Efficient Diagnosis of Liver Disease using Deep Learning Technique

Abstract

The diagnoses a patient receives can have significant repercussions for enhancing patient safety, investigation, and policymaking. Medical practitioners employ a variety of pathologic techniques to arrive at conclusions about their patients’ states in clinical information. The field of medical diagnosis has seen renewed efforts from clinicians in recent years. When Artificial Intelligence (AI) and Deep Learning (DL) are used in tandem with clinical data, they can greatly enhance the accuracy of disease diagnoses. The use of computers and internet has made it possible to acquire data and visualize previously inaccessible findings, such as addressing the issue of missing values in clinical research. Decision-making can be aided by problem-specific Deep Learning algorithms. In order to automatically identify illness specimens, effective predictive methods are essential. In this regard, this work employs techniques of deep learning to distinguish liver patients from normal persons. In this research, we make a prediction of liver illness using a Deep Learning model called BiLSTM. This model is able to keep track of long-term relationships in both the forward and the backward direction. The efficiency of the model’s predictions came out to be 93.00% overall. According to the findings, the implementation of a hybrid model seems to have enhanced the predictive accuracy.