A game theory based many-objective hybrid tensor decomposition for skin cancer prediction

The development of skin cancer can be influenced by the abnormal expression of certain microRNAs (miRNAs). Current prediction models for miRNA-skin cancer associations have difficulties in maintaining both accuracy and comprehensiveness. To construct a comprehensive and interpretable skin cancer prediction model, various advanced tensor decomposition methods are organically combined, and a many-objective hybrid tensor decomposition model is proposed. In addition, due to the high computational cost of tensor decomposition, a many-objective optimization algorithm based on game theory was designed to solve the model. The game theory was used to dynamically adjust the diversity and convergence of the population, alleviate the pressure of solution selection, and improve the performance of the algorithm. The performance of the proposed algorithm is tested on a benchmark, and the prediction results of the many-objective hybrid tensor decomposition model are evaluated by a fivefold test method, and a case study of the prediction results is also presented. Experimental findings reveal that the proposed model and algorithm enhance overall performance by approximately 5.3%, compared to current advanced models.

Automated summary

This study proposes a many-objective hybrid tensor decomposition model and a many-objective optimization algorithm based on game theory for skin cancer prediction. Experimental findings demonstrate significant performance improvements compared to existing models.