Enhancing of Classification and Prediction of Gastric Cancer by Using Lime AI

Abstract

Artificial Intelligence is improving healthcare by allowing more accurate prediction and identification of diseases. Deep Learning, which is a part of AI, is good at examining complicated medical information. Stomach cancer is a serious health issue that often leads to high death rates. Detecting this disease early and correctly categorizing it are key to effective treatment. The use of AI and Deep Learning increases the accuracy of diagnosis and helps doctors take promptly. In existing systems, stomach cancer prediction relies on deep learning models like EfficientNet-B5 for medical image classification. These models often suffer from overfitting, reducing accuracy and lack explainable AI, making results hard to trust. Techniques such as data augmentation, regularization and explainable AI are applied to improve accuracy, reliability and interpretability for clinical use. Even though existing models like EfficientNet-B5 provide cancer prediction, they face issues such as overfitting and lack of explainable AI, reducing accuracy and trust. These limitations make it difficult for doctors to rely on automated predictions. The proposed system uses a hybrid model combining machine learning and deep learning techniques to improve accuracy. It integrates LIME (Local Interpretable Model-agnostic Explanations) for interpretable and transparent predictions. This approach enhances diagnostic reliability and supports effective clinical decision-making.