A Multidimensional Model for Recommendation Systems Based on Classification and Entropy

The proliferation of false and redundant information on e-commerce platforms as well the prevalence ineffective recommendations other untrustworthy behaviors has seriously impeded healthy development these platforms. To address issues enhance prediction accuracy user trust, contemporary recommendation systems often utilize additional (i.e., side information). In this work, we propose a model to improve quality by employing entropy user-item ratings. was used reflect global rating behavior item. We also utilized classification item heuristic quality. our best result, achieved significant improvement 8.2% in accuracy. classified items users’ actual preference, which is more trustworthy for users. evaluated with three real-world datasets. performance proposed significantly better than baseline methods. similarity calculation method employed present potential mitigate data sparsity problem associated correlation-based similarity. weight matrix zero sparsity. Furthermore, favorable computational complexity compared conventional k-nearest neighbor method.