非正常波动统计数据修匀方法研究——以福建省消费数据为例

摘要

Statistical data are often subject to the influence of regular or irregular fluctuations caused by abnor-mal factors, thus distorting the fundamental changes in short-term time series and making it difficult to conduct in-depth study and correct interpretation of the economic laws. Provided that scientific methods are used to revise and smooth these statistical data of abnormal fluctuations by determining, separating, offsetting and adjusting these abnormal factors in economic time series can the basic development trends be better reflected. The paper takes for example the indicators of the total retail sales of social consumption in Fujian Province and relative indicators, conducts a comparative analysis before and after revising and smoothing them. It is found that the curve has been smoother and the effect more reasonable. By extrapolation prediction and simulation, dynamic simulation model and static prediction results come out, the dynamic simulation model of the chain CPI emerges, which better reflects the trend of the static forecasts and shows better short-term fluctuations.%统计数据经常会受到定期或不定期的非正常波动因素的影响，因此而扭曲的时间序列短期的基本变动，使得我们难以深入研究和正确解释经济规律。如果利用科学的方法将这些非正常波动因素从经济时间序列中测定、分离、抵消和调整，对这些非正常波动统计数据进行修匀，则能更好地反映其基本的发展趋势。以福建省社会消费零售总额指标和相对指标为例，对其进行了修匀处理和并进行修匀前后的对比分析，发现修匀后的曲线较平滑，修匀效果比较合理。进行外推预测和模拟，得到模型的动态模拟结果以及静态预测结果，得到的环比CPI的动态模拟结果较好地反映了CPI的走势，静态预测较好地显示出短期波动情况。