Performance Prediction of Medical Appointment System Using Improved SVM Based on Loadrunner

In high concurrency environments, problems such as slow system response speed and poor stability often occur, seriously affecting the patient's medical experience and the hospital's operational efficiency. Therefore, this study takes the medical appointment system as the research object, designs a performance testing plan for the medical appointment system based on the LoadRunner, and collects the system response time under different states. Next, an improved SVM prediction model is proposed to achieve accurate estimation of response time based on four system parameters: the number of time period users, concurrent users, resource utilization, and data volume. The experimental results show that the proposed prediction method has a high correlation coefficient of 0.9979 on the test set, a low MBE of 0.0665 s, an absolute error (MAE) of 0.1391 s, and a RMSE of 0.2257 s. The overall testing performance is better than other methods. The response time prediction model proposed in this article is helpful for predicting the performance of the medical system in various situations, identifying performance bottlenecks in a timely manner, and ensuring the stable operation of the system.