텍스트 마이닝을 활용한 시중은행 영업점 직원의 만족도 및 재무적 가치 간의 시차상관분석

초록

In the present study, the authors, by drawing on the methodology of text mining and time lagged cross correlation analysis, analyzed voice of employees who work at the branches of the four major domestic commercial banks (Kookmin Bank, Shinhan Bank, Woori Bank, and Hana Bank) and the correlation with corporate value (stock price). To the end, we collected the corporate review data (i.e., voice of employees including thought and feeling) from Job Planet(https://www.jobplanet.co.kr) by using web scraping method and the corporate value data (market capitalization) from the Korea Stock Exchange with Python. The Job Planet’s corporate reviews including unstructured texts were processed in natural language to analyze high-frequency words and slang. Employees’ thought and feeling were analyzed with the topic modeling method and the emotional analysis, respectively, to quantify qualitative data. The results showed that there was a statistically significant time-lagged cross correlation between the voice of employees and stock prices. Furthermore, the results raised a question about Eugene Fama's (1970) efficient market hypothesis, which is the dominant theory that the market is as efficient as it immediately reflects all information , by suggesting a possibility that there is a time delay between reflection of all information in the stock market and stock price.

키워드