Method and system for customer order weight prediction based on xgboost model

Abstract

The invention discloses a method and system for predicting the weight of a customer's order based on an XGBoost model, wherein the method includes: judging whether the current ordering customer is an unapproved customer; if the current ordering customer is an unapproved customer, obtaining the The customer's current order and historical order data; combine the current order and historical order data to carry out feature engineering to construct the mold entry feature variable associated with the current order; The input feature variable is input into the corresponding trained XGBoost model to obtain the predicted value of the order weight. By confirming that the customer is a customer who is not allowed to place an order, the present invention constructs a model-entry feature variable associated with the current order, and inputs it into a corresponding prediction model for prediction, so as to obtain a relatively accurate order weight prediction value, avoiding the need for customers to provide The order weight is inaccurate and the pickup vehicle arranged is unreasonable, which reduces the empty run rate of the pickup vehicle and saves transportation costs.

Description

Technical field [0001] The invention involves the field of logistics and cargo forecasting technology, especially a customer -based order weight prediction method and system based on the XGBOOST model. Background technique [0002] In the logistics industry, when customers fill in the approximate weight of the entrustment, due to various reasons (errors, disorderly filling), they often cannot accurately estimate the actual weight of the entrusted goods The deviation is too large. For example, the actual weight of a client's entrusted object is 500 kg, and the customer only fills in 50 kg due to personal negligence when placing an order. The direct consequence of this is that the load of the vehicle with the scheduling arrangement is not matched. However, the goods only arranged a small van. When the driver arrived at the customer, he found that the van could not load all the supporting objects, and he could only return to the point to replace the car again. For example, the actua...

AI Technical Summary

Problems solved by technology

[0003] The invention provides a method and system for predicting the weight of an order placed by a customer based on an XGBoost model, so as to solve the problem in the existing logistics industry that in the existing logistics industry, when the customer places an order, the order weight is incorrectly filled in, resulting in unreasonably arranged pick-up vehicles and waste of resources.

Images