| Presentation | 2016-03-03 Malicious-Spam-Mail Detection System with Autonomous Learning Ability Shogo Osaka, Jun Kitazono, Seiichi Ozawa, Tao Ban, Junji Nakazato, Jumpei Shimamura, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Recently, damages caused by spam mails that guide receivers to maliciousweb pages become more and more serious. In this study, we propose anautonomous learning system to detect such malicious spam mails. In theproposed system, the main body of a mail is transformed into a featurevector based on the tf-idf weight, and the feature vector is classifiedby machine learning classifiers. In addition, to keep up with the trendof mail-body contents, we adopt one-class SVM to detect outliers, andcontinuously update the classifiers. In the experiments, we usedouble-bounce mails collected by NICT for 120 days. We demonstrate thatthe proposed system can detect malicious spam mails with a highF-measure of $97.25%$ on average. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | machine learning / support vector machine / spam mail / autonomous learning |
| Paper # | ICSS2015-50 |
| Date of Issue | 2016-02-25 (ICSS) |
| Conference Information | |
| Committee | ICSS / IPSJ-SPT |
|---|---|
| Conference Date | 2016/3/3(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Academic Center for Computing and Media Studies, Kyoto University |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Information and Communication System Security, etc. |
| Chair | Yutaka Miyake(KDDI R&D Labs.) |
| Vice Chair | Takashi Nishide(Univ. of Tsukuba) / Yoshiaki Shiraishi(Kobe Univ.) |
| Secretary | Takashi Nishide(Mitsubishi Electric) / Yoshiaki Shiraishi(NII) |
| Assistant | Katsunari Yoshioka(Yokohama National Univ.) / Kazunori Kamiya(NTT) |
| Paper Information | |
| Registration To | Technical Committee on Information and Communication System Security / Special Interest Group on Security Psychology and Trust |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Malicious-Spam-Mail Detection System with Autonomous Learning Ability |
| Sub Title (in English) | |
| Keyword(1) | machine learning |
| Keyword(2) | support vector machine |
| Keyword(3) | spam mail |
| Keyword(4) | autonomous learning |
| 1st Author's Name | Shogo Osaka |
| 1st Author's Affiliation | Kobe University(Kobe Univ.) |
| 2nd Author's Name | Jun Kitazono |
| 2nd Author's Affiliation | Kobe University(Kobe Univ.) |
| 3rd Author's Name | Seiichi Ozawa |
| 3rd Author's Affiliation | Kobe University(Kobe Univ.) |
| 4th Author's Name | Tao Ban |
| 4th Author's Affiliation | National Institute of Information and Communications Technology(NICT) |
| 5th Author's Name | Junji Nakazato |
| 5th Author's Affiliation | National Institute of Information and Communications Technology(NICT) |
| 6th Author's Name | Jumpei Shimamura |
| 6th Author's Affiliation | clwit Inc.(clwit) |
| Date | 2016-03-03 |
| Paper # | ICSS2015-50 |
| Volume (vol) | vol.115 |
| Number (no) | ICSS-488 |
| Page | pp.pp.19-24(ICSS), |
| #Pages | 6 |
| Date of Issue | 2016-02-25 (ICSS) |