EndoInsight: A Machine Learning Analysis of Endometriosis Data

Endometriosis is one of the gynecological conditions which is a very complex one. It is not easy to diagnose using conventional clinical methods. Any kind of delay in treating a patient with this kind of condition can lead to some aftereffects. To address this kind of challenge the current study evaluates and compares a variety of machine learning algorithms to automate the endometriosis by classifying baseline model such as logistic regression, decision trees, support vector machines and K nearest neighbors and then compare it with ensemble methods such as random forest, adaboost and xgboost. The models were trained on the clinical dataset which has demographic, symptomatic and biochemical attributes. The cross-validation technique is used on each classifier to assess an ensure stability. The performance for each of these classifiers are measured using accuracy precision recall and defense score along with the area under the Roc curve. After performing the classification, it is found that the ensemble algorithms have performed well than the individual based learners giving high accuracy in diagnosis hence the decision support tools for the detection of endometriosis may help clinicians in improving their diagnostic efficiency.