本研究提出以知識本體做為解析新聞的方法,將新聞轉換為系統可處理的競爭智慧,並將解析後的新聞納入決策協助AHP的塑模。現階段決策,多以量化資料做為決策依據,受限於外在資訊的取得及不易轉化為系統可處理的形式等原因,因此決策系統較少納入如新聞這類的非量化資訊,然而新聞擁有許多與決策相關的資訊,為競爭智慧帶來了機會,但也造成了資訊過載的問題。過去常以關鍵字的方式來處理資訊過載,但關鍵字主要依賴比對,不容易取得符合使用者想法的結果。本研究以知識本體做為解析新聞的方法,以期可以取得符合使用者想法的決策資訊,而將解析新聞的過程分為過濾階段與決策階段兩個階段,在第一個階段以過濾本體過濾不相關新聞,第二階段再以決策本體將新聞解析成為可納入AHP決策的資訊;決策本體以AHP的決策要素為基礎建置而成,將AHP的決策要素分為正負面消息,新聞經由本體的解析會被歸類到定義好的概念中,如此就可將新聞納入決策參考。知識本體中的知識,以心智圖法(Mind Mapping)及正規概念分析法(Formal Concept Analysis, FCA)來對專家做擷取,再將擷取後的知識轉換成知識本體的型式,並以OWL的知識本體語言儲存。最後以股票投資做為系統的實例,研究結果顯示,本研究所提出的過程,確可將新聞轉換為系統可處理的形式,並納入AHP的決策程序。
This study proposed ontology-based competitive intelligence (CI) for on-line news classification. That can be applied to AHP(Analytical Hierarchy Process)-based decision support system. Quantitative data or structure data is the most common way for decision-making. Computer system has not much ability to process qualitative data or unstructured data such as on-line news. There is more valuable information inside on-line news for decision-making. Information overload is the problem of on-line news. Usually, Keywords-matching is used to solve this kind of problem, but the results at the end are not good enough to satisfy user’s mind. According to the above situation, this study proposed ontology-based competitive intelligence (CI) for on-line news classification. The first step is to sort the interrelated news using Filtering-Ontology. The second step is to decompose the news into the structure information for supporting AHP system with Decision-Ontology. Decision-Ontology is composed of AHP factors in good news or bad news. Mind Mapping and Formal Concept Analysis are used for extracting knowledge from domain expert. We use the case study of stock investment. The results show that the process of ontology-based news classification could transform news into competitive intelligence for AHP-based decision-making.