產業創新活動調查可幫助我們了解產業界中新知識的創造以及現有知識的新應用,尤其是我國中小企業在研究發展以外的創新活動是相當活躍的。Knell 與Nas (2006)及Raymond (2006)等人指出在研究創新過程以及「投入與產出」會有一些限制,即在收集創新調查資料時,調查資料時間點是一項根本的問題,必須以長期追蹤資料(panel data)才能完整呈現「投入與產出」的關聯性。目前歐盟CIS 調查已進行到第五次,我國創新調查也已進行兩次,由於調查樣本的不同,或調查問卷中問項題目會隨著不同梯次進行增修,造成了不同梯次的創新調查所得的資料並不相同,各梯次的資料庫無法直接串聯合併成長期追蹤調查資料庫。本研究嘗試以資料庫的函數映射(functional mapping)技術進行產業創新資料庫中變數與資料擴增,期能達成建構長期追蹤的創新資料庫之目標。研究結果顯示資料庫映射結果的相對誤差率RER 維持在1 上下,並沒有造成太多的誤差,結果令人滿意且穩定。表示利用資料庫函數映射技術確能將原有的資料庫不足之處加以擴增,解決資料短缺問題,形成長期追蹤資料,提升資料庫的價值。 Innovation is the key of survival for small and median enterprise (SME). It was important to understand the relation between innovation activities. However, Knell and Nas (2006) and Raymond et al(2006) indicated the limitation on studying the innovation process, such as the “timing” is the essential problem as collecting data. It was impossible to draw a whole picture about the relations of input and output without having panel data. Nowadays Community Innovation Survey (CIS) had been conducted the fifth time in Europe and the second time in Taiwan. The questionnaire using in Taiwan Industry Innovation Survey had been modified 2 times. The contents and variables were not the same within these 2 databases. As the consequence, the 2 databases cannot be combined as a panel database. This study was trying to use functional mapping technology to expand the data and variables and to build a panel database for Taiwan innovation survey. According satisfied outcomes of this research, RER varied around 1. It meant using functional mapping technology to expand Taiwan innovation survey database was practicable.