Method for constructing address classification model, address classification method and related equipment

Abstract

The invention provides a method for constructing an address classification model, an address classification method and related equipment, and the method for constructing the address classification model comprises the steps: dividing a to-be-classified region into a plurality of sub-map regions, so as to obtain map category data; dividing the historical delivery addresses of the to-be-classified area into a plurality of delivery categories to obtain delivery address category data; and training an address classification model based on the map category data and the delivery address category data, the address classification model being used for predicting a delivery category to which a to-be-classified address belongs. According to the method, the address classification model is optimized, so that accurate address classification can be carried out on address positions or buildings without historical delivery records, the logistics timeliness is improved, and meanwhile, the logistics experience of users is improved.

Description

technical field [0001] The invention relates to the field of computer applications, in particular to a method for constructing an address classification model, an address classification method and related equipment. Background technique [0002] In the logistics scenario, after the user generates a logistics order, the logistics platform or other platforms with logistics scenarios such as e-commerce platforms need to infer the various circulation links that need to go through from the place of origin to the place of receipt according to the address information filled in by the user. , for express delivery. This series of circulation information is mainly identified by the three-segment code printed on the logistics face sheet. However, the three-segment code needs to deduce the node that dispatches to the address according to the detailed address filled in by the user. The commonly used inference model relies on a large number of dispatch records in the area to learn. For ...

AI Technical Summary

Problems solved by technology

[0003] However, for addresses or buildings that do not have delivery records in historical logistics data, it is difficult for common inference models to learn the corresponding delivery nodes

For example, [Changfang International Plaza] has not appeared in the historical dispatch records, and the reasoning model is difficult to construct the mapping relationship between [Changfang International Plaza] and any node

The current method will lead to the writing of logistics orders for addresses or buildings that have no historical delivery records. In the reasoning model, the delivery node cannot be determined, so the three-segment code cannot be determined, and the three-segment cannot be printed on the logistics form. Therefore, when sorting logistics packages, manual judgment and sorting is required, which reduces the timeliness of logistics and affects user experience

Images