Comparison of Statistical Methods for Classification of Human Semen Data

Abstract

Abstract Classification is quite an important task in medical data analysis. The goal of classification is to build a concise model of the distribution of the dependent attribute in terms of the predictor attributes. Statistical classification is the problem of identifying the subpopulation to which new observations belong to whereas probabilistic classification returns a probability distribution for the class label attribute rather than returning a single class label. The aim is to classify human semen sample either as fertile or infertile using probabilistic and statistical classification algorithm. The accuracy of classifying the semen sample is assessed using Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis, Naive Bayes and Logistic Regression to distinguish infertile men from normal individuals. The comparisons were made in terms of sensitivity, specificity, precision, accuracy and error to determine the suitability of the statistical methods. The results have shown that QDA outperforms Logistic Regression, Naive Bayes and LDA in classifying the semen samples.