earticle

영어

This study has modeled numerous confounding parameters of seminal quality for the purpose of digging out the hidden relationship between these seminal parameters using Bayesian Belief Network (BBN). The data source for this study was retrieved from UCI machine learning repository. Etiological patterns were derived out of complex relationship of nine related attributes. We have shown that as compared to conventional statistical measures, BBN is quite useful in analysis of seminal quality as well as classifying an unknown instance. The outcome is composed of a predictive probabilistic model which can classify any new instance whether the seminal quality is altered or not. The observed accuracy of the model is highest (91%) whereas the previous highest accuracy was reported to be 86% only.