Abstract:
In recent years, the fields of artificial intelligence have achieved significant advances in medicine. From detection to diagnosis, everything is now done utilizing artificial intelligence. Among all illnesses, lung cancer could be considered the most lethal as it is difficult to diagnose it in its preliminary stages. So, most people are usually unaware of having this disease. Early identification of lung cancer is critical for patient survival, and machine learning-based prediction models could predict lung cancer. Ensemble approaches are effective methods in Machine Learning for improving prediction accuracy. This study utilized the Deep Learning approach Convolutional Nural Network (CNN) to identify lung cancer, from various CT Scan pictures provided to the model. An Ensemble technique was created for this study to solve the issue of lung cancer detection. Instead of utilizing a single Deep Learning model, we create three CNNs to perform and forecast the outcome with more accuracy, and then assemble them using a voting mechanism. Our original dataset had 2500 images; after augmenting the dataset, it became around 7200 images. This study demonstrates the worth of ensemble approaches in improving the accuracy of lung cancer diagnosis with deep learning methods. We obtained better performance than individual CNN models by pooling their predictions via a voting process. Our findings emphasize the importance of the detection of lung cancer and highlight the capabilities of machine learning-based technologies.
