Online Fake Drug Detection System in Heterogeneous Platforms Using Big Data Analysis

摘要

The widespread use of internet provides extensive heterogeneous platforms for drug sales. The internet has greatly facilitated the development of merchandise sales, meanwhile, many fake drug sellers that has been strongly restricted in the market by law enforcement agencies build their own platforms for sales via the internet. In order to against the fake drug website, reduce time and human resource consumption, it is necessary to screen and identify the drug information on the internet automatically. In this paper, we develop an automatic drug information screening and content analytical system which online distract the information and mine the hidden relationship and find the source of the sellers. Our major contributions lie in those aspects as follows: (1) We apply focused crawler technique to transform the unstructured data on the drug website into structured data, and stored the data in the local database. (2) An integrated fake drug identification method is proposed which consists of an image recognition module and information retrieval module. Based on this method, the fake drug website is not only identified one by one, but we also extract the hidden connections of multiple platforms. Experimental results demonstrate that our system can successfully identify a large number of fake drug website.