Thesis supervisor: Gábor Kiss
Location of studies (in Hungarian): Óbuda University - Donát Bánki Faculty of Mechanical and Safety Engeneering - 1081 Budapest, Népszínház str. 8. Abbreviation of location of studies: ÓEBGK
Description of the research topic:
With the widespread use of the internet and mobile phones, people all over the world are using online social media websites to share knowledge, form opinions, gather information and view personal social media profiles.
Social network usage has increased by 73% since 2012. Anyone can easily create online profiles on social media, which in the case of a fake profile, can create opportunities for identity fraud with potentially fatal consequences.
Social network analysis is a useful tool to provide information on social behaviour. Social network analysis can be used, among other things, to detect malicious behaviour and protect people's privacy, resulting in increased security of social networks.
Research purposes:
This research study discusses the theory and applications of existing soft computing techniques for data mining in web-based social media.
The proposed algorithms are ArtiNeuro Fuzzy Clustering Algorithm (AFCA) and Hybrid AntNeuro Fuzzy Clustering Based Association with Classification (AFCAC), for social media data mining to identify fake profile which pose security risk. A framework using soft computational methods such as ANN and Fuzzy Set Data Mining for social network data mining can overcome the major limitations of previous web mining methods by handling large-scale unstructured and dynamic weblog data from social media sites. The framework is based on a hybrid soft computing model that uses hybrid 3-layer approaches to test the datasets.
The scope of the research study is limited to:
1. proposal of some advanced algorithms based on soft computing approaches such as ANN, Fuzzy Logics and ACO for social network analysis.
Propose hybrid advanced algorithms based on several parameters as a combination of ANN, fuzzy logic and ACO methods, and design testing tools using R or Python and other simulation tools for efficient user classifications to identify problems such as users and groups with fake profiles, hate groups or terrorist groups covertly exchanging information.
Required language skills: English B2 Recommended language skills (in Hungarian): English C1 Further requirements: Programming knowledge, using of MATLAB and R as data mining tools
Number of students who can be accepted: 1
Deadline for application: 2022-08-31
2024. IV. 17. ODT ülés Az ODT következő ülésére 2024. június 14-én, pénteken 10.00 órakor kerül sor a Semmelweis Egyetem Szenátusi termében (Bp. Üllői út 26. I. emelet).
Login
