如何有效保存與擴散人類的知識向來為知識管理的關鍵議題，過去以來，要理解人物的思想脈络往往需從專人訪談或專書獲得，但通常這個過程耗時且耗成本，有時因為人為或環境的因素而無法順利達成。因此，本研究嘗試應用現代新興文字採掘技術中的人工智慧分群方法—Growing Hierarchical Self-Organizing Map(簡稱GHSOM)，探討該方法在人物知識管理應用上的可行性，並以管理領域中的主導管理邏輯(dominant managerial logic)之應用為例。由於台塑集團為台灣第一大的民營企業集團，其精神領袖王永慶先生更被喻為台灣的經營之神，為享譽國際的企業家，其成功的經營管理經驗因此成為管理學界及管理實務上極欲探討的管理邏輯思想，因此，本研究將以王永慶先生的言論為資料源，建構王永慶先生的知識地圖（又稱主題地圖）。研究結果初步發現該技術有助於輔助建構王永慶先生的知識地圖，但仍有其應用上的限制，尤其在抽象的精神層次之觀點上，自動分群並無法具體呈現這類的知識，僅在功能層次的觀點上，獲得不錯的分群效果，可明確將知識顯示於知識地圖中。

The storage and diffusion of corporate leaders’ knowledge has long been a major management issue. In the past, to learn the dominant logics of these leaders, we either rely on storytelling by face-to-face interview or reading their autobiographies. That usually is time-consuming and costly and feasible only if you get chances to have in-depth interviews or their autobiographies do exist. Recently, some text mining approaches have been proposed to help extracting knowledge from various document sources with some degree of success.
In this research, we apply growing hierarchical self-organizing map (GHSOM), a widely used text mining tool, to analyze the dominant managerial logics of Mr. Y.C. Wang, the Chairman of Formosa Plastics Group, (also known as the God of Business in Taiwan). A collection of speeches given by Mr. Wang in various occasions is used to construct topic maps, representing his management philosophy by using GHSOM. The research results do show some usefulness of GHSOM in clustering documents based on their main management concepts. While for concepts with functional orientations the results are quite promising, those of more abstract nature do pose some difficulties to cluster.