Optimized Predictive Models in Health Care Using Machine Learning

18TPLSTM-Based Deep ANN with Feature Matching Prediction of Lung Cancer

Thaventhiran Chandrasekar*, Praveen Kumar Karunanithi, A. Emily Jenifer and Inti Dhiraj

School of Computing, SASTRA University, Thanjavur, Tamilnadu, India

Abstract

Early detection of lung cancer is essential for accurate diagnosis and treatment recommendation. Lung cancer prognosis is the most critical issue in healthcare, especially given the exponential expansion of medical data. As a result, early cancer identification lowers mortality rates. However, this requires a lot more accuracy and effort. A model approach called TP-LSTM was created to quickly and correctly predict lung cancer. The suggested deep-learning model for analysing input patient data is composed of several layers. To forecast lung cancer at an early stage, many strategies are used at each strata. In deep neural learning, the input layer delivers the hidden layer features and data it has received from the dataset. Target Projection matching pursuit is employed in the feature selection process of the first hidden layer to quickly and precisely find the most crucial qualities for cancer prediction. Lung cancer is predicted using the patient data classification approach using LSTM at hidden layer 2 and the selected important features based on Czekanowski’s proximity factor. A number of patient data and parameters, such as accuracy, false-positive rates, and detection accuracy timeare empirically compared between the unique TP-LSTM methodology ...