| Title |
Kenksmingų interneto tinklalapių identifikavimo modelis / |
| Translation of Title |
Malicious web page identification model. |
| Authors |
Dapševičius, Andrius |
| Full Text |
|
| Pages |
60 |
| Keywords [eng] |
malicious web pages ; classification ; PageRank ; logistic regression model ; support vector classifier |
| Abstract [eng] |
Web pages can be classified into two categories. One category is for websites that profit from some service provided for the customer, and the other category is for web pages that profit by harming the user. Web pages that fall into this category are malicious web pages. This paper describes a tool and methodology for malicious web page identification. Identification is based upon addresses of the web pages and their position in the web graph, which allows to consider a tendency of malicious websites to contain links to other malicious websites. Three machine learning models were being trained – a decision tree, support vector machine, logistic regression model. The best model reached 97% accuracy. Model was used to create a software tool that is designed to be used in other applications and application systems. Also, this paper raises some concerns regarding use machine learning models for malicious website identification because of problems that were noticed during creation of the model. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2017 |
