CAJ | 학술논문

为了解决传统指数平滑法在滤波后的数据与原始数据之间存在一个因累积误差而导致右偏移的问题，提出一种基于平衡指数平滑的交通流数据预测方法。该方法对传感器采集的异常数据和缺失数据进行处理，对于超过阈值的数据，则用前一时刻数据取代该数据；对于缺失数据，则插入前后时刻的均值。对处理过的数据采用平衡指数平滑法进行滤波。通过三次指数平滑预测模型对数据进行预测。实验仿真表明，该方法能有效消除传统指数平滑的右偏移误差，提高数据预测的准确性。
위료해결전통지수평활법재려파후적수거여원시수거지간존재일개인루적오차이도치우편이적문제，제출일충기우평형지수평활적교통류수거예측방법。해방법대전감기채집적이상수거화결실수거진행처리，대우초과역치적수거，칙용전일시각수거취대해수거；대우결실수거，칙삽입전후시각적균치。대처리과적수거채용평형지수평활법진행려파。통과삼차지수평활예측모형대수거진행예측。실험방진표명，해방법능유효소제전통지수평활적우편이오차，제고수거예측적준학성。
Proposes the prediction method for traffic flow data based on the balance of exponential smoothing in order to tackle the problem that tra-ditional exponential smoothing exists a right offset between filtered data and original data caused by the cumulative error. Processes the abnormal data and missing data. Replaces the data of exceeding the threshold data by the previous timing data, and inserts the missing data by the mean data of the front and rear moment. Filters the processed data by using balanced exponential smoothing method. Predicts the traffic flow data by using three exponential smoothing. Simulation results indicate that this method can effectively eliminate the right offset error in the traditional exponential smoothing, and improve the accuracy of prediction data.