對壽險公司而言,從現有的保戶篩選出投資型保險的潛在客群,是極為有效的行銷策略。本文試圖利用遺傳演算法,找出最佳的羅吉斯迴 歸模型,預測現有保戶對投資型保險的購買決策。由於投資型保險的保戶必須自行承擔投資的風險,其風險態度應為影響其購買行為的重要變數。本文同時考慮了財務風險態度與一般風險態度。另外,對於保戶風險態度的分類,除了傳統的平均數標準差法,灰色統計提供了另一個可能的選擇。本文發現:灰統計較適合結合基因演算法的羅吉斯迴歸模型;平均數標準差法較適合結合逐步羅吉斯迴歸模型。最後,財務風險顯然是保戶在購買投資型保險時所相關的。
For life insurance companies, selecting the potential buyers of investment-linked insurance out of the existing policyholders is an effective and economic strategy. We try to use logistic regression model combined with genetic algorithms and grey clustering statistic to forecast the policyholder's purchase decision of investment linked Insurance. Because policyholder of unit linked insurance bears the investment risk, their risk attitude should have a great impact on their purchase decision of investment-linked insurance. We take general risk attitude and financial risk attitude into account at the same time. Grey clustering statistic offers an alternative to the traditional methods of classification for risk attitude. We find that grey clustering statistic is more suitable for GA-based logistic regression; the mean-and-standard-deviation method is more suitable for stepwise logistic regression. Finally, financial risk attitude is more relevant for policyholders' purchase decision than general risk attitude.