Development of an adaptive learning system about information security education

The aim of this study is development of an adaptive web based learning system about information security education by investigating relationship between users' actual behaviors and their intentions on risky and conservative behaviors about information security considering individual differences which are demographics, socio-demographics, internet usage routines, personality, risk perception and exposure to offense. According to literature and real life practices, some kind of technical solutions are applied in order to provide information security in which attempts still cannot stop information systems' to be compromised. As being the weakest link, human behavior and awareness on information security needs to be considered and studied. The literature also shows that end users show different information security behaviors because of individual differences which causes to be compromised with different threats. Developmental research method is used in this study. At the first stage of this study, demographic, socio-demographic, internet usage routines, personality, risk perception and exposure to offense attributes, risky and conservative behaviors are collected through self-reported "Information Security Awareness Scale" with the attendance of 619 volunteer end-user. After that, these volunteers were asked to show 11 of the risky and conservative behaviors which can be measured through a web site and defined with the expert opinions. Then 301 of this 619 users' actual behaviors related with their answers to scales are examined through a web application. This web application pretends for users to evaluate a web application's usability in order to prevent Hawthorne effect. In order to explain the differences between the intensions and actual behaviors of the participants about risky and conservative behaviors with independent variables of demographic, socio-demographic and personality test scores, simple lineer regression model was used. Different decision tree algorithms have been tried to predict which individual characteristics of the participants affect their intension and behavior, and how they may act on a particular information security topic. C5.0 decision tree algorithm, which gaved the best result on the data set, has been used. As a result of these analyzes, a personalized online learning content was presented to the user in this direction. Arrangements were made on the adaptable web based learning system prepared using the rapid prototyping method by presenting expert opinions on design and content. Adaptable web based learning system that has been implemented has contributed to the literature and real life applications in terms of modeling and modeling the real behavior regarding information security, reaching large audiences with its online structure, providing personalized educational content and preventing over-information and orientation problems.