Medical Data Crawling Algorithm Based on PageRank

Abstract

In order to establish a database of characteristics related to physical conditions and then build a remote health intelligence-assisted diagnosis model based on the deep learning training mechanism, it is necessary to perform deep mining of medical data. In addition to the structured medical data stored in medical institutions, there are a large number of doctors and patients on the Internet about the interaction of the disease, and these are important sources of medical data. PageRank algorithm is an efficient link-based Web page sorting algorithm. This algorithm considers the Internet as a whole and uses links between pages as an important indicator. Through the relationship between Web pages pointing to each other, the algorithm calculates the importance of the page. However, it also has some problems, such as the heavy emphasis on old Web pages, the theme drift, and so on. In this paper, based on the characteristics of medical data crawling, an improved PageRank algorithm based on PageRank is designed. The algorithm introduces time factors and potential correlation factors, and solves the problems of the original algorithm. Experiments show that the algorithm presented in this paper has good performance, both in terms of operating speed and accuracy.