Research Article
Research on Prediction Algorithm of College Entrance Examination Filing Line Based on ARIMA and LSTM
@INPROCEEDINGS{10.4108/eai.29-3-2024.2347733, author={Yifan Yan and Zuxu Dai}, title={Research on Prediction Algorithm of College Entrance Examination Filing Line Based on ARIMA and LSTM}, proceedings={Proceedings of the 3rd International Conference on Educational Innovation and Multimedia Technology, EIMT 2024, March 29--31, 2024, Wuhan, China}, publisher={EAI}, proceedings_a={EIMT}, year={2024}, month={6}, keywords={filing line; arima model; lstm model; arima-lstm combined model}, doi={10.4108/eai.29-3-2024.2347733} }- Yifan Yan
Zuxu Dai
Year: 2024
Research on Prediction Algorithm of College Entrance Examination Filing Line Based on ARIMA and LSTM
EIMT
EAI
DOI: 10.4108/eai.29-3-2024.2347733
Abstract
To improve the prediction accuracy of college entrance filing line, this study uses ARIMA-LSTM combined model to predict the rank of college entrance filing line based on score-to-rank conversion table. The model forecasts the rank of filing line for colleges, upon which the admission filing line is predicted. The ARIMA model is utilized to analyze linear relationships in the data, and its autoregressive coefficients set the time steps for the LSTM model, which addresses the nonlinear aspects of the forecast. The predictive results of the combined model are compared with those of the standalone ARIMA and LSTM models. The experimental results show that at the 90 % confidence level, the prediction error confidence interval of the ARIMA-LSTM combined model is (0.2, 3.6), which surpasses the ARIMA model's interval of (3.5, 6.6) and the LSTM model's interval of (-6.3, -2.7). This demonstrates the combined model's efficiency and accuracy in forecasting college entrance filing line.
