透過您的圖書館登入
IP:3.133.142.145
  • 學位論文

應用文字探勘技術於消費者產品使用狀況之研究-以手機遊戲線上評論為例

Apply Text Mining Technology to Study Consumer Product Usage – An Example of Online Reviews of Mobile Games

指導教授 : 曾世賢

摘要


隨著手機遊戲的盛行,遊戲開發商以及玩家的數量皆大幅度成長,許多玩家皆會在論壇上討論經驗與建議,使用者的線上評論逐漸成為重要的訊息來源,然而,這些使用者的論述對於開發者而言,是最直接與即時的資訊,卻鮮少被整理並有效利用;過去許多企業或研究透過問卷來蒐集消費者的使用狀況,但消費者只能依問卷的框架與限制進行回答,而文字探勘則可以由消費者開放式的線上評論掌握使用需求,不受時間、空間與樣本數的影響,因此,本研究藉由文字探勘技術萃取消費者在PTT所分享的Pokémon Go線上評論,透過Python所撰寫的網路爬蟲程式進行資料蒐集,並且經過資料預處理、中文斷詞、去除停用詞、關鍵字詞雲、建立字典、LDA主題模型與解釋與評估等步驟進行分析,透過LDA主題模型分析字詞可能指涉的主題範疇,以及對應的關鍵字詞出現的頻率,再由趨勢圖繪出隨著時間所產生的變化,最後,根據由線上評論所分析的消費者產品使用狀況,提供開發商做為行銷策略上重要的依據。

並列摘要


With the prevalence of Mobile Games, numerous of game developers and players are growing significantly. Many players are willing to discuss experiences and suggestions at the forum and thus the online reviews from users become an important source of marketing and game development information. In fact, the discussions from the users are the most direct and real-time information for the developer, but it is rarely sorted out and effectively used. In the past, many enterprises or studies collect the use of consumers from questionnaire, but the answer is limited by the framework. The application of Text Mining can overcome the disadvantage of questionnaire by collecting information from open online review without the limitation of time, space and number of samples. Therefore, in this study, we apply Text Mining to analyze online review of mobile game Pokémon Go from PTT and utilize the web crawler software written in Python to collect the data. We also go through the process of preprocessing, tokenizing, deleting stop word, creating word cloud, dictionary and LDA Topic Model and further to explain and evaluate the analysis results. We categorize the terms of subject and calculate the corresponding appearing frequency of key words and then draw the trend graph to understand the popularity of these key words. Finally, according to the analysis of online reviews, we provide an important basis of marketing strategy for game developers.

並列關鍵字

Mobile Game Online Review Text Mining LDA

參考文獻


陳怡廷,2016,從公司與顧客的認知價值初探企業潛在無形資產:以自然語言處理的方法為應用,中山管理評論,二十四卷三期:565-608。
陳譽晏,2015,運用 R Shiny 建立文字探勘平臺之語意分析及輿情分析,數據分析,10卷6期:51-78。
張銀益、劉上嘉、陳松輝、陳慧玲、蔡幸蓁,2010,手機線上遊戲消費者使用意願影響因素研究,輔仁管理評論,17卷3期:55-84。
張凱迪,2014,應用潛藏面相評分分析於中文評論:使用局部潛藏狄利克雷分配方法,國立臺灣大學資訊管理學研究所碩士論文。
張宗浩,2015,網路遊戲之跨社群資料探勘,國立臺灣大學資訊管理學研究所碩士論文。

延伸閱讀