無線感測網路（ZIGBEE）文獻發展的趨勢、類別與熱點之探討──使用文獻內容書目對與共現字以發掘研究問題的機會

EXPLORING AMONG TRENDS, CLUSTERS AND HOT TOPICS OF WIRELESS SENSOR NETWORKS (ZIGBEE)-USING BIBLIOGRAPHIC COUPLING AND CO-WORD FOR RESEARCH QUESTIONS

無論是企業或學術界，不斷的創新是營運與學術生涯保鮮的根本。智慧街燈的發展是智慧城市中多議題中之一個有趣新興議題，在企業界與學術界永續發展的目標中，皆已投入甚多基礎與應用研究，研究成果首先皆已撰述論文或專利文件，投稿期刊或各國智慧財產專利局，歷經多位同儕盲目審查，經公開發表或取得核准且具排他權之著作權或專利。這些成果揭示與時俱進的新知識，為促進學術界與業界技術研發的快速擴散，皆已收錄在電子資料庫中，在學界，如WOS，Scopus，Elsevier；在業界如各國專利資料庫，如USPTO，EDO等。這些電子資料庫在顯示各種領域系統化知識，且已儼然成為全人類社會的資產。然在智慧街燈的研究與開發中，含有多項次技術領域，本研究旨在探討無線感測網路（ZIGBEE）論文科學性論文文獻計量分析。過去文獻停留在傳統調查性之質性內容分析或計算前後數年前間比率的簡易計量分析，顯然對投入學術與產業研究者所需精確專業領域資訊，付之闕如。若能採用使用文獻內容書目對及共現字之科學性計量分析以發掘研究問題的機會，如文獻發展的趨勢、類別與熱點等變項分析，應可提供創造一個新的有效性的研究問題來增強研發前沿的機會。本研究目的在建立一個探討文獻內容發展的趨勢分析模型，並使用CATAR內容分析模型做為發掘無線感測網路（ZIGBEE）文獻發展的趨勢、類別與熱點分析。本研究的第一階段採用檢全原則，建立相關鍵字，進行檢索WOS資料庫，接者使用檢準原則，經過篩與資料清洗以取得具高效信度茲分析資料集。再以統計分析，來建構文獻內容的概觀模型。再使用文獻書目對與共現字之相似以形成相似矩陣，再以數與構面觀念，使用奇異值分解（Singular Value Decomposition，SVD）進行相似矩陣分析，再使用階層式凝聚合之完全連鎖法（Hierarchical-Agglomerative-Complete Clustering）集群分析方法以進行文獻的主題集群分析，再以HHI來鑑別出研究主題的熱點。分析結果揭出期刊、國家、年代之發展趨勢之競爭態勢。另也區分出七個主題群與七個共現字群，且兩者間有很強的共通性，少數個獨特性。是故；此研究架構與流程能發掘研究的前沿（Frontier）之內在與外在效度。

Continuous innovation is the foundation of both business and academic career preservation. The development of smart street lights is becoming one of the most interesting modernization tools in smart cities. A lot of basic and applied research has been invested in achieving sustainable development in business and academic circles. The results of these studies have been published in academic papers or patent documents, submitted to journals or intellectual property patent offices of various countries, and after blind review by many peers, they were published copyrights or the patents were approved with exclusive rights. These achievements revealed new knowledge that helped businesses to keep pace with the times. To promote the rapid diffusion of technology research and development, they have been included in the electronic databases of the academia, e.g., WOS, Scopus, and Elsevier, and in the industry, e.g., the national patent databases including USDTO and EDO. These electronic databases contain systematic knowledge in various fields and have become the assets of the entire human society. However, in the research and development of smart street lights, many technical fields can be found. The present study explores the bibliometric analysis of scientific papers on wireless sensor networks (ZIGBEE). In the past, the literature mainly focused on the traditional qualitative content analysis of investigative nature or simple quantitative analysis of calculating the ratio between the years and the years before and after, which lacked the accurate professional field information required by academic and industrial researchers. If the scientific quantitative analysis using bibliographic pairs and co-occurrence words of literature content can be used to explore opportunities for research problems, such as the analysis of variables including trends, categories, and hot topics of literature development, it could provide a new and effective research problem to enhance opportunities at the forefront of R&D. The CATAR content analysis model was used to explore the trends, categories, and hot topics in the development of the wireless sensor network (ZIGBEE) literature. The first stage of this research adopted the principle of verification, established relevant keywords, searched the WOS database, and used the principle of verification to obtain a data set with high reliability after screening and data cleaning. Statistical analysis was then used to construct an overview model of the literature content. After, the similarities between bibliographic pairs and co-occurring words were identified to form a similarity matrix, and then reduce the variation. Additionally, the number and aspect concept and the SVD (Singular Value Decomposition) were employed to conduct a similarity matrix analysis. Next, the Agglomerative Hierarchical Clustering method was used to perform a subject cluster analysis of the literature. Finally, the HHI was utilized to identify the hot topics of research themes. The analysis results showed the competitive situation of the development trends of journals, countries, and years. In addition, seven theme groups and seven co-occurring word groups were distinguished, and it was found that there is a strong commonality between the two; however, a few are unique. Therefore, this research structure and the process could explore the internal and external validity of the research frontier.