Enhancing breast cancer diagnosis accuracy through genetic algorithm-optimized multilayer perceptron

This research paper investigates the application of a genetic algorithm (GA) to optimize the performance of a Multilayer Perceptron (MLP) model for breast cancer diagnosis. Various configurations of hidden layers in the MLP are explored, and their corresponding accuracies range from 0.92 to 0.972. To ensure robust evaluation, k-fold cross-validation is employed, alongside comprehensive preprocessing of the dataset, including normalization, scaling, and encoding. These methodologies contribute to the model's consistent performance and its enhanced ability to generalize. However, the incorporation of a genetic algorithm significantly improves the accuracy range, resulting in values between about 0.97 and 0.99 across different generations. The genetic algorithm optimizes the MLP model by evolving a population of potential solutions (individuals) over multiple generations. Each individual represents a specific set of MLP parameters, such as the number of hidden layers, neurons per layer, and learning rate. The fitness of each individual is evaluated based on the MLP model's accuracy on the breast cancer dataset. The fittest individuals are selected for reproduction, and genetic operators like crossover and mutation are applied to generate new offspring. This iterative process of selection, crossover, and mutation gradually enhances the MLP model's performance. The primary objective of this research is to improve breast cancer diagnosis accuracy by leveraging the strengths of MLP and the optimization capabilities of the genetic algorithm. The results demonstrate that the combined approach effectively enhances the accuracy of breast cancer prediction compared to using MLP alone.