Abstract
In this study, we analysis Yahoo! stock BBS posting by Text Mining using specialized dictionary and investigate the relation between posting and stock price. Morphological analysis using general dictionary has two problems. First, general dictionary extract a noise which seems to have no correlation with stock price like mental abuse. Second, technical terms, such as PER and M&A, can't be extracted.
So we solve these problems by using an economic specialized dictionary and a financial specialized dictionary in the case of a morphological analysis. We compare the correlation of principal component analysis factor and stock price for each dictionary, and show the usefulness of specialized dictionary.
