A Gen AI and Deep Learning Based Approach for Liver Disease Detection

Early detection of liver disease is important for proper treatment, as it is a public health issue on the rise worldwide. With the Liver Patient Dataset (LPD), this paper suggests a deep learning method, in particular, using a feed- forward artificial neural network (ANN), and applying feature engineering procedures such as encoding, scaling, and model performance improvement. The dataset was split into training and testing samples, and hyperparameter adjustment was done in order to ensure maximum accuracy. The model was tested against some of the important key performance indicators (KPI), such as accuracy, precision, recall, and F1 score. The findings confirmed that the model works competitively in comparison to other models. Deep learning models like the feedforward ANN used here have outperformed conventional machine learning models in classifying liver diseases and even offer a larger benefit in employing web-based applications. This was done through the development of a Flask application to facilitate quick evaluation of the liver in real-world healthcare environments. This aspect highlights the high suitability of deep learning towards medical diagnosis and the scope for the model’s further extension with the addition of more clinical parameters and the use of a higher sample size for improved generalization. In all, the model is expected to assist the healthcare sector with enhanced timely detection and required intervention of liver disease.