Paper Title
An Ai Framework to Aid the Diagnosis of Hepatic Encephalopathy, EEG as a Case Study

Abstract
Hepatic Encephalopathy (HE) is a disease with problematic complications that plague patients with liver cirrhosis. The clinical presentation of HE includes elevated blood ammonia levels and neuropsychiatric and cognitive dysfunction. Sometimes, the clinical presentation combined with motor activity and electroencephalography (EEG) abnormalities. The treatment efficacy of HE is suboptimal, mainly because the early changes are not noticeable. With the rapid development of big data and artificial intelligence (AI), experts and scholars have developed several AI models for assisting diagnosis. This paper proposes a framework for helping HE diagnosis. Suppose a potential patient can report their biomedical data more quickly. In that case, the doctor can diagnose as early as possible and better treat the patient. KaiDiag provides exclusive preprocesses, feature extractions, and analyzers for EEG signals. We design an adaptive scoring module to derive the final score so that a doctor can take the score as his reference during diagnosis. In addition, we create a series of hardware devices, including an aggregator and a portable EEG device, for daily monitoring and early detection. We conduct several experiments to verify the performance of KaiDiag, including the performance impacts of different data preprocessing methods and AI models. The experiment results show that KaiDiag (with Multi-Layer Perceptron) achieves 89.2% accuracy, 89.4% sensitivity, and 88.1% specificity. Compared with the existing research, KaiDiag can analyze the enormous variety of data and maintain a good performance in terms of accuracy, sensitivity, and specificity. Keywords - Hepatic Encephalopathy Diagnosis, Portable EEG Device, Artificial Intelligence