Predicting the Development Trend of University Sports Meeting Results Using LSTM Model

Result prediction is valuable in sports because it can help athletes improve performance and support the optimization of training efficiency. Similar time-series prediction methods are also important in engineering fields such as electromagnetic testing, antenna performance evaluation, and communication system monitoring, where trend forecasting can support resource allocation and process adjustment. As an important component of quality-oriented education, university sports meetings provide measurable data for evaluating students’ physical fitness and improving physical education programs. Changes in sports meeting results can reflect long-term training effects and provide data support for teaching reform and training plan optimization. Long Short-Term Memory (LSTM), as a representative recurrent neural network, has advantages in processing long-sequence data and capturing nonlinear temporal dependencies. This study takes university sports meeting results as the research object and constructs an LSTM-based prediction model for result development trends. Experimental results show that the proposed model achieves higher prediction accuracy across different events than the ARIMA model and BP neural network. By analyzing the development trend of university sports meeting results, the model provides a scientific basis for optimizing training plans and event organization. The study also demonstrates the applicability of LSTM-based time-series prediction to engineering data analysis, including electromagnetic measurement sequences and antenna performance trend prediction.