A new model for lung cancer prediction based on differential evolution algorithm and effective feature selection

Abstract

Lung cancer is one of the most dangerous and fatal diseases worldwide. By using advanced machine learning techniques and optimization algorithms, early prediction and diagnosis of this disease can be achieved. Early identification of lung cancer is an important approach that can increase the survival rate of patients. In this paper, a novel method for lung cancer prediction is proposed, which combines two important techniques: Support Vector Machine (SVM) and Differential Evolution (DE) algorithm. Firstly, using the differential evolution algorithm, important and suitable features for lung cancer prediction are extracted. Then, using the SVM classifier, a classification model is built for prediction. The proposed approach is implemented on two lung cancer databases and achieves a good level of accuracy, which is compared with four other methods: C4.5 decision tree, neural network, Naive Bayes classifier, and logistic regression. The proposed model, with high accuracy and generalization power, is a suitable model for lung cancer detection and can serve as a strong decision support system alongside medical professionals.

Keywords

• lung cancer
• support vector machine
• differential evolution algorithm
• feature selection