Prediction method of urban PM10 concentration based on characteristic expansion fusion neural network

Abstract

The invention relates to a prediction method for urban PM10 concentration based on a characteristic expansion fusion neural network, which comprises the following steps of: step S1: constructing a model of the concentration prediction of the urban PM10 based on the feature-expanded stack self-encoder and LSTM network; step S2: selecting training data and test data from monitoring data of pollutants and meteorology; step S3: carrying out the training on the feature-expanded stack self-encoder by using the training data; step S4: processing the feature vectors of the output of the stack self-encoder based on the Gaussian function, calculating the corresponding influence weights for the feature vectors of different cities, and weighting and summing to obtain a new feature vector; step S5: inputting the new feature vector into the LSTM to carry out integral training of the model; step S6: inputting test data into a trained model, and measuring the error of the prediction result generated by the test data; step S7: adopting the training and fine-tuning models for air pollutant concentration prediction. Compared with the prior art, the invention has the advantages of accurate predictionand the like.

Description

technical field [0001] The invention relates to a method for predicting PM10 concentration, in particular to a method for predicting urban PM10 concentration based on a fusion neural network of feature expansion. Background technique [0002] Air pollution is a problem that has been widely concerned in daily life, and with the increasingly serious problem of air pollution, the types of air pollutants are becoming more and more diverse, the formation and diffusion of air pollutants are becoming more and more complicated, and the prediction of pollutant concentration is no longer necessary. It is single-point, but dynamic and regionally linked. Therefore, in the current situation, in order to make more accurate predictions of pollutant concentration, prevent the occurrence of heavy pollution incidents, and improve the level of environmental management and decision-making, we should make full use of the monitored pollutants and meteorological big data, and fully mine and learn ...

AI Technical Summary

Problems solved by technology

These methods have played great advantages in the past when the environmental situation was simple and the data scale was small. However, in the case of massive pollutant data and meteorological data, these methods lack in-depth analysis of data characteristics and cannot fully learn At the same time, these methods regard the change of pollutant concentration as a discrete event, and do not consider and cannot perform correlation analysis in time and space, so that accurate pollutant concentration prediction cannot be performed

[0004] On the other hand, considering that it is difficult to obtain a complete data set for the prediction of air pollutant concentration, most of them have missing features and insufficient feature dimensions, which makes the prediction model unable to fully learn data features and mine the relationship between data, so that it cannot make accurate predictions

At present, there are very few technical researches on expanding the feature dimension in the prediction model in the academic circle, but this technique is very effective in dealing with the problem of insufficient feature dimension

Images