以社群探勘與偏最小平方法結構方程歸納影響消費者接受智慧手環之關鍵要素

Abstract

近年來，由於預防醫療、健康管理意識抬頭，加上物聯網和雲端運算技術進步，智慧穿戴裝置之市場隨之崛起，並且發展蓬勃。科技產品是否能廣普及，取決於消費者對該產品、技術的接受程度，因此，許多學者提出有關消費者行為之理論。近三十年來，技術接受模式已廣泛運用於分析、解釋消費者對新科技之接受行為，但現有研究受到樣本及取樣的限制無法針對母體分析，對消費者行為的探討，仍多所限制。此外，雖然運用文字探勘分析社群之研究日增，但關於智慧穿戴裝置的研究非常稀有，現有之研究，仍局限於歐美國家之少數個案，且少有以理論研證者。為解決此研究缺口，本研究擬定義一新穎的分析架構，使用網路爬蟲技術萃取社群網站有關智慧手環之貼文，並以現有情感詞典詞彙作為基礎，然後使用Jieba進行斷詞、提取特徵詞及意見詞，接著進行情感分析和主題建模，得出影響消費者接受智慧手環之關鍵因素。之後依據技術接受模式提出假設，再以偏最小平方法結構方程（PLS-SEM）驗證影響關係是否顯著。本研究以探勘 PTT 社群網站驗證分析架構之可行性，結果顯示除知覺易用性（Perceived Ease of Use, PEU）、知覺有用性（Perceived Usefulness, PU）會影響消費者接受、使用智慧手環外，情緒（Sentiment）於知覺易用性（Perceived Ease of Use, PEU）對知覺有用性（Perceived Usefulness, PU）之調節作用亦為顯著，負面情緒尤其如此。未來，本研究所定義，且驗證完善之分析架構，可用於探勘影響其他產品消費行為之因素，亦可做為產品行銷與設計次世代消費電子產品之依據。In recent years, due to the rising awareness of preventive medical treatment and health management, coupled with the advancement of the Internet of Things and cloud computing technology, the market for smart wearable devices has risen and developed rapidly. The vigorous development of the smart wearable device market depends on consumers' acceptance of the product and technology, and many scholars have put forward theories about technology adoption behavior. During the past decades, the technology acceptance model has been widely used to analyze and explain consumers' acceptance behavior of new technologies. However, existing research is limited by samples and sampling, and therefore cannot represent the population and fully reflect consumer behavior. Also, although the literature on the use of text mining and analysis of social platforms is increasing due to the flourishing of social media, research on the topic of smart wearable devices and smart bracelets is very rare, and the existing research is still limited to a minority. Among them, most of the areas analyzed by text mining are a few cases in European and American countries, and there are few theoretical kinds of research. Therefore, this research aims to propose a novel analysis framework to cross the research gap. The web crawler technology will be adopted to extract posts on social media websites related to smart bracelets based on the vocabularies of existing sentiment dictionaries. Then use the Jieba to segment words, extract feature words and opinion words, then perform sentiment analysis and topic modeling, and get the key factors that affect consumers' acceptance of smart bracelets. After that, a theoretic framework based on the technology acceptance model will be verified by using, the Partial Least Squares-Structural Equation Model (PLS-SEM). This study is to explore the feasibility of the analysis framework by exploring the PTT social website. The results demonstrate that Perceived Ease of Use (PEU) and Perceived Usefulness (PU) affect consumers’ acceptance and use of smart bracelets. Further, sentiment mediates the correlation relationship between the PEU and PU; the effects of negative sentiments are especially significant. In the future, based on the results of empirical studies, the analytical framework proposed and verified in this research can be used to mine factoring influencing consumers’ acceptance of novel technologies, as well as the basis for marketing and design of next-generation consumer electronics.