摘要
目的对有异常波动的时间序列资料进行预测。方法采用两阶段预测方法,先对异常波动时点进行预测,然后用该预测值代替原始值正式预测。结果若用2003年受SARS影响的实际收容数进行趋势季节模型预测,得到的预测趋势明显不符合规律,修匀后预测取得理想效果。结论存在异常波动时,应首先对异常点进行处理,然后进行预测。
Objective To predict the data on the points in time with irregular fluctuation. Methods Prediction is made in two phases. In the first phase, we predict the data at the points in time with irregular fluctuation. In the second phase, we make a formal prediction by replacing the original data with the prediction result in the first phase. Results If we predict the season tendency with the actual inpatient number in 2003, which was affected because of SARS, the result did not conform to the law. But after smoothing, we can get a more reliable season tendency. Conclusion When irregular fluctuation happens, first of all, we should deal with the data at the points in time with irregular fluctuation, and then we can make a prediction.
出处
《中国医院统计》
2006年第2期144-145,共2页
Chinese Journal of Hospital Statistics
关键词
时间序列
预测
修匀
收容数
Time series Prediction Smoothing Inpatient number
