A bank customer intelligent marketing method based on neural point process

Abstract

The present application relates to the technical field of bank customer intelligent marketing, and discloses a bank customer intelligent marketing method based on a neural point process, which comprises the following steps: converting preprocessed customer event data into vectorized customer event data required by a neural network model input; training a neural network to obtain an optimal neural point process model; outputting the probability of each customer purchasing a target product, and saving the purchase probability prediction results of all customers; forming an intelligent marketing scheme; calculating the performance indicators of the optimal neural point process model, and incorporating the calculation results into the iterative training of the next round of neural point process model to form continuous optimization of the neural point process model. The neural point process model is used for self-adaptive learning of event sequence patterns, which can not only accurately predict the purchase behavior of customers, but also predict the most likely purchase opportunity of customers, so as to more deeply mine the purchase intention of customers and provide more abundant customer information for marketing personnel.

Images