Using PTT BBS Research of Job-Related Issues by Text Mining & Sentiment Analysis

• Main Authors: GONG,WEN-SIN, 龔雯馨
• Other Authors: LIN,YU-SEN
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/72n73q

碩士 === 國立高雄師範大學 === 人力與知識管理研究所 === 106 === Bulletin Board System (BBS) is widely used to discuss current events and information exchange in the internet era. The purpose of this thesis is to explore the keywords and sentiment analysis of work-related issues in PTT BBS which is the most popular BBS in Taiwan. The data is mainly based on the Life_Job board of PTT . The data analyzed was from April 1, 2016 to March 31, 2018, which divided into eight quarters. A standardized process of a 4-step approach was used: text/data pre-processing, processing, analysis, and visualization of information. R language was used for text extraction, text mining and sentiment analysis. The results show that: (a) there are 263 keyword words in 10 boards. Finance board has the least keywords (n=15). Part-Time board has the most words (n=34); and (b) For the comprehensive analysis of sentiment analysis, most boards show positive sentiment relations except PublicServan board is negative and the Medstudent board is neutral. Based on our findings, job seekers can appreciate the relevant issues in particular job or positions from the keyword analysis results. Job seekers can prepare well for interview, and therefore increase the possibility of acceptance. From the employer’s point of views, the results of keywords can be combined with the four steps of the SEO strategy. The results of sentiment analysis can be applied in conjunction with the seasonal variations. Finally, keyword and sentiment analysis can be integrated into an online management platform for human resource personnel in recruitment operations, or for company/enterprise to search keywords in certain issues, and even deal with related businesses of corporate image management.