本研究為台灣人工智慧產業公司經營績效之研究，因此取樣商業週刊、大和證、鏡週刊、CMoney等推薦近年來人工智慧相關產業概念股，並由公司產業性質分為人工智慧硬體設備相關類股、半導體相關類股及其他相關之上市櫃的公司，共計 53家，為2014-2016年之636筆季資料，資料來源為台灣經濟新報（TEJ）。依據資料包絡分析法進行投入與產出間之經營效率分析，其中員工人數、營業費用及固定資產為三項投入變數，產出變數則為市場附加價值，並納入股權結構的董監事持股、經理人持股及十大股東持股等變數，檢視影響效率值之因素。本研究得到以下幾點結論。（1）員工人數、營業費用、固定資產皆對MVA有正向顯著關係。（2）人工智慧產業製造業的公司相對無效率數值偏高，資源配置有嚴重失當的現象。（3）分析各公司近3年的MVA及經營效率值後得知，人工智慧製造業中分類為半導體相關類別及人工智慧硬體設備類別的公司皆較其他相關類別的公司整體表現來的優異。（4）董監事持股對於人工智慧製造業經營效率呈現負向關係，與十大股東持股呈正向關係，並且結果都顯示顯著。

This research studies efficiency analysis of Taiwan artificial intelligence Industry. The findings are as follows. First, the number of employees, operating expenses, and fixed assets all have a positive and significant relationship to MVA. Second, the relative inefficiency of companies in the artificial intelligence industry is too high, and there seems to be resource misallocation issue. Third, it is found that companies in the semiconductor-related category and the artificial-intelligent hardware category were all superior to those of other related companies. Finally, it is shown that top ten main shareholders' holdings of the underlying companies are positively correlated with the efficiency of the companies.