The Evaluation Of The Reliability Of Data In GDP With Robust Principal Component Regression Method

In recent years，with the rapid development of economy in our country,the GDP have grown highly for several years. Many scholars doubt aboutthe China’s economic growth data and the reliability of some official figuresreleased by the department of statistics was also doubted by socialcitizens.China’s economic growth data covers many aspects in thedevelopment of national economy.In order to estimate the reliability ofChina’s statistical data, we need to assess the reliability of data on economicgrowth first. This paper adopts a robust statistical method--robust principalcomponent regression method to evaluate the reliability of the data onChina’s economic growth. Firstly the paper compares the robust principalcomponent regression method with the classical principal componentregression method in the ability to resist outliers and the diagnosis results ofabnormal data. According to the outcome of comparison of two method, therobust principal component regression method can obtain more dependableresults and identify outliers better. Secondly we make the study of regionalgrowth data in three years.The result shows the regional growth data inChina is consistent with the data of ten indicators which is closely relatedwith the regional growth data.and further more we diagnose the reliabilityof regional growth data and the results show that most of the region’seconomic growth in China is reliable. Growth figures of little areas werediagnosed as abnormal point and a reliability problems may be exist in these areas. In order to needs to determine whether the region’s economic growthdata is reliable, we need to consider the background of abnormal point andits regional particularity.