Regional Development Stage Division Method Based on LSTM Model

Abstract

Due to the differences in resource endowments and development foundations, the development trends of regions are different. As the electricity transmitted by the power grid continues to increase, accurately predicting the development trend of regional power grids under limited data and dividing the development stages of power grids can effectively identify the needs of power grid construction and investment in different regions. To accurately identify regional development potential and optimize power grid decision-making, a regional development stage division method based on long short-term memory (LSTM) model is proposed. Firstly, the historical data of power grid development in each region is collected. Secondly, the data is divided into test set and training set, and input to the LSTM unit for data prediction to form an LSTM development prediction model. Taking the historical development data of power grids in different regions as input, the LSTM model is used to predict the development data and form the regional development curve. The principal component analysis method is used to reduce the dimensionality of the high-dimensional development data. The Logistic growth curve model is used to fit the development data and divide the development stages of the regional power grid. Finally, three typical cities are used as practical examples to analyze the applicability of the method.