Multiple Sclerosis Detection via Wavelet Entropy and Feedforward Neural Network Trained by Adaptive Genetic Algorithm

摘要

Multiple sclerosis is a disease that damages the central nervous system. Current medical treatments can only prevent or relieve symptoms. The target of this study is to improve the detection efficiency and classification accuracy. We propose a method based on wavelet entropy and feedforward neural network trained by adaptive genetic algorithm that is implemented over 10 runs of 10-fold cross validation. In which the wavelet entropy serves as a feature extractor and the feedforward neural network is employed as a classifier. Adaptive genetic algorithm work as a training algorithm. We also use the three-level decomposition of db2 wavelet to make a frequency analysis. According to the experimental results, the global optimization capability of adaptive genetic algorithm is more powerful than ordinary genetic algorithm. Comparing to the HWT-LR method, the accuracy of our method detection is higher.