Research on Model and Algorithm Optimization of Site Area Clustering Problems in Mobile Communication Networks

In mobile communication network planning, base station planning acts a significant role in the performance of the mobile communication network. Among them, site planning is particularly critical. In practical work, it is required to execute regional clustering of weak coverage points in order to better handle the weak coverage problem. Therefore, this paper designs a variety of clustering algorithm models and analyzes the relevant parameters in the clustering process to calculate the time complexity and evaluate the models. First, the second-order clustering method is selected to determine the number of clusters, and then K-means clustering analysis is used to classify all points. However, when the amount of data is large in K-means cluster analysis, the algorithm has a high level of temporal complexity. Therefore, it is further optimized based on K-means clustering, and it is converted from numerical clustering to density clustering to be more suitable for actual data. The data is clustered by DBSCAN clustering method, and a clustering model with relatively good anti-noise and robustness is obtained. But the time complexity and space complexity of DBSCAN clustering algorithm are not low. So we choose the K-nearest neighbor algorithm and implement it with a K-d tree, and use the K-nearest neighbor algorithm to test the K-d tree. Finally, the time complexity of each clustering method in the classification process is compared and compared, and it is concluded that the K-d tree algorithm has the lowest time complexity. The Internet era's and mobile communication technology's rapid development, it provides a new idea for upgrading and optimizing the regional clustering of mobile communication networks.