Translator Disclaimer
18 December 2019 Analysis of access network requirements based on similar service feature values clustering
Author Affiliations +
Proceedings Volume 11340, AOPC 2019: Optical Fiber Sensors and Communication; 113401H (2019)
Event: Applied Optics and Photonics China (AOPC2019), 2019, Beijing, China
This paper proposes an access network requirements analysis method based on similar service feature values clustering. The network service feature values data are collected from the real access network, including devices and users’ information. The K-means ++ algorithm is adopted to cluster the PON ports, based on the similar service feature values of users connected to them. 3 classifiers are selected for classification training and prediction. They are K-nearest neighbors, SVMs and perceptron. In the case of no clustering, K-nearest neighbor algorithm performs better, and the classification correct rate is about 91.1%. And if the data is divided into 5 groups by K-means++ algorithm, it can be calculated that the average accuracy is 94.3%. Compared with the result without clustering of service feature values, it achieves more accurate prediction and the correct rate is increased by 3.3%. With this method, the operator can determine whether the PON port needs to be expanded, making the expansion planning more forward-looking. At the same time, combining the prediction results with the PON port geographic information, it can distinguish areas with sufficient or insufficient broadband resources, and helps operators adjust service allocation plans to match user needs.
© (2019) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jinglei Sun, Hui Li, Yuefeng Ji, Guangquan Wang, Wu Jia, Yudan Su, and Yan Shao "Analysis of access network requirements based on similar service feature values clustering", Proc. SPIE 11340, AOPC 2019: Optical Fiber Sensors and Communication, 113401H (18 December 2019);


Back to Top