%0 Conference Paper %1 hardy2006 %A Hardy, Hilda %A Kanchakouskaya, Vika %A Strzalkowski, Tomek %B Proceedings of the 2006 AAAI Workshop on EventExtractionand Synthesis %D 2006 %K dataset_own event_extraction sentence_classification %P 36--41 %T Automatic Event Classification Using Surface Text Features %U http://www.aaai.org/Papers/Workshops/2006/WS-06-07/WS06-07-008.pdf %X Extracting events from documents quickly and accurately is an important goal for many tasks that require language understanding, such as question answering. We present a data-driven method for discovering events and their attributes in a corpus. We further demonstrate that a carefully chosen set of textual features, when used to train some well-known learning algorithms, can approach or exceed the accuracy of hand-crafted patterns for event classification, requiring far less time and expertise. The features can be gathered using lightweight text processing tools. Overall classification accuracy reaches 59.76% for a set of 11 event types.

@inproceedings{hardy2006, abstract = {Extracting events from documents quickly and accurately is an important goal for many tasks that require language understanding, such as question answering. We present a data-driven method for discovering events and their attributes in a corpus. We further demonstrate that a carefully chosen set of textual features, when used to train some well-known learning algorithms, can approach or exceed the accuracy of hand-crafted patterns for event classification, requiring far less time and expertise. The features can be gathered using lightweight text processing tools. Overall classification accuracy reaches 59.76% for a set of 11 event types.}, added-at = {2009-10-22T15:20:02.000+0200}, author = {Hardy, Hilda and Kanchakouskaya, Vika and Strzalkowski, Tomek}, biburl = {https://www.bibsonomy.org/bibtex/213e7643eca5a2fb1ab1cc5308ea901d1/jnothman}, booktitle = {Proceedings of the 2006 AAAI Workshop on EventExtractionand Synthesis}, description = {Interested in identifying events for QA, based on verbs, nouns, adjectives. Compare manual patterns to automatic classification. Own annotated dataset. Consider only sentences containing multiple NEs (NER with BBN Identifinder). Weka for classifying sentences to NONE or 11 event types (3996 total instances in corpus). ML techniques not suited to large feature spaces. Features: top k words per {verb, noun, adj} POS; POS tags, sent length, NEs, chunk patterns. Chunk patterns were not useful. Performance up to 60% accuracy.}, interhash = {067cb43924be046d0b65eb9861273e35}, intrahash = {13e7643eca5a2fb1ab1cc5308ea901d1}, keywords = {dataset_own event_extraction sentence_classification}, pages = {36--41}, timestamp = {2009-10-22T15:20:03.000+0200}, title = {Automatic Event Classification Using Surface Text Features}, url = {http://www.aaai.org/Papers/Workshops/2006/WS-06-07/WS06-07-008.pdf}, year = 2006 }