Yuuki Hotta, Haruki Ogata, and Ryuya Uda
Recommendation, Semantic Web, Lifelog, Ontology
Recently, popularity of online shopping has raised crime rate on the web such as identity theft, phishing, etc. Impersonation fraud by malicious people on the web finally leads to stealing money, e.g. illegal usage of someone's credit card or web money. One of the countermeasures against impersonation on the web is identification with lifelog which is collected in daily behavior. However, ideas and methods on existing researches have not strong relation between personal favorites and lifelog ontology models. For example, the ontology models are generated with users' positioning information by GPS. Therefore, in this paper, we propose new lifelog ontology models strongly related with personal favorites. In our proposal, all ontology models consist of favorite vectors which are independent each other, and the vectors are generated by analysis of lifelogs on shopping sites. Our ontology models can eliminate impersonated users from online shopping site, when one of the users has different favorite from a correct user's favorite.
Important Links:
Go Back
