基于LSTM神经网络算法互联网电视EPG业务隐患预测的应用研究
Application Research on Prediction of Hidden Dangers in Internet Television EPG Business Based on LSTM Neural Network Algorithm
班雪飞 1倪峰 1周玮 1马占婕 1孙忠岩1
作者信息
- 1. 中国移动通信集团内蒙古有限公司,内蒙古 呼和浩特 010010
- 折叠
摘要
伴随三网融合的深入开展以及智慧物联网技术的发展,家庭宽带加互联网电视已成为新的智慧家庭入口.为了解决互联网电视业务质量监测手段在时序上的先天滞后无法先于用户发现隐患故障的问题,通过引入神经网络LSTM算法实现数据注智业务运维能力,聚焦EPG业务服务质量,从业务历史波动识别异常并做出预测预警,实现EPG业务质差隐患识别和预测,隐患发现时长缩短至0.5小时,隐患识别及时率和准确率均在90%以上.
Abstract
With the deepening of the integration of the triple play and the development of smart Internet of Things technology,home broadband and internet television have become a new entry point for smart homes.In order to solve the problem of congenital lag in the timing of internet television business quality monitoring methods that cannot detect hidden faults before users,a LSTM neural network algorithm is introduced to achieve data injection intelligence business operation and maintenance capabilities.It focuses on the quality of EPG business services,identifies anomalies from business historical fluctuations and making predictive early warnings,and achieves the identification and prediction of EPG business quality hidden dangers.The length of hidden dangers discovery time is reduced to 0.5 hours,and the timely and accurate rates of hidden danger identification are above 90%.
关键词
互联网电视/EPG业务质量/LSTM算法/隐患预测Key words
internet TV/EPG business quality/LSTM algorithm/prediction of hidden dangers引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
维普
万方数据