기상데이터를 결합한 경부고속도로 시계열 교통량 예측 모델링 = Time Series-Based Traffic Volume Forecasting on the Gyeongbu Expressway with Weather Data

다국어 초록 (Multilingual Abstract)

As road traffic volume has steadily increased, the need for efficient traffic management and accurate prediction, particularly on expressways, has become increasingly critical. However, existing traffic prediction studies are often limited by their reliance on historical traffic data, failing to adequately consider external factors such as weather conditions. To address this limitation, this study integrates meteorological variables with traffic data and evaluates the performance of the Temporal Fusion Transformer (TFT) model. Experimental results indicate that incorporating weather data improves prediction accuracy across all models, with the TFT model outperforming LSTM. Specifically, the TFT model achieved a reduction in MAPE compared to LSTM by 4.27 percentage points in Suwon and 1.44 percentage points in Daegu. These findings demonstrate the effectiveness of integrating weather data in traffic volume prediction and suggest the potential for future research to incorporate additional external factors.