Predicting Liver Cancer Patients on Covid-19 Pandemic Using Deep Learning (DL) Method

Abstract

Liver cancer (LC) is one of the most rapidly growing types of cancer worldwide. Lower death rates result from the early identification of liver cancer. Due to COVID-19's high infectiousness, a large influx of patients can enter hospitals at once for both detection and treatment, which posed a significant challenge to the nation's public healthcare institutions. The COVID-19 epidemic has adversely impacted cancer patients. The COVID19 pandemic's negative effects on cancer patients who contract the virus, its effects on the accessibility of cancer treatment, and the significant interruption to cancer research are only a few examples of this effect. The population of cancer patients is diverse, and current research has now identified characteristics that enable risk categorization of cancer patients in order to improve therapy. The severity of the symptoms based on initial assessment frequently determines the priority of the treatment. Clinically, it has been difficult to predict outcomes for patients with liver cancer having COVID-19. The severity of the disease has been retrospectively linked to a large number of clinical characteristics, but it is still unclear how well these variables predict disease progression or how many variables interact to enhance risk. This research focuses on sequential analysis to extract features using activation function in the Long Short Term Memory (LSTM) as Deep Learning (DL) techniques. The COVID-19 effect on Liver Cancer Prediction dataset is used to predict LC during COVID-19 impact. Additionally, the findings of this experiment demonstrated that DL models performed better in terms of accuracy with leaky relu activation function classifier than other activation function classifiers.