DESIGNING A NEURAL NETWORK MODEL USING K-MEANS CLUSTERING FOR RISK ANALYSIS OF LUNG CANCER DISEASE

Abstract

According to the World Health Organization report in 2004, lung cancer belongs to highest mortality rate cancer type compared to others. Genetics and early starting smoke etc. become the basis for lung cancer risk. In recent years, lung cancer cases are increasing with the use of cigarettes at younger ages. One of the most important factor in the treatment of the disease is early diagnosis. Artificial intelligence methods, which have been used in many areas in recent years, are also used for early diagnosis and imaging of diseases. In this study, a hybrid artificial neural network (ANN) model was designed to bring a different perspective to the use of multilayer ANN in the literature for lung cancer risk prediction. Lung cancer risk factors were used as input data in predicting the disease. We tried to estimate the results using clustered data by K-means clustering algorithm and multi-layered ANN method. When the results obtained from the normalized and clustered data set are compared with the results in the literature, the proposed model has a higher accuracy value than the other methods.