A hybrid recommendation algorithm based on weighted stochastic block model

TL;DR

This paper introduces a hybrid recommendation system leveraging a weighted stochastic block model to enhance accuracy and address cold-start issues by combining content-based and collaborative filtering techniques.

Contribution

It presents a novel hybrid recommendation algorithm based on WSBM that improves prediction accuracy and effectively handles cold-start problems.

Findings

Outperforms traditional hybrid recommendation methods in accuracy

Effectively addresses cold-start problem

Shows improved prediction and classification results

Abstract

Hybrid recommendation usually combines collaborative filtering with content-based filtering to exploit merits of both techniques. It is widely accepted that hybrid filtering outperforms the single algorithm, thus it has been the new trend in electronic commerce these years. In this paper, we propose a novel hybrid recommendation system based on weighted stochastic block model (WSBM). Our algorithm not only makes full use of content-based and collaborative filtering recommendation to solve the cold-start problem but also improves the accuracy of recommendation by selecting the nearest neighbor with WSBM. The experiment result shows that our proposed approach has better prediction and classification accuracy than traditional hybrid recommendation.