SISTEM MULTI AGEN UNTUK EKSTRAKSI WAKTU DAN TEMPAT EVENT DARI DATA TWITTER UNTUK WEB KALENDER

Fajri Profesio Putra, Dr. Azhari S.N., M.T.

2014 | Tesis | S2 Ilmu Komputer

-

Twitter users every day to share event information but no other user at any time to update the information. A system is required to present information of time and place as well as the category of an event. Computational approaches to doing data processing is multiagen intelligent. twiter characteristics of the agent that is able to work independently and mobile on their environment, are expected to play a role in processing event information from twitter. The Data is processed to obtain information on the time and place of activity and determine the category of the event.Naive bayes classification classificatiom a method to determine the category of the document or object, consist of training and testing process. The type of the event consisted of three categories of entertainment, sports, and education. To design agent uses methodology prometheus. Agents developed consisting of collection agent for collect and extraction data, preprocessing agent to clean data and classifications agent for determining the category of event. Extraction date and location based on writing of date and location format. The output data shown in a calendar containing information of categories, time and place of events.

Kata Kunci : twitter, event, data extraction, naive bayesian classifier, multiagent.