| Title |
Žalingo turinio atpažinimo metodas |
| Translation of Title |
Harmful content detection method. |
| Authors |
Pečiulis, Mantas |
| Full Text |
|
| Pages |
58 |
| Keywords [eng] |
harmful content detection ; classification of text ; harmful content ; supervised learning |
| Abstract [eng] |
Different technologies make it possible to access or share information. What most people do not realize is that online activities can be creative and educational and destructive. Harmful sites are becoming increasingly easy to access as they try to mask or mislead and circumvent blocking, all of which affect safe web browsing. Most web filtering systems are based on blacklists or keyword mechanisms that do not consider the content of the opened page, which does not always ensure that the content on the page can be identified as harmful or otherwise. In the master thesis, a prototype has been implemented that determines the category of the entered page. A dataset is built according to the methodology developed. A machine learning algorithm was selected and compared with other algorithms using the constructed dataset. The dataset consists of the content of the selected pages, which has been normalized and features extracted using a vectorization algorithm. The trained algorithm was tested and compared with other algorithms. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2022 |
