Prediction of Disease Using Machine Learning and Deep Learning

摘要

Big data plays an important role in multiple industries, but it has a greater impact and is important to the rapidly growing healthcare industry. Big data plays a crucial role by providing a robust system wherein a better result in disease detection can be achieved. Initially, the predictions are made on the data available, but the lack of incomplete data leads to a reduction in the quality of accuracy. In addition to incomplete data the different characteristics of specific regional diseases, which vary with respect to the region can weaken the prediction more. In this paper, we compare machine learning algorithms and Convolutional neural networks (CNN) to show the increase in efficiency of predictions related to chronic disease outbreaks in communities. We implement and check the efficiency of the proposed models over real-life medical data. To avoid complications that arise due to missing data we use a latent factor model to reconstruct incomplete information. We perform predictions on a regional chronic disease lung infection. We propose a convolutional neural network-based multimodal disease prediction that performs predictions based on both structured as well as unstructured data.