A Comparative Study of Ovarian Cancer Prediction Using Machine Learning Method

The growth of cells in the ovaries is the cause of ovarian cancer. Healthy body tissue can be destroyed quickly by the cells, which can rapidly multiply. There are two ovaries in the female reproductive system, one on each side of the uterus. Eggs and the hormones estrogen and progesterone are produced by each of the almond-shaped ovaries. This work aims to perform predictive analytics for early diagnosis by combining statistical techniques and machine learning models with clinical data from 349 patients. The significant blood biomarkers are determined by applying the student’s t-test and the log fold changes of two groups in statistical analysis. In addition, a collection of machine learning models, such as Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), Extreme Gradient Boosting Machine (XGBoost), Logistic Regression (LR), Gradient Boosting Machine (GBM), and Light Gradient Boosting Machine (LGBM), are used in the creation of classification models to differentiate between patients with malignant ovarian cancer and those with benign ovarian cancer. Serum samples of human epididymis protein 4, carbohydrate antigen 125, carbohydrate antigen 19-9, and carcinoembryonic antigen were found to be the most useful biomarkers using both methods. Additional general chemistry tests included neutrophil ratio, thrombocytosis, hematocrit, calcium, indirect bilirubin, uric acid, and atriums. Additionally, the predictive analysis results demonstrate that, up to 98 percent of the time, the machine learning models can correctly distinguish between patients with benign and malignant conditions. Machine learning detection has the potential to significantly aid in the diagnosis of cancer because early-stage detection is typically unavailable.