Stratifying Success Of In-Vitro Fertilization Using Chi-Kruskal’s Hybrid Feature Extractor And Machine Learning Classifiers

Abstract

Fertility problems have become more common in recent times, with many challenged with quotidian routines that can contrive infertility amongst couples. The pivotal causes to these detriments can be the unhealthy living style, high stress levels and less to no exercise. The realm of medical sciences is burgeoned extensively to render progressive results, nonetheless simulative and swifter analysis of various parameters that trigger the success and failure of the In-Vitro Fertilization (IVF) process remains to defy the demands. The currently established empirical studies delineate various statistical and computational implementation effectuated toward the prediction of IVF, but fails to elaborate on the correlation of attributes in relevance to augmenting classifier accuracy of machine learning models. The proposed study thereby effectuates a stepwise analysis from pre-processing the dataset to efficiently unsheathing the features of vitality, and further to incorporate an assortment of machine learning classifiers to explicitly comprehend the performance of the networks. This paper proposes a fused feature extractor using the Chi-Square and the Kruskal Wallis feature extraction methods to form a hybrid feature extractor. Furthermore, the machine learning classifiers such as Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), Logit Boosting, Gradient Boosting Machine (GBM), Decision Tree (DT) and an ensemble classifier are entailed to cognize the stratification efficacy of the IVF data. The process of hyperparameter tuning, along with classification indagation is simulated in MATLAB, and the results procured in each phase of implementation aids to render better medical decisions for clinicians in IVF treatment formulation, and customized healthcare provisions for patients.