Prediction System for the Lung Cancer Patients and Classification Accuracy Enhancement Using Ensemble Method

Lung cancer is the biggest challenge in the research world. Blood samples for lung cancer patients are significant for the prediction method in the research environment. Accurate prediction is essential to increase the survival period and may help to take proper medication. The feature selection is an essential part of this prediction system. In this method, the essential features are analyzed from lung cancer patient’s blood. Biomarkers can be identified from blood and other tests. Biomarkers are used in many scientific fields. The Serum CKAP4 levels were analyzed from lung cancer patient’s blood at the TNM stages. White blood counts and Hemoglobin are viewed as an analysis of immunity, and CKAP4 is located at an immunity level that has been recognized in the body fluid of patients with lung cancer. The system is analyzed immunity from collected real health information by a proposed Gaussian K_b Ensemble Weighting Classifier method. The novel attributes are composed of a cancer research center and cancer sanatoriums. The measurement of weighting value {low = 0, high = 1} of blood in Serum CKAP4, Hemoglobin, and the WBC are classified by immunity range. High immunity people’s survival rate is high at TNM stage 1. If not, high immunity was survival rate is low at TNM stage 1. Lung cancer patient’s survivability classification was performed by various supervised machine learning models such as Ada boost, neural networks, and SVM model. The Gaussian K_b Ensemble Weighting Classifier method helped to improve the classification accuracy by feature selection and the algorithm was implemented in the MATLAB environment. Kaplan–Meier curve method used to analyzes results. The classified labels are compared with existing results. The area proved the Gaussian K_b Ensemble weighting classifier algorithm under the Curve through the ROC method.