| abstract |
The present invention provides a dynamic high-risk customer group detection algorithm, which is used to detect the risk degree of new users belonging to different customer groups, including the following steps: i. Using a clustering algorithm based on multiple historical data of multiple historical users Building a plurality of historical user groups to form a plurality of customer groups, the clustering algorithm is related to the type of the historical data; a. determining the risk degree of each customer group based on the evaluation index, and dividing the risk degree The customer group greater than the risk threshold is determined as a high-risk customer group; b. Determine the similarity between the new user and the high-risk customer group based on a distance function, and the new user whose similarity is greater than the first similarity threshold Determined as a risk user, the type of user data of the new user described in the distance function is related, and the beneficial effects of the present invention are redefined here. The present invention is simple to operate and easy to use, and provides a dynamic risk customer group detection algorithm and corresponding system with high commercial value. |