Research On Individualized Campus Calendar System Based On Information Extraction

With the dramatic progress of Internet, the scale of information is becoming more and more diversified and complicated. This brings many difficulties for persons to search for useful information. How to extract the needed information from the daily emerging data has become the focus research of natural language processing. Information Extraction which is researched in this paper was generated for this reason, it makes disordered and irregular information become structured information to save and plays an important role in promoting the development of the information technology.The feature of this paper is that to research the information extraction technology centre on the events and time, and design and realize an individualized campus calendar system. The main innovation and research are as follows:Firstly, design and actualize a method that combines regular construction and statistical models together to extract Chinese entity relationship. This method extracts the exact result by the use of regular expression. And then takes use of CRF and ME to give the additional results, this method can improve the accurate rate. It also makes good result In SlotFilling task for TAC-KBP Track.Secondly, this paper propose、 design and actualize the individualized campus calendar system. This system extracts the information of event and collates the time expression in the event to provide clues for people fully understand the event.This system extracts the time expressing by the model of the rules and makes it normally. On this basis, a method is put forward to identify the start and end time expression of the event based on Word Activation Force. The start and finish time of the event provides more information for understanding the development of the events. This system is having a successful application on the system of COSE and on-line.Thirdly, this paper propose a method to extend the opinion word vocabulary based on WAF and a method for sentiment orientation judgment based on machine learning. This method makes a good result in the task one of2011COAE track. This algorithm gives the campus calendar system an adding sentiment orientation judgment function. This function can be used for the campus public opinion monitoring.