An Enhanced Hybrid Model for Liver Disease Detection Utilizing Deep Learning and Machine Learning

The liver is our largest internal organ and controls all bodily metabolic processes, including transforming dietary nutrients into compounds that may be used by the body, storing those substances, and then providing them to the cells as needed. Ailments of the liver are among the most devastating disorders in many countries. The prevalence of liver disease has progressively risen due to excessive alcoholism, exposure to dangerous gases, eating foods laced with poison, and drug use. The majority of people worldwide experience mild to severe liver disorders as a result of bad lifestyle choices. Liver diseases continue to post a significant global health challenge, and the need for improved detection methods is crucial. Here, we propose a hybrid model to predict liver maladies utilizing machine learning & deep learning modes. Researchers study datasets of patients with liver disorders in order to help in the creation of classification models for forecasting liver illness. Making use of such datasets can ease the burden on medical practitioners and speed up the diagnosing process. An ensemble stacking model is used in the first phase with ML algorithms such as Naïve Bayes, Decision Tree, KNN & SVM. A logistic regression model functions as meta learner for predicting liver diseases utilizing clinical data. In the second phase, ensemble stacking model is used with VGG 16, ResNet and Inception V3 as the base learners and logistic regression as meta learner for the analysis of image dataset. Combining multiple models, especially using ensemble methods, often enhances predictive performance by leveraging the strengths of individual models.