Research on the Identification Method of Cells with Deteriorating Wireless Network Indicators

The severe degradation of wireless network performance indicators in cells seriously affects user perception. Rapidly identifying small cell performance degradation based on fixed thresholds set by network optimization personnel through business experience can lead to misjudgment and omission. An anomaly detection algorithm is used to remove abnormal indicator values to eliminate sudden fluctuations caused by sudden situations. Then, the DBSCAN clustering algorithm is used to calculate the dynamic threshold of the indicators, and a method for quantifying the degree of indicator degradation is proposed. Finally, combined with historical poor quality cell label data, the Lightgbm classification algorithm is used to identify cells with degraded wireless network performance indicators. To verify the effectiveness of this method, the experimental results show that the accuracy of identifying cells with degraded wireless network performance indicators using this method is higher than that of empirical rules and traditional classification algorithms.