Data model generation method and device, data model calling method and device, equipment and storage medium

Abstract

The invention discloses a data model generation method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring various different customer data; performing data processing on the plurality of customer data to obtain a plurality of corresponding sample data; inputting the plurality of sample data into a deep neural network for training to obtain a universal customer data coding model; performing model post-processing on the universal customer data coding model; and adding a classifier on the processed general customer data coding model, and inputting customer response data of the corresponding task into the general customer data coding model for model fine tuning to obtain a data model of the corresponding task, so that the data modeling efficiency is improved, and the modeling cost is reduced.

Description

technical field [0001] The present application relates to the field of computer technology, and in particular to a data model generation method, calling method, device, equipment and storage medium. Background technique [0002] The bank's refined operation of customers is increasingly dependent on modern data modeling technology. For different data of different banks, different modeling teams will choose different modeling methods based on past personal experience. The method of bank customer data modeling often needs to formulate a targeted modeling scheme for a specific task of a specific bank, and formulate multiple corresponding models for multiple tasks, which consumes a lot of time. At the same time, the model is vulnerable to problems such as data imbalance and less labeled data, which seriously affects the performance of the model. Moreover, the prepared model can only be applied to a specific task, and the validity period of the model is relatively short. After th...

AI Technical Summary

Problems solved by technology

The method of bank customer data modeling often needs to formulate a targeted modeling scheme for a specific task of a specific bank. It takes a lot of time to formulate multiple models corresponding to multiple tasks

At the same time, the model is vulnerable to problems such as data imbalance and less labeled data, which seriously affects the performance of the model

Moreover, the prepared model can only be applied to a specific task, and the validity period of the model is relatively short. After the validity period, it needs to be remodeled according to the latest data, which is costly.

Images