The technology of new events extraction on a defined topic from Twitter social network

• Видавець: Інститут проблем реєстрації інформації НАН України
• Дата: 2017
• Автор: Pryshchepa, S. V.
• Формат: Стаття
• Мова: Ukrainian
• Опубліковано: Інститут проблем реєстрації інформації НАН України 2017
• Теми: event extraction; event extraction method; block diagram of the extraction of new events; экстрагирование событий; метод экстрагирования событий; блок-схема экстрагирования новых событий; екстрагування подій; метод екстрагування подій; блок-схема екстрагування нових подій
• Онлайн доступ: http://drsp.ipri.kiev.ua/article/view/126542

It is presented an overview of our approach to information extraction, which is a field of natural language processing that deals with new events from Twitter social network. There are formed methods and approaches into technology for detection events and determine their novelty. On the main base of our approach we offered to use technology of extracting events based on special dictionaries with event triggers, event classification event or not event by the method of classification of naive Bayes and extraction of special concepts and analysis their duplication with past events in some database of events. The approaches and algorithms of the technology based on these methods are presented. There is a steady interest in effective event extraction, due to its conceptual simplicity from one hand, and its potential utility on the other. The task of processing natural language has many problems, some of which cannot be solved within the framework of this technology, but the solution of which could improve the effectiveness of detecting new events.