| Summary: | 碩士 === 國立高雄應用科技大學 === 金融資訊研究所 === 102 === The main objective of this article is to investigate attributes of customer affecting an insurance contract. Therefore, cross analysis, One-sample t test, One-way Analysis of Variance, Logistic Regression model and Decision Tree Model analysis are applied to analyze the correlation between the variable of attribute of customer and the content of insurance contract.
The empirical results are as follows:
1. Cross analysis: “Gender” and “the insurance age of the insured” are significantly correlated to “insurance contract status”; however, “occupation” is not significantly correlated to “insurance contract status”.
2. One-Sample t test: “The insurance age of the insured” and “insurance contract status” are significantly different.
3. One-way Analysis of Variance: “Age” shows significant difference to “insurance value”, “insurance fee” and “insurance contract status”; moreover, “occupation” shows significant difference to “insurance value” and “insurance type”.
4. The result of Logistic Regression model shows that: “Age” is significant variable of “insurance p contract status”.
5. Decision Tree Model: The result shows that “gender”, “age” and “insurance type” are significant variables affecting insurance contract status.
6. When the prediction accuracies of models are compared, it was found that the prediction accuracies of Logistic Regression model and Decision Tree Model on insurance contract status are all 89.4%.
|
|---|
