Development And Application Of Intelligent Community Portrait Algorithm Based On Machine Learning

With the development of data analysis technology,it has become an important means to integrate cell wireless network data and construct cell portrait with data analysis to provide support for network optimization.This application is of great significance for improving operator network optimization efficiency and user experience.The cell portrait system in this paper is based on the network optimization platform from a network operator in China.The algorithms in it integrate and analyze the key quality indicators data and key performance indicators data.Many machine learning algorithms and data mining algorithms are used to form a variety of community portrait labels and complete many functions in this system.It also provides data support for network optimization personnel.This system is of great value to the intelligent upgrading of mobile communication,forming network intelligence optimization strategy,promoting the Intelligent iteration of mobile network.In addition,it is very important to enhance the efficiency of human and time as well as improve the experience of using network.This paper mainly studies the algorithm part of the cell portrait system.First of all,on the basis of the product requirement description document,this paper analyzes the demands of the community portrait system.At present,network optimization has drawbacks such as time consuming and lag.In order to solve these problems,the network optimization platform needs to predict and judge the wireless user perception of the cell in advance,output the business scenes of the cell,and establish the correlation between key quality indicators and key performance indicators,so as to realize the timely regulation of network problems.At the same time,the system also needs to provide data processing functions,including outlier statistics,missing value ratio and so on.Secondly,on the basis of demand analysis,this paper further studies the product architecture of the community portrait and designs the logic flow of the overall community portrait algorithm in detail.The algorithm studied in this paper is based on the historical key quality indicators data,combined with the K-means clustering of the autoencoder and data mining in machine learning and Apriori association analysis algorithm to realize the wireless sensing intelligent discrimination and indicators association of the cell.Matplotlib package in Python is used to render the results visually.Thirdly,this paper tests the functions of the cell portrait system and the performance of the core algorithms.Functional tests such as data processing and business scene division are all passed.By changing the K value in the clustering algorithm,the optimal performance solution of the cell wireless sensing discrimination algorithm is obtained by comparison.The final algorithm model recognition rate reached 93.7%,which meets the requirements of operators' requirements documents.At the last,this paper describes in detail the design and implementation process of the algorithm model of association analysis,and proposes a creative method,the sliding window method for the discretization of indicator data,which plays a vital role in the process of association analysis.The research in this paper provides important reference for this operator to accurately locate the residential business and realize network optimization,and also provides reference for other operators to build similar portrait system.The high precision of portrait algorithms greatly improves the efficiency of network optimization,which is an innovation and breakthrough of network optimization toward intelligence.