利用NLP算法校准宽带小区资源地理位置

打开文本图片集

中图分类号：TN91；TP3 文献标志码：A 文章编码：1672-7274（2026）02-0050-03

Calibrating the Geographic Location of Broadband Community Resources Using NLP Algorithms

LI Jiran1， CHE Bin²， WANG Yonggang³， JI Shangpeng³， SONG Zirui'， RENYangbo1，SEN Chenghu1，YUANXizheng1

（1.Beijing Telecom Planning&Designing Institute，Beijing 10oo37， China; 2. China United

Network Communications Group Co.， Ltd.， Beijing 10oo33， China; 3.China United Network

Communications Group Co.，Ltd.Network Operation Department，Beijing 10oo37， China）

Abstract： As a core digital asset for broadband resource coverage，the accuracy of standard addresses directly determines the efficiency of service acceptance and activation.To address current pain points in terminal resource management for household broadband--such asambiguous descriptions，duplicate entries，hierarchical confusion，and spatial coordinate deviations of standard addresses.This paper proposes a social governance approach for standard addresses byconstructing anauthoritative，multi-source integrated social addressreference database.Utilizing NLP algorithms，a governance mechanism incorporating hierarchical rule-based cleansing， GIS coordinate alignment，and dynamic feedbackcalibrationis established.

Keywords： social addresses; standardized addresses; broadband resource coverage; address standardization

研究背景

从网络建设转为网络运营为主的通信大背景下，作为精准营销、精准运营的载体，标准地址数据是贯穿B域和O域的唯一生产要素[1，是宽带基础设施与数字服务的关键空间纽带，其核心价值已从基础资源记录升级为全业务链路的智能调度中枢。（剩余3584字）