隨著網際網路技術的快速發展、資訊量的暴增，使用者在按索網路資料從往往需再審視大量的按索結果，以發掘出符合需求之資訊，此種現家突顯出目前之資訊搜尋技術和知識獲得間存在著缺口，其間需技合更精準的訊採擷技術方竟其功，本研究遂以隱匿式馬可夫模式為基礎，開發出強效精準的資訊採擷工具，研究過程係透過網路取得樣本，在武器獲得方面，從「國防工業預測情報 j以關鍵字取得分類領域資訊，擷取廠家名稱、武器型式、金額等資訊 ﹔同時在情報蒐集方面，從中國的人民網國際軍事新聞上，建立以簡體中文為分類領域的擷取樣本，擷取人名、組織名稱、武器型式、軍事議題、國家地名等資訊，透過隱匿式馬可夫模式探討適用於軍事領域的資訊擷取模型，研究實證均有 80%以上的擷取成果。

As internet technology is fast developing and web information volume is also dramatic increase, users usually have to put considerable effort to review again the outcome of web search so as to dig out the real information required by the users. Such phenomenon reveals the fact that there is gap hold between the currently exist searching technology and knowledge acquirements. One therefore has to involve more accurate information extraction technique in order to achieve accurate information extraction in further. This study thus uses hidden Markov model (HMM) as basis to build up robust and accurate information mining tools for digging defense acquisition related information. The training samples for constructing the HMM are derived from the web. The contents of "Forecast International" are used for extracting the weapon-relating information including the company, weapon types, and cost. While the samples obtained from Chinese web "People’s Daily Online" are used for extracting information regarding names, organizations, weapons, military issues, and nations. The outcomes of study reveal that, by using HMM for information extraction from military field, the accuracies can achieve up to 80%.