Merchant returned customer prediction method based on big data

Abstract

The invention requests to protect a merchant returned customer prediction method based on big data. The method comprises the following steps: 101, carrying out preprocessing operation on historical behavior data of a consumer; 102, training set data and verification set data are divided according to historical behaviors; 103, performing feature engineering operation on the historical behavior dataof the consumer; 104, performing feature selection on the sample set with the constructed features; 105, establishing a plurality of machine learning models, and carrying out model fusion operation;106, predicting whether the new buyer buys the commodity again in the same merchant in six months in the future according to the historical behavior data of the consumer through the established model.According to the invention, pretreatment and analysis are carried out to extract features, features are selected, a plurality of machine learning models are established, and whether a new buyer purchases commodities in the same merchant again in six months in the future is predicted according to consumption behavior data of the consumer in half a year before Double Eleven, thereby providing services for the merchant to accurately locate potential loyalty customers, reduce the promotion cost and improve the investment return.

Description

technical field [0001] The present invention relates to the field of machine learning and big data processing, especially feature construction, model design, model fusion scheme and the like. Background technique [0002] With the rapid development of the global economy, the era of economic information is advancing steadily. As a new way of business circulation and communication with people, e-commerce has emerged in the era of information economy. The rapid development of e-commerce has also brought huge competitive pressure to various merchants, so merchants sometimes conduct big promotions on certain dates and time points, and attract a large number of new buyers in this way. But unfortunately, many buyers who are attracted by promotional activities are "deal hunters" who only spend one-time consumption. As a result, these promotional activities carried out by merchants cannot obtain long-term benefits and fail to achieve the real purpose of promotion. [0003] During th...

AI Technical Summary

Problems solved by technology

The traditional linear weighted fusion is used in model fusion, which mainly assigns weights to each model manually based on model performance and personal experience, which will bring large errors

Images