Alcoholism Detection by Wavelet Entropy and Support Vector Machine Trained by Genetic Algorithm

Nowadays, alcoholism becomes a more serious social problem, and we proposed a method to help doctors to detect the alcoholism patients. In our method, wavelet entropy (WE) was used to extract the features, support vector machine (SVM) was used classify the samples, and genetic algorithm (GA) was used to optimize our classifier. In our method, the average sensitivity is 88.42±1.74%, the average specificity is 88.93±1.62%, and the average accuracy is 88.68±0.30%, which is better than the three state-of-the-art approaches: FA-PNN, HMI, and FRFT. Our method is effective in alcoholism detection and can help doctors to reduce the massive detection work.