- 學位論文
網路電話問卷調查法之使用意願行為模式分析－使用兩階段集群法和C5.0決策樹
Behavior Intention to Use Internet Phone Survey－ Using Two-stage Clustering Approach and C5.0 Decision Tree
摘要
網路電話是網路應用上創新的科技。科技接受度模型（TAM）可以探知使用者使用新科技的接受度。其中重要的因素有感覺有用（PU）和感覺易用（PEOU），這兩項因素可以決定使用者對於新科技是否有強烈的使用意圖，本研究再加入感覺焦慮和電腦自我效能這兩項外部變數來探討網路電話問卷調查法的接受度。創新擴散理論（DOI）有五項重要的「創新特性」包含相對利益（Relative Advantage）、相容性（Compatibility）、複雜性（Complexity）、可試用性（Trialability）與可觀察性（Observability）。本研究使用科技接受度模型（TAM）和創新擴散理論（DOI）做為研究的兩大基礎理論。 本研究有效樣本共330份問卷。問卷題目主要是探知使用者對於網路電話問卷調查法的使用意圖；分析問卷資料的技術包含多元迴歸分析和資料探勘技術，根據研究問題一使用多元迴歸分析來分析網路電話問卷使用意圖，而研究問題二採用資料探勘的方法此方法包含兩階段集群法和c5.0 tree，C5.0決策樹可以來建立使用者行為模式的決策樹。 實驗結果顯示，在第一部分，電腦自我效能與「感覺易用」有顯著的影響，感覺易用與「感覺有用」有顯著影響，感覺有用與「感覺喜歡」有顯著影響，感覺有用與「意圖使用」有顯著影響，感覺喜歡與「意圖使用」。 第二部份，分群結果群數為兩群，分別命名為低意願與高意願，再利用分群後的結果建立C5.0 tree，低意願的使用者年紀小於21歲男性且為醫藥、管理、文學院者，使用電腦時間在7年以上者；高意願的使用者年紀介於21?25歲的女性且為資訊、工程、設計學院，使用電腦時間在7年以下者，未來在使用網路電話做問卷調查時，可以透過年齡、科系、性別、使用電腦時間即可瞭解使用者使用網路電話問卷調查法的意願。 關鍵字：科技接受度模式（TAM）、創新擴散理論（DOI）、網路電話問卷調查法（Internet Phone Survey）、兩階段集群法（Two-stage Clustering Approach）、C5.0 決策樹演算法
並列摘要
Internet Phone is an innovative internet application. Technology Acceptance Model (TAM) has been employed to predict user acceptance on new computer technology, and it stresses the importance of perceived usefulness (PU) and perceived ease-of-use (PEOU) as key determinants. Additionally, this study increases the perceived anxiety and computer self-efficacy as external variables to explore the user acceptance for the Internet Phone survey. Diffusion of Innovation Theory (DOI) has five important innovation characteristics, including: Relative Advantage, Compatibility, Complexity, Trialability and Observability. Then, TAM and DOI are the two foundation theories applied in this study. The total samples of 330 valid questionnaires were collected from the survey. The questionnaires are used to exploring users’ intentions of the Internet Phone survey. The aim of this study is to apply the multiple-regression-analysis and data mining techniques to analyze the questionnaire data. The first research issue in this study is to employ the multiple-regression-analysis to examine users’ intention of the Internet Phone survey. The second research issue is to adopt the data mining analysis methods that include both two-stage clustering approach and C5.0 decision tree to find different decision models for users’ intention. The first part results of this study show that the computer self-efficacy had been significantly correlated with easy-to-use; and easy-to-use had been significantly correlated with perceived usefulness. Perceived usefulness had been significantly correlated with perceived liking. Perceived liking had been significantly correlated with intention to use Internet Phone. Perceived usefulness had been significantly correlated with intention to use Internet Phone. The second part results illustrate that the questionnaire data can be cluster into two groups: low-intention and high-intention. The samples within low-intention group are male, less than 21 years old, major in medicine, management, literature and computer experience above than 7 years. The high-intention samples are female, belong to 21~25 years old, major in information science, engineering, design and computer experience below than 7 years. According to the results of this study, the researches can only stand on users’ age, department, sex, and computer experience to find different decision models for users’ intention of Internet Phone survey in the future. Keywords: TAM, DOI, Internet phone survey, two-stage clustering approach, C5.0 tree