一個好的財務規劃不僅可以建立好的資產配置更可滿足顧客的需求。而個人理財服務的評估不是一件容易的工作，係因為要考慮到個人背景、觀察能力及個人資產配置的決定。而相關知識的建立可從觀察樣本而得知。從樣本中歸納規則取得知識可用於特別用途。本文提出一個約略集方法，除了針對歸納規則作加值應用外，並解決在財務或保險相關資訊不確定的問題。此混合約略集方法內容共有三種：(1) FSBT方法可以增加規則的解釋性，(2) 精簡度可是作為規則篩選門檻設定的方法，(3)利用正規概念分析以取得預知知識以增進決策制定的效果。因為混合約略集方法主要是針對規則的使用，所以可適用於任何方法所產生的規則。本研究利用實際財務資料作測試，其結果顯示所提方法確實能提供理財專家作更好的服務。 A well-designed financial plan can help to achieve good asset allocation and satisfy customer needs. Assessing personal investment portfolio was not an easy task due to concern about personal background, personal perspective and personal asset allocation decisions. That relative knowledge was usually acquired from observed data. Rule induction extracted knowledge from a set of observations which represent a data pattern may help to discover knowledge for special demand. This research was proposed hybrid rough set approaches based on the Rough Set Theory which intend value-added in the decision rules and also solve the imprecise data for finance or insurance investment analysis. Three objectives in the approaches were described as: (1) FSBT (Forward Search and Backward Trace) can increase the rule interpretable; (2) Compactness rate was a rule index selection and pruning method, and (3) applied Formal Concept Analysis to get the prior knowledge which can increase the performance of finance portfolio making. The hybrid rough set approaches focus on the decision rules which can generate from any method. One finance practical data sets were tested by the proposed methods. The results illustrate that the proposed method can induce important factors for finance experts to make optimal decision support.
