Абстрактный

Toward an improved collaborative filtering algorithm for omics data adjusted by double factors

Li-Ping Li, Guang-Li Xu, Wen-Xia Ding

A comprehensive recommendation algorithm adjusted by double factor based on improved Particle Swarm Optimization (PSO) and K-means was proposed to further improve algorithm performance in the high throughput omics data filtering. It uses User Behavior Factor (UBF) to adjust similarity. Meanwhile, it also introduces Global Supplement Factor (GSF) to adjust parameters in the adjacent phase selection and supplement items. The experiment shows that the improved algorithm can achieve good efficiency and recommendation accuracy. The application of this algorithm in biomarker feature set filtering has also been evaluated in this study.