Author(s)

Zhou Bei

Corresponding Author

Zhou Bei

Abstract

The traditional information extraction is mainly for the news, the extraction operation is relatively simple, and the characteristic of micro blog post - short refining needs to select a new and trial way to conduct the event extraction. In the course of these questions, this is a new algorithm that can think about the key word and time, set up the LDA model of the micro blog event, which is the first thing to set up the terms of the event, and calculate the semantics of the key words by the method of the DTW algorithm, the corresponding similarity matrix, and then use the method of co-training to do the same thing as the final eigenvector and finish the selection of the event. It is shown that that algorithm proposed in this paper has high accuracy and high efficiency compared with the previous algorithm.

Keywords

Micro blog, text data mining, eigenvectors, similarity matrix.