Research on the Safety Accidents Prediction for Smart Laboratory Based on Statistical Analysis

摘要

With the development of information technology, the Smart Laboratory is an core research topic of the modern university, which is responsible for the laboratory safety. To this problem, we focus on the human management since many safety accidents are often occurred by improper operations. This paper proposes the concept of security credit, which consists of user characteristic, behavior preference, learning ability, test score, course attendance and accident history. It is an evaluation system to personal security knowledge and skills which can reflect the potential risks of laboratory. Thus, the users will be first clustered based on their safety credits. After that, the risk group and the potential group are defined by K-means. The relationship between the learning content and the type of safety accidents is defined. Then, the security accidents are predicted according to the situation of the user's learning contents and exam results. Finally, the experiments are carried out based on the datasets supported by Ankai WebSite (SHUAnKai) of Shanghai University, by which the results are demonstrated to the feasibility to the management of Smart Laboratory.