Abstract:To solve the rural revitalization of the hollow village resident population prediction problem of the strategic planning, for the state promotes the development of rural, rural construction, rural governance provide auxiliary reference. In this paper, GBDT regression algorithm is used to predict the permanent resident population of hollow village by using electricity and meteorological data. Five characteristics with the strongest correlation between the permanent population of hollow village were screened by eigenvalue importance analysis, and model training and prediction were used to predict the permanent population of hollow village according to these characteristics. In this paper, the data set is divided into training set, cross validation set and prediction set in the ratio of 3:1:1 by using the five-fold cross validation method. After obtaining the prediction results of permanent population, the accuracy of the prediction results was verified by means of mean square error and R square value combined with visualization method. The verification results show that the GBDT regression based permanent resident population prediction algorithm has a good prediction result for the permanent resident population of the hollow village.