Optimisation and research of Los Angeles communication network hotspots based on machine learning and social software data

Abstract

With the development of the times, people have become more and more dependent on their mobile phones, while instant messaging has become a new and booming field recently, and social media plays an important role as a communication medium for people to communicate with each other and share their lives. Using Twitter data with various keywords, researchers can analyse when, where and what the tweets were posted. Machine learning can further analyse the local user traffic and the location where users often send data so that new fifth-generation (5G) base stations can be deployed. The innovation of this simulation is to use machine learning algorithms to cluster the Twitter user data in Los Angeles. Two clustering algorithms, DBSCAN and K-means, are used to optimise the simulation through parameter settings and contour functions to determine the value of K to determine the number of clusters needed in the local area. The clustering results are analysed and processed, and the 5G base stations are deployed in the city with fourth-generation(4G) base stations already covered, thus helping the operator to deploy base stations more accurately and saving the company's overhead. The optimal K-value for the algorithm was obtained through simulation, and targeted optimisation was made for some areas based on the local geographical location and the distribution of hotspots and blackspots of passenger traffic. The simulation improves the targeting of the signal base station deployment of the telecom operator company. Unlike the previous aim of pursuing full coverage of the area, this simulation hopes to analyse the hotspot and blackspot of the whole city through the existing communication data, so that the base station deployment can be guided by the data, which is a more direct and efficient method.