Mining actions from reports on flood
| Authors | |
|---|---|
| Year of publication | 2006 |
| Type | Article in Proceedings |
| Conference | Proceedings of IS-LTC'06 |
| MU Faculty or unit | |
| Citation | |
| Field | Use of computers, robotics and its application |
| Keywords | text filtering; information extraction; term extraction |
| Description | This paper focuses on mining in short reports that describe a situation in a given area and actions performed as reaction to that situation. Such texts are frequent in crisis management in situations like earthquake, fire or flood. For further analysis it is necessary to filter the relevant pieces of text. We found that common machine learning algo rithms fail for filtering such sentences. We describe a novel method based on inductive logic programming which yields in high precision and recall. This method has been successfully used for analysis of reports on flood in Central Europe in 2002. We also discuss different domain knowledge and also various natural language processing tools that we used for preprocessing the documents. |
| Related projects: |