以多準則決策與專利探勘方法定義機器視覺之技術地圖

Abstract

大數據概念迅速發展，成為跨越資訊技術顯學，在眾多樣態、良莠不齊的巨量資料中，精萃出有潛在價值、商業趨勢的資訊，提供企業改善與經營決策之參考依據，進一步提升市場上競爭力，真正實現大數據的價值。專利是推動科技進步的重要機制，在法律嚴密的保障下，激勵人們發明創新與增進經濟發展，並可作為競爭分析及技術發展基礎。以往技術路徑圖的相關研究通常聚焦於特定技術的結果，或是企業技術發展歷程，較少有研究探討如何分析專利探勘結果產生技術路徑圖藉以預測未來技術趨勢。因此，本研究透過文字探勘技術收集專利關鍵字，並依據專家評估意見，進行關鍵字分組。再透過優勢關係為基礎的粗略集理論推導技術關鍵字與產品關鍵字間之推理關係，再導入正規化概念分析，將技術關鍵字歸納為技術之概念，最後利用決策實驗室分析法訂定關鍵字及概念之影響關係，參考組織現況定義出技術路徑圖。實證研究以機器視覺技術為例，進行專利文字探勘並定義出技術路徑圖，驗證此方法架構的可行性，期許本研究結果可作為產業技術發展之依據。

The concept of big data has developed rapidly, it is important to find out meaningful information from big and different quality data, and then transform these information into visualized chart is an important issue for all companies. Patent is an important mechanism to push scientific and technological progressive, because of its highly exclusive and well protect by the law that will encourage people. Many researches of technology roadmap usually focus on specific technology or on the development of technology in enterprises, seldom on deriving the technology trends in the future. Therefore, this research will collect keywords of patents by using text mining and grouping these keywords by experts’ opinions. Then, deriving the inference relationship among the keywords of technology, product by using DRSA, conducting the technology concept from the keyword by using FCA and deriving the influence relationship between the keyword and concept by using the DEMATEL. Finally defining the technology roadmap with consideration of the current situation of the organization. Empirical studies based on the patent analysis of the Machine Vision defined the technical roadmap which can be used to demonstrate the feasibility of the proposed analytic framework. The analytic framework and results can serve as the industry develops technology in the future.