我國宏觀經(jīng)濟統(tǒng)計數(shù)據(jù)質(zhì)量診斷方法與實證

摘要： 文章以我國數(shù)據(jù)質(zhì)量評估現(xiàn)狀作為研究數(shù)據(jù)協(xié)調(diào)性的起點和背景,透過數(shù)據(jù)生產(chǎn)視角并結(jié)合我國實際情況,建立對異常值有較強抵抗性的穩(wěn)健MM回歸模型,并在趨勢擬合診斷法與統(tǒng)計診斷法的基礎(chǔ)上構(gòu)造出綜合診斷法,實證發(fā)現(xiàn)趨勢擬合診斷法與統(tǒng)計診斷法的異常值診斷結(jié)果并不是完全一致的。最后運用綜合診斷法對湖南省歷年地區(qū)生產(chǎn)總值數(shù)據(jù)進行了數(shù)據(jù)協(xié)調(diào)性研究。
This paper takes the reality of China's data quality assessment as the beginning and background of research data coordination, and through the perspective of data production and China's actual situation, constructs a robust MM regression model with strong resistance to outliers. And then the paper develops a comprehensive diagnosis method on the basis of trend fitting diagnosis method and statistical diagnosis method. The empirical study shows that the result of the trend fitting diagnosis method is not consistent with that of statistical diagnosis method. Finally, the paper utilizes the comprehensive diagnosis method to conduct a data coordination research on the GDP data of the past years in Hunan Province.