Park Ji-yeon of the Department of Construction Systems Engineering, Kookmin University (Convergence Water Resources Engineering Lab, Professor Ju-Young Shin) won the Best Paper Award at the 2024 Korean Society of Civil Engineers Convention.

Park presented a paper titled 'Evaluation of the applicability of the LTSF model of Chungju Dam inflow for long-term time series prediction'. For long-term time series forecasting, there are long-term simulation of physical models or artificial intelligence forecasting techniques through data statistics and pattern recognition.

Existing physical models have limitations that can reduce accuracy due to uncertainties in the characteristics and parameters of the target basin during long-term forecasting. To solve this problem, Park analyzed the efficiency and accuracy of Chungju Dam inflow forecasting using an AI model, the Long-Term Time Series Forecasting (LTSF) model. It was emphasized that the LTSF model shows excellent performance by better extracting long-term trends and periodicity of data. Using these characteristics, the LTSF model was applied to Chungju Dam inflow forecasting to optimize the forecasting performance on long-term time series data.

The key achievements of the study are the improvement of pattern recognition and prediction accuracy of long-term time series data by applying two forms of the LTSF model, NLinear and DLinear, and the in-depth analysis of the technique for long-term forecasting data, and the validation and verification of the prediction performance of the data-based AI model to enhance the validity and applicability of the model.

Park's research is expected to make practical contributions to various fields such as hydrology by improving the prediction accuracy of long-term time series data.