Countrywide anomalous incident detection from aggregated mobile phone data
| dc.contributor | Graduate Program in Computer Engineering. | |
| dc.contributor.advisor | Salah, Albert Ali. | |
| dc.contributor.advisor | İncel, Özlem Durmaz. | |
| dc.contributor.author | Gündoğdu, Didem. | |
| dc.date.accessioned | 2023-03-16T10:01:55Z | |
| dc.date.available | 2023-03-16T10:01:55Z | |
| dc.date.issued | 2014. | |
| dc.description.abstract | Mobile phones are extensively used both in developed and developing countries. Richer countries tend to use it as a computer, whereas in developing countries it is a kind of replacement of inadequacy of infrastructure, or it is used to cover more crucial needs. This work is an explorative analysis of the nationwide Call Detail Record (CDR) in Cote d'Ivoire. Our aim is to detect anomalous incidents and to explore the possibility of early detection of severe incidents from mobile phone data. Beside this, this work is a kind of roadmap who would like to work on CDR data, from obtaining open data set to visualization tool selection. Data is collected and anonimized in Cote d'Ivoire by Orange Telecom, from real call data. It has been published as a part of a scienti c challenge, Data for Development (D4D). We explored irregularity of phone usage by probabilistic Markov modulated Poisson process (MMPP) method. During the data collection period (from December 2011 to April 2012) Cote d'Ivoire was su ering from civil war, which caused hundreds of deaths and many injuries. Validation of the experiments has been done through United Nations Security Council's report covering these dates. | |
| dc.format.extent | 30 cm. | |
| dc.format.pages | xvii, 107 leaves ; | |
| dc.identifier.other | CMPE 2014 G86 | |
| dc.identifier.uri | https://digitalarchive.library.bogazici.edu.tr/handle/123456789/12276 | |
| dc.publisher | Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2014. | |
| dc.subject.lcsh | Mobile communication systems. | |
| dc.title | Countrywide anomalous incident detection from aggregated mobile phone data |