Health Care Automation in Compliance to Industry 4.0 Standards: A Case Study of Liver Disease Prediction

Venkata, Manjula Devarakonda; Lingamgunta, Sumalatha ; Murali, K


The industrial internet contributes to the standards of Industry 4.0, which involve handling large volumes of data using advanced soft-computing techniques. Machine Learning (ML) is an advanced soft-computing technique that plays a critical role in predicting and detecting serial chronic diseases, thereby automating the diagnosis. The process constitutes and uses several data mining algorithms and methods for efficient medical data analysis. Recent studies on several chronic diseases, liver disorders and diseases associated with the organ have been fatal. In this paper, the liver patient dataset from India is considered and investigated for developing a classification model. Liver disease is a dangerous, life-threatening disease often diagnosed false positive. Mild liver enlargement, improper or ambiguous functionality over a brief period, is prominent even in healthy people, which has become the main reason for ignoring the same at the early stage. It is essential to predict liver disease through the parameters and their values from the liver functionality test sensing the behavior of similar patients who were ignored in the initial stage. In this paper, the machine learning technique is demonstrated to predict liver disease using the liver function test data of the 580 patients as training data. The model has been developed with an accuracy of approximately 75%. The simulation-based experiment is based on the publicly available dataset and can be extended to any native set to predict the patients' health quickly. The Random Forest Algorithm is used to develop the model in Matlab, and the analysis is carried out using parameters like total bilirubin, alkaline phosphotase, alamine aminotransferase, total proteins, and A/G ratio.


Advanced soft-computing techniques, Indian dataset, Industrial internet, Machine learning, RF algorithm

Full Text: PDF (downloaded 330 times)


  • There are currently no refbacks.
This abstract viewed 788 times