Movie Review Selection Algorithm Based on Improved Spherical Fuzzy Semantic DEMATEL and TODIM Methods

In view of the abundance of movie reviews of varying quality available on movie review websites, current selection algorithms are ineffective, leading to ‘information overload.’ This study proposes a novel approach that begins by enhancing the score function to more precisely capture the fuzzy information encoded in the evaluation language. Utilizing spherical fuzzy sets, the method establishes an assessment index to evaluate movie reviews based on textual and non-textual features. It then determines the weights of these indicators using the Improved Spherical Fuzzy Semantic DEMATEL (Decision Making Trial and Evaluation Laboratory) method, followed by the application of the TODIM (an acronym in Portuguese for interactive and multicriteria decision making) method to rank the usefulness of movie reviews, considering the psychological factors influencing decision-makers across multiple groups. Sensitivity analysis and noise experiments are conducted to validate the approach. Sensitivity and noise analyses are conducted to validate the robustness of the approach. Comparative experiments confirm the effectiveness and scientific rigor of the proposed SFS-DEMATEL-TODIM method in assessing the usefulness of movie reviews.