LDA-LSTM based on variational entropy weight and multi-dimensional conditional grey association

This paper studies the LDA-LSTM model based on variational entropy weight and multidimensional conditional grey relation, and applies it to the actual case of urban traffic congestion in China. Firstly, the paper collects traffic data from 20 cities in China, determines the weight of indicators through variational entropy weight, and innovatively integrates the multidimensional conditional grey relation model to identify the key indicators affecting urban traffic. Then, the paper proposes the LDA-LSTM hybrid model to perform topic analysis on the data, extracts 8 topic words as indicators for LDA-LSTM, and predicts the traffic congestion situation in the first quarter of 2024 in cities, finding that the prediction values fit well and overall achieve good prediction results.