PSO Optimized RBFNN Classifier For Lung Cancer Identification System

Lung cancer is the maximum collective cause of mortality in the world, with a greater fatality rate than other cancer types because of its late stage of development and difficulties in diagnosing symptoms. Therefore, early lung cancer identification improves the chances of a favourable outcome and is essential for diagnosis. It is the trickiest method to increase a patient's odds of surviving. This paper presents the construction of an adaptive structure for lung cancer estimation based on a combination of architectural evolution with weight learning using neural networks and Particle Swarm Optimization (PSO) optimized and Radial Basis Function Neural Network (RBFNN) classification methods. This approach offered various versions and hybridized it with an optimization approach to boost its efficiency. It employs global searching of PSO and local searching capabilities of neural network to calculate lung cancer as cancerous or non-cancerous. The organization is completed, and the results are compared to the performance of other algorithms. The parameters derived from the feature extraction Grey Level Co-occurrence Matrix (GLCM) technique are used to build the architecture of the RBFNN model. Based on the patient's state, the forecasting system helps physicians make appropriate judgments. The RBFNN has an accuracy of 93.5%.