A Study of the Advantages of Clustering and Gray Prediction Algorithms for the Prediction of Data Metrics

In this paper, the clustering algorithm is used for the contraction city classification algorithm. First of all, the big data are dimensionless, and by analyzing the growth rate of each factor, the cities are divided into five categories according to the contraction characteristics of each factor. Through the gray correlation analysis, the weights of different factors affecting the contraction of the city are ranked, and the gray correlation can better solve the problem of uncertainty of influencing factors between multiple indicators, and can supplement the defects of the traditional model. Finally, the gray prediction method is used to predict the future development trend. Compared with the traditional prediction model, the gray prediction model requires less data volume, and only a small amount of incomplete information is needed for model building and prediction. Gray system theory is able to deal with poor information systems, with lower requirements on the smoothness of data, and with a wider scope of application. The gray prediction model adopts gray mathematics to deal with uncertainty, which can make the urban development prediction results closer to reality and improve the prediction accuracy. Finally, curve fitting is used to get the development trend of urban shrinkage.