以文字探勘技術分析一般員工於網絡社交平台抱怨內容之研究

隨著大數據和人工智慧的推行，企業各部門積極推動數位變革，設法利用龐大的資料來改善本身部門的運作，過去人力資源談到數據管理，不外乎離職率、員工滿意度及KPI等，然而人資相關的相關研究，以往都是藉由問卷方式進行抽樣調查，分析的結果及範圍容易因為採用各個針對不同研究所設計出的問卷構面而受到限制。故本研究透過公開社群媒體所分享的言論，蒐集員工在不受任何壓力與牽制的情況下所產生的抱怨言論，並以文字探勘技術加以分析瞭解其中關鍵因素，藉由分析結果及數據之間的關係，探討勞資雙方經營觀念的認知落差，並瞭解員工內心需求，進而提升勞資關係，增加員工留職意願並提升員工投入程度。研究結論有助於進一步瞭解，員工抱怨次數及關鍵詞彙是否會隨著法規修正及節慶時間而有所變動，以及管理者對於員工人才管理以及認知落差的處理。

關鍵字

社群媒體；線上評論；員工抱怨；文字探勘；隱含狄利克雷分模型

並列摘要

With the tide of big data and artificial intelligence, many companies and their functional departments start to take advantage of the digital change and try to apply big data analysis to improve operations. In the past, the research field of human resources always focus on data management, including turnover rate, employee satisfaction, KPI, etc. Also, most of the relevant researches are sampled by questionnaire and the results and scope of the researches may be limited. In this thesis, we collect complaints or whines from employees without any pressure or containment which are posted on the social media and apply text mining to analyze them. Using the analysis results to discuss the cognitive gap between employers and employees, the psychological needs of employees, enhance labor relations, increase the employees intention to stay, and job involvement. The conclusion of this research help us to further develop the variations of employee’s complains and as well their keywords between the generation of regulation or the upcoming festivals. Also, to dig dipper on how HR manager handles the different requirement between board and employees.