User Authentication in Training Systems

Abstract

The problem of verifying the identity of a student who works in an online educational system is discussed in the article. This is necessary to prohibit the substitution of one user for another upon receipt of information or in the process of performing certification work. Verification is carried out on the basis of the student's keyboard handwriting, formed according to the cluster principle. Symbols with similar characteristics are clustered. For each cluster, statistical characteristics are calculated. The authentication method in the process is very important. It is proposed to consider the methods of character-by-character data control or verification after entering the entire answer in order to ensure the correct decision. For this, approaches were used to analyze each symbol individually or average values for various combinations of symbols. During the study, a number of experiments were carried out and the data obtained are presented in the work. Studies have shown efficiency of the proposed user authentication method. Monitoring cluster average values was considered impractical, since large confidence intervals almost always include the average. The most successful is checking after entering the whole word. Next, you should remove the "glaring errors" or calculate the average values for each character, if there are several of them in the answer. In the latter case, the confidence interval must be constructed, based on the probability of occurrence, equal to 0.95.