Prediction of Repeat Customers on E-Commerce Platform Based on Blockchain

Table 2

The top 10 impact features of the DeepCatboost individual model.

Ranking of features	Feature detail
1	The ratio of shopping times during promotion to nonpromotion days
2	The ratio of purchasing times to browsing
3	The total historical shopping times of a buyer from some seller
4	The ratio of purchasing times to collecting
5	The time interval between the first purchase and the last purchase of a buyer from a seller
6	The interaction frequency between a buyer and a seller measured in months
7	The total number of goods purchased by different age groups of buyers from a seller
8	The total number of buyers browsing some seller
9	The total number of bands purchased by a buyer from a seller
10	The total number of goods purchased by different sexes of buyers from a seller