Seminal quality prediction using data mining methods.

摘要: Background Now-a-days, some new classes of diseases have come into existences which are known as lifestyle diseases. The main reasons behind these changes in the people such alcohol drinking, smoking, food habits etc. After going through various diseases, it has been found that fertility rates (sperm quantity) men considerably decreasing last two decades. Lifestyle factors well environmental mainly responsible for change semen quality. Objective objective this paper is to identify and features affects seminal quality also rate man using data mining methods. Method five artificial intelligence techniques Multilayer perceptron (MLP), Decision Tree (DT), Navie Bayes (Kernel), Support vector machine+Particle swarm optimization (SVM+PSO) machine (SVM) applied on dataset evaluate predict person either normal or having altered rate. While eight feature selection support (SVM), neural network (NN), evolutionary logistic regression (LR), plus particle (SVM+PSO), principle component analysis (PCA), chi-square test, correlation T-test methods used more relevant affect These contains 100 instances with nine attribute classes. Results experimental result shows SVM+PSO provides higher accuracy area under curve (AUC) (94% & 0.932) among multi-layer (MLP) (92% 0.728), Vector Machines (91% 0.758), (Kernel) (89% 0.850) 0.735) parameters. This focuses process i.e. how select important prediction In paper, find out a set good features. investigational results childish (0.079) high fever (0.057) less impact while age (0.8685), season (0.843), surgical intervention (0.7683), consumption (0.5992), smoking habit (0.575), number hours spent setting (0.4366) accident (0.5973) impact. It observed increase above mentioned (multilayer 92%, 91%, 94%, 89% decision tree 89%) compared without 86%, 85%, 83% 84%) applicability prediction. Conclusion lightens application medical domain. From can be concluded disease based parameters/features rather than undergoing test. provide accurate tree.