Optimized Deep Learning Model for Lung Cancer Prediction Using ANN Algorithm
Deepak Rawat, Meenakshi Meenakshi, Lokesh Pawar, Gaurav Bathla, Ravi Kant
Abstract
Cancer is disease in which there is unnatural growth of abnormal cells that infest healthy cells in the body. Lung cancer invokes imbalance in cells which affects the lungs. Prediction of lung cancer is very necessary at the early stages especially in populated and less income countries. Specialists use clinical traditional procedures such as blood test and other therapies. Artificial intelligence era has started, now with the help of machine learning algorithm and deep learning algorithms it is possible to develop computer aided diagnosis mechanism. This research study has considered four machine learning algorithms such as Bayes Net, Naive Bayes, Decision Tree, Random Forest and one deep learning algorithm Artificial Neural Network to predict lung cancer at early stages. For this purpose, state of art parameters are measured for different algorithms and evaluated on lung cancer dataset. Accuracy as a prime parameter is evaluated and compared for all-the five algorithms. As per experimental result, Artificial Neural Network is best learning algorithm with accuracy of 92.23%. Further accuracy in Artificial Neural Network is evaluated for one, two and three hidden layers and compared. Artificial Neural Network with one layer has highest accuracy in experiment conducted.