Model for Predicting Survivability of Soft Tissue Sarcoma Cancer Patients Using Stacking Classifier

Abstract

Soft tissue sarcoma is type of cancer that affects the blood vessels, deep skin tissues, muscles, fat, cartilage, tendons, ligaments or any other tissue that connects, supports or surrounds any other body organs. Most of the soft tissue sarcomas originate in arms or legs. Initially patient discoversanovergrown massthat may or may not be painful. According to an estimate about 13,460 people will be diagnosed with Soft tissue sarcoma in the year 2021 in United States of America. Out of this about 5,350 people will die because of Soft tissue sarcoma this year.

This study proposesa modelthat uses a stacking machine learning classifier for predicting the 5 years survivability, after diagnosis, of a Soft tissue sarcoma patient. Initial stageof this stacking classifier constitutes the Random Forest and LinearSVC classifiers and the LogisticRegression is used as the final estimator. To express the validity of the derived model the study uses the accuracy, recall and precision scores. Further this study also compares the various performance metrics from individual classifiers used.