A cosine adaptive particle swarm optimization based long-short term memory method for urban green area prediction

Abstract

In the quest for sustainable urban development, precise quantification of urban green space is paramount. This research delineates the implementation of a Cosine Adaptive Particle Swarm Optimization Long Short-Term Memory (CAPSO-LSTM) model, utilizing a comprehensive dataset from Beijing (1998-2021) to train and test the model. The CAPSO-LSTM model, which integrates a cosine adaptive mechanism into particle swarm optimization, advances the optimization of long short-term memory (LSTM) network hyperparameters. Comparative analyses are conducted against conventional LSTM and Partical Swarm Optimization (PSO)-LSTM frameworks, employing mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE) as evaluative benchmarks. The findings indicate that the CAPSO-LSTM model exhibits a substantial improvement in prediction accuracy over the LSTM model, manifesting as a 66.33% decrease in MAE, a 73.78% decrease in RMSE, and a 57.14% decrease in MAPE. Similarly, when compared to the PSO-LSTM model, the CAPSO-LSTM model demonstrates a 58.36% decrease in MAE, a 65.39% decrease in RMSE, and a 50% decrease in MAPE. These results underscore the efficacy of the CAPSO-LSTM model in enhancing urban green space area prediction, suggesting its significant potential for aiding urban planning and environmental policy formulation.

Keywords: Cosine adaptive; Long series prediction; Long-short term memory; Particle swarm optimization; Urban green area.

Figure 1. Architecture diagram LSTM.

The image depicts an LSTM (long short-term memory) neural network architecture, illustrating the flow and transformation of data within. It shows the internal gating mechanisms—forget, input, and output gates—of an LSTM cell, how they process the input X, and generate an output Y. The LSTM layer connects to a fully connected layer that integrates the features, leading to the final output layer where the result is produced.

Figure 2. Flowchart of CAPSO-LSTM.

The process begins with normalizing the input data, followed by initializing the particle velocity and position. It then calculates the particle fitness and checks if the predefined number of iterations has been reached. If not, it updates the individual and global optima using CAPSO. This loop continues until the iteration condition is met. Once completed, the process outputs the optimal parameters for the LSTM model, which are then used to predict the green area. The flow is sequential and iterative, with a decision point that loops back until the stopping criterion is satisfied.

Figure 3. Train and prediction results of LSTM.

The black line indicates the predicted values generated by an LSTM model, while the green line represents the actual observed values.

Figure 4. Train and prediction result of PSO-LSTM.

The black line indicates the predicted values generated by an PSO-LSTM model, while the orange line represents the actual observed values.

Figure 5. Train and prediction result of CAPSO-LSTM.

The black line indicates the predicted values generated by an CAPSO-LSTM model, while the purple line represents the actual observed values.