随着移动互联网行业的高速发展，无线网络业务量呈爆发式增长，网络压力突增，传统模式下高度依赖人工静态配置基站软硬件能力，难以长时间适配当前业务需求。同时，面对海量历史业务数据，缺乏智能化精确小区业务趋势评估能力，无法对网络容量保障与扩容做出提前投资预判，难以保障运营商小区业务健康发展。该文基于时序模型统一语言模型UniLM(Unified Language Model)，建立了小区级未来流量长时间预测方案，旨在赋能网络优化领域快速定位高低业务量区域。
Research on Business Volume Prediction Method Based on UniLM Generative Formula
With the rapid development of the mobile Internet industry,the volume of wireless network business is growing explosively,and the network pressure is increasing sharply.In the traditional mode,it is highly dependent on the ability to manually and statically configure the base station software and hardware,which is difficult to adapt to the current business needs for a long time.At the same time,faced with massive historical business data and a lack of intelligent and accurate trend assessment capabilities for community business,it is difficult to make advance investment predictions for network capacity guarantee and expansion,making it difficult to ensure the healthy development of operator community business.This article is based on the Unified Language Model(UniLM)of the time series model,and establishes a long-term prediction scheme for future traffic at the community level,aiming to empower the field of network optimization to quickly locate high and low traffic areas.
万方数据
维普
