CAJ | 학술논문

时间序列的预测有着广泛的应用，如何提高预测的准确程度是人们关心的问题。将海水位数据作为研究对象，对多组不同长度的样本，采用AR模型预测，将预测误差与样本长度的关系进行定量分析。另外，实验得到的预测误差与样本长度的函数关系呈现出一定的幂函数特征，这有助于提高时间序列预测的准确程度。
시간서렬적예측유착엄범적응용，여하제고예측적준학정도시인문관심적문제。장해수위수거작위연구대상，대다조불동장도적양본，채용AR모형예측，장예측오차여양본장도적관계진행정량분석。령외，실험득도적예측오차여양본장도적함수관계정현출일정적멱함수특정，저유조우제고시간서렬예측적준학정도。
Time series prediction has a widespread application. How to improve the accuracy of the prediction is concerned. Taking the sea level data as the study object,the AR model is used to predict the multigroup samples with different length,and the relationship between forecasting errors and sample length is quantitatively analyzed. The function relationship between them presents some characteristics of power function,which will contribute to improving the accuracy of time series prediction.