Slope One Predictors for Online Rating-Based Collaborative Filtering

Abstract

Rating-based collaborative filtering is the process of predicting how a user would rate a given item from other user ratings. We propose three related slope one schemes with predictors of the form f (x) = x + b, which precompute the average difference between the ratings of one item and another for users who rated both. Slope one algorithms are easy to implement, efficient to query, reasonably accurate, and they support both online queries and dynamic updates, which makes them good candidates for real-world systems. The basic SLOPE ONE scheme is suggested as a new reference scheme for collaborative filtering. By factoring in items that a user liked separately from items that a user disliked, we achieve results competitive with slower memorybased schemes over the standard benchmark EachMovie and Movielens data sets while better fulfilling the desiderata of CF applications.

Does Wikipedia Information Help Netflix Predictions?

“…They are similar to the results from a slope one predictor [8] (RMSE roughly 0.984), which is equivalent to our implementation of k-NN with all similarities set to 1. This suggests that the movie similarities computed using Wikipedia content are only weakly indicative of user rating patterns.…”

A Neural Networks Approach for Improving the Accuracy of Multi-Criteria Recommender Systems

“…• Slope one algorithm: The slope one is a model-based CF algorithm proposed by D. Lemire et al [19], which was derived from item-based CF technique. It was named slope one algorithm and was proposed to overcome some of the issues encountered in CF-based RSs.…”

An item/user representation for recommender systems based on bloom filters

“…This can be used also in collaborative filtering methods. Indeed, in memory-based collaborative filtering techniques, users are represented by a vector of preferences where each dimension corresponds to one item and each value is the degree of interest in this item [4], [5], [10], [11]. Then in order to find out similar users, correlations and similarities among these vectors are performed.…”