Method and device for carrying out user grouping and quota withdrawing based on variable influence degree index and electronic equipment
A technology with influence and user, applied in data processing applications, instruments, finance, etc., can solve problems such as risk strategy incompatibility
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0071] figure 2 It is a schematic diagram of a method for determining a quota increase strategy based on user grouping according to an embodiment of the present invention. Such as figure 2 As shown, this embodiment is to obtain a risk control strategy for increasing the amount. In order to obtain the risk control strategy for the amount increase, first formulate a grouping rule, group the user data obtained from the historical financial user concentration, and obtain four user groups: user group 1, user group 2, user group 3, and user group 4.
[0072] In this embodiment, we use the discrimination index to establish grouping rules. That is, we first calculate the discrimination index of each variable of the user data in the historical financial user set. In order to calculate and find out the variant with high discrimination degree from multiple variants, we use the decision tree model here. The decision tree model belongs to the machine self-learning classification mode...
Embodiment 2
[0096] Different from Embodiment 1, this embodiment adopts different grouping rules. In this implementation example, the influence index of the independent variable on the dependent variable of the user data in the historical financial user set is calculated, and the grouping variables and grouping rules are determined according to the influence index. Here, the Boruta algorithm is used to calculate the degree of influence of the independent variable on the dependent variable.
[0097] Boruta is a variable selection algorithm. To be precise, it's a wrapper algorithm around Random Forest. We know that feature selection is a critical step in predictive models. This step is especially important when building a model where a dataset contains multiple variables.
[0098] Using the Boruta algorithm to calculate the degree of influence of the independent variable on the dependent variable, including: establishing a model for the entire historical financial user set, and using the ...
Embodiment 3
[0110] image 3 It is a module composition diagram of an apparatus for determining a quota increase strategy based on user grouping in the third embodiment of the present invention. Such as image 3 As shown, the device includes a rule establishment module, a grouping module, a model establishment module and a policy determination module.
[0111] The rule establishment module is used to determine the grouping rules for the historical financial user set.
[0112] Therefore, to determine the grouping rules, we must first determine the variables of the group, and then determine the classification rules according to the variables. The present invention proposes various schemes for determining grouping rules.
[0113] A preferred implementation manner is to calculate the importance index of each variable of the user data in the historical financial user set, and determine the grouping variable and the classification rule of the dependent variable according to the importance ind...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


