Bias Mitigation in News Recommender Systems Using Preference Correction

This paper presents a novel approach for mitigating biases in news recommendation systems while also incorporating a user's personal preferences to provide a personalised recommendations. The proposed hybrid content and collaborative-based filtering system ranks news categories according to weights based on user history, resulting in a recommendation system that recommends news articles from a purely unbiased perspective. The method of “preference correction” is incorporated to further recommend articles that the user will also prefer to read, improving the overall user experience. This approach can be applied to any news dataset that collects user impressions and has the potential to enhance awareness among readers and burst filter bubbles.