AI-based 5G Resident Ratio Prediction for Multi-Frequency Heterogeneous Networks.

This paper studies the prediction and evaluation scheme of the 5G resident ratio of multi-frequency heterogeneous 4G/5G networks. It proposes a prediction method for the 5G resident ratio in combination with multidimensional data sources. Based on the calibration of the traditional 5G wireless propagation model, this paper introduces the Gradient Descent Algorithm. Based on the massive DT/CQT test data, it accurately evaluates the improvement and impact of 900M low-frequency network construction on the 5G resident ratio. Based on the intelligent prediction model, the theoretical limit value of the 5G resident ratio is evaluated accurately, which will guide network planning and 4/5G network collaborative optimization.