| 摘要: | 企業不易及時與敏銳地察覺技術的變動，將導致其研發成果被取代性高或研發過時的技術，造成研發資源的排擠與浪費。利用專利分析可以及時辨別與確認技術的變化以及增加企業的技術競爭力而獲取利益。視覺化的專利地圖能用最少的文字說明技術情境，或在較短的時間內簡化與表達隱藏的技術情報。目前專利分析軟體的功能如PatentGuider 或PatentTech 係依個別變數（發明人、技術分類號、國家別、權利人等）進行敘述統計分析，但無法在專利地圖中同時呈現多變數間的多重對應關係與揭露隱藏的技術情報。統計方法中的多重對應分析(Multiple Correspondence Analysis, MCA)能以距離遠近（或相似程度）與分佈密度的方式有效地挖掘出多變數間對應的關聯性，並以視覺化繪圖效果揭露出隱藏之意涵。本研究提出整合對應分析與集群分析方法製作具有多變數對應關係的技術專利地圖，利用幾何空間的直觀概念更能有效地呈現專利權利人、技術分類號與申請日期的對應關係與隱藏意涵。為了洞察台灣企業技術佈局的全球情勢，本研究亦將融入國家別變數於技術專利地圖中，揭露的技術情報能協助研發人員掌握企業與其競爭對手的技術研發歷程，察覺產業當前的技術缺口與預測該技術的發展趨勢。鑑於台灣薄膜太陽能電池產業發展之重要性與日俱增，本計畫將針對此個案產業為例，發展其多重對應關係之技術專利地圖。完整期間的技術專利地圖除能顯示企業專利技術的分佈情形，並可獲知企業在各技術領域研發能力的優劣。藉由比較多期間多重對應關係之技術專利地圖異同可洞察創新技術的演化歷程，探索產業的技術缺口，以及預測未來技術發展的走向。本研究為二年期的計畫案，第一年將完成薄膜太陽能電池專利文件的蒐集、整理、建檔，以製作多個不同時期之多變數多重對應關係的技術專利地圖。第二年將針對各期間技術專利地圖的權利人、國家別、分類號與申請日期變數進行資訊整合分析，並與專家共同討論解析各地圖的技術情境。本計畫研究成果將可提供太陽能電池產業發展與廠商擬訂研發策略之依據。Enterprises that are unable to perceive technological changes quickly or keenly enough could end up for having replaceable or outdated technologies that result in the crowding out and waste of R&D resources. Patent analysis allows timely identification and confirmation of technical changes and helps an enterprise staying at technically competitive edge and profitable. The patent mapping exploited least efforts to mining technical situation and extract the simplified underlying technical information by graphical displaying. Conventional software of patent analyses, like Patent Guider or Patent Tech, provide the individually summarized function over categorical univariate such as the righter, patent classification, country code, however, the associations or correspondence among variables and the latent patent information are not yet unveiled. Multiple correspondence analysis (MCA), statistical approach, allows to explore the associations among variables in accordance with the distance/similarity of geometry and visualize the results onto a lower-dimensional graphical displays simultaneously. In this study, the integration of correspondence analysis and cluster analysis is proposed for making the technical patent maps, which use the concept of geometric space, for more effectively showing the corresponding associations among the righter, patent classification, and year of application. Moreover, in order to further investigate the global scenario for the technical deployment of Taiwanese enterprises, the country code of patent information is also incorporated into the proposed technical patent maps to provide overall view of technological advancement for researchers in monitoring competitors’deployment, mining the techniques gap and forecast new trends of technology. Due to the significant increasing value on the evolution of thin film photovoltaic industry in Taiwan, the aim of this study will develop the multiple correspondence technical patent maps for this industry. The proposed patent map made from the whole period patent data may reveal the deployment information of the enterprises and show its strength and weakness in various domains of technology. By comparing the patent maps with different periods, the insights on the progress of technical development, technical gap, and forecast technical development can be obtained. This research is a 2-year project; the expected outcomes in first year are patent data gathering, filing, conducting correspondence analysis, clustering analysis, and building up the technical patent maps for different observed periods respectively. In the second year, the scenario analysis with experts will be conducted with the full use of integration information among righters, countries, categories, and years of applications. The results of this study can serve as a basis for formulating the R&D strategies of Taiwanese enterprises and industrial policy upon the thin film photovoltaic industry. 研究期間 : 9808 ~ 9907 |
