Project funded by:
Project results
Characteristics of methods and a prototype being developed
- Sources of information of the SMC system are public content (e.g. articles and comments on web pages, posts on discussions forums, ads) from the shallow and the deep Web. Project includes processing of natural language unstructured data.
- Data being processing comes from heterogeneous, structured and unstructured sources.
- Information sources are constantly monitored to detect irregularities that may be effect of occurred threat.
- Information about a certain class of a threat is stored in threat metaprofiles, that enable filtering of information from sources being monitored.
- Threat profiles are built and evolve according to rules defined by experts.
- Domain experts contribution is necessary only while defining rules, not during the process of monitoring for certain threat class.
Unique features
- Information extraction from chosen Internet sources, like social network sites or on-line auction services.
- Integrating data from diverse sources, from public web sites as well as from internal databases.
- Automatic detection of content-related cyber threats in monitored sources.
- A threat is identified based on rules defined by experts. Thus proposed solution can be used to detect different threat types.