Air quality prediction method based on deep bidirectional long-short-term memory network

Abstract

The invention discloses an air quality prediction method based on a deep bidirectional long short-term memory network. Aiming at the current situation that air pollutants are difficult to predict, theinvention designs a high-precision prediction algorithm. According to the algorithm, the time series data of historical pollutant indexes is innovatively processed by using the deep bidirectional long short-term memory network, so that prediction and analysis of pollutants are carried out; and the algorithm is built by a PYTHON 3.6.5 Keras tool. The method comprises the following steps: S1, dividing time series data of the plurality of pollutant indexes into a training set, a verification set and a test set; S2, inputting the data of the training set into the deep bidirectional long short-term memory network for training until the network converges; S3, inputting the data of the verification set into the network for verification, adjusting the parameters of the network, and finally obtaining optimal parameters; S4, applying the network to the test set to evaluate the model to obtain a high-accuracy effect; and S5, storing the model and applying the model to the actual situation. The algorithm provides a new solution for air pollutant prediction, and is further widely applied to the field of air pollution prediction.

Description

【Technical field】 [0001] The invention relates to an air quality prediction method based on a deep bidirectional long-short-term memory network, belonging to the field of air pollution prediction. 【Background technique】 [0002] The prediction of the concentration of air pollutants has a strong interdisciplinary nature and has always been a research hotspot in the fields of environment, meteorology, mathematics, geography and computer science. Due to the diversity and complexity of pollution components, pollution indicators often have a highly nonlinear relationship. It is difficult to establish an accurate prediction model with traditional mathematical model methods, and requires a large amount of data collection and analysis of motion mechanisms. The interference of changes increases the difficulty of air pollutant prediction. [0003] Common methods are mainly divided into theoretical methods and statistical methods. Statistics-based prediction methods are getting more ...

AI Technical Summary

Problems solved by technology

Due to the diversity and complexity of pollution components, pollution indicators often have a highly nonlinear relationship. It is difficult to establish an accurate prediction model with traditional mathematical model methods, and requires a large amount of data collection and analysis of motion mechanisms. Varying disturbances increase the difficulty of forecasting air pollutants

Images