Incremental Learning of Triadic PLSA for Collaborative Filtering

PLSA which was originally introduced in text analysis area, has been extended to predict user ratings in the collaborative filtering context, known as Triadic PLSA (TPLSA). It is a promising recommender technique but the computational cost is a bottleneck for huge data set. We design a incremental learning scheme for TPLSA for collaborative filtering task that could make forced prediction and free prediction as well. Our incremental implementation is the first of its kind in the probabilistic model based collaborative filtering area, to our best knowledge. Its effectiveness is validated by experiments designed for both rating-based and ranking-based collaborative filtering.