodity also have a certain effect on user purchasing behavioral prediction.…”

Recommendation System Using Autoencoders

“…Today, with the concept of deep learning becoming more important, several researchers have begun to test its usability in a collaborative filtering approach in order to achieve better results [20]. For example, Van den Oord et al made use of deep convolutional neural network to give music recommendations [21].…”