Application and Practice of Data Mining Algorithms in Power Evaluation System

With the continuous expansion of the scale of power systems and the increasing complexity of the operating environment, how to efficiently and scientifically evaluate the performance and operating status of power systems has become an important issue that the power industry needs to solve. To meet this challenge, this paper builds a power evaluation system based on data mining technology, and improves the evaluation accuracy and efficiency of the system by applying a variety of data mining algorithms. Firstly, the clustering algorithm is used to classify the historical power data to identify the characteristics of different operating states. Then, the decision tree method is applied to establish a classification model for power performance indicators to reveal the key influencing factors. Finally, the association rule mining technology is used to explore the relationship between various variables in the power operation process and to mine potential rules and correlations. The experimental results show that when the number of concurrent users is less than 150, the system can maintain high processing performance and response speed, with an average response time of less than 3 seconds and a transaction processing volume (Transaction Per Second, TPS) of more than 400; but when the number of concurrent users reaches more than 200, the system’s CPU usage is close to full load, the response time of data entry requests increases significantly, reaching 4.5 seconds, TPS drops to 300, and the business failure rate reaches 5%. These results show that the system’s performance bottleneck under high concurrent load, especially the responsiveness of data entry operations, needs to be optimized to improve the overall concurrent processing capability and stability of the system.