Research on air quality prediction based on improved long short-term memory network algorithm

Abstract

Air quality is changing due to the influence of industry, agriculture, people’s living activities and other factors. Traditional machine learning methods generally do not consider the time series of the data itself and cannot handle long-range dependencies, thus ignoring information relevant to the predicted items and affecting the accuracy of air quality predictions. Therefore, an attention mechanism is introduced based on the long short term memory network model (LSTM), which attenuates unimportant information by controlling the proportion of the weight distribution. Finally, an integrated lightGBM+LSTM-attention model was constructed based on the light gradient boosting machine (lightGBM), and the prediction results were compared with those of 11 models. The experimental results show that the integrated model constructed in this article performs better, with the coefficient of determination (R2) of prediction accuracy reaching 0.969 and the root mean square error (RMSE) improving by 5.09, 4.94, 4.85 and 4.0 respectively compared to other models, verifying the superiority of the model.

High-Precision Microscale Particulate Matter Prediction in Diverse Environments Using a Long Short-Term Memory Neural Network and Street View Imagery

“…This novel gating system enhanced information discernment and enabled effective simulation of multiple inputs. As supported by research studies, , the LSTM model exhibited better performance in processing time-series problems. Furthermore, these findings indicated that the LSTM model is better equipped to capture the complex spatiotemporal variability associated with various PM metrics …”

“…Menares et al (2021) employed both a deep feedforward neural network (DFFNN) and LSTM models to forecast the PM 2.5 concentration value, with the LSTM model outperforming the DFFNN model and achieving the highest accuracy of 0.87 . While these models enhance the precision of predictions to varying extents, they face a common limitation: as the length of the time series grows, data points that are further away from the present information become increasingly disregarded . This can lead to a loss of valuable information that may be crucial for accurate forecasting.…”

“… 4 While these models enhance the precision of predictions to varying extents, they face a common limitation: as the length of the time series grows, data points that are further away from the present information become increasingly disregarded. 16 This can lead to a loss of valuable information that may be crucial for accurate forecasting. However, street view images obtained within urban street canyons provide a human-centric perspective of the built environment, offering a promising avenue to access information about the surrounding environment.…”

Computer Simulation and Innovation of Virtual Reality Technology in Film and Television Animation Scene Construction

“…Reverse modelling technique, based on multi-sensor measurement, has steadily emerged as a new solution with development of CAD film as well as television scenes. A thorough overview of the fundamentals, benefits, and uses of this technology in the creation of motion pictures and television shows was given by author [15]. Reverse engineering techniques are used in CAD film as well as television scene reverse modelling based on multisensor measurement, which transforms sensor data into a 3D CAD model.…”