A Neural Network Prediction Method of Air Quality Based on Decision Tree Index

Abstract

The invention relates to a method for predicting air quality based on a neural network index of a decision tree, comprising the following steps: establishing a time series data set of relevant meteorological factors, air quality and air pollutant emissions; using a decision tree DT algorithm to train the obtained training The samples are classified to generate the optimal tree structure T oriented by air quality characteristics α and the corresponding classification results; according to the classification results, establish a BP neural network model for each classification, and carry out model training; input the prediction data set, carry out classification index based on the decision tree, and select the DT-BP neural network model after training Or integrated BP neural network to predict air quality; obtain continuous air quality prediction results based on iterative algorithm; record the number of data sets that do not meet the classification and matching rules of decision tree, and automatically start model update if the set value is exceeded. The invention is suitable for air quality prediction and forecast of normal weather, sudden change weather and heavy pollution weather.

Description

technical field [0001] The invention belongs to the technical field of data processing, and relates to a method for forecasting air quality suitable for conventional weather, sudden change weather and heavily polluted weather, in particular to a method for predicting air quality based on a neural network based on a decision tree index. Background technique [0002] With the rapid growth of our country's economy and the continuous development of urbanization, the problem of environmental pollution has increasingly seriously affected the space on which people live, and even caused major vicious accidents, which greatly endangered people's health and production and construction. For a long time, researchers have conducted comprehensive and systematic research on the change characteristics and trend forecast of regional ambient air quality. However, since air pollution is affected by various factors such as weather background, topography, transportation and convergence, and the ...

AI Technical Summary

Problems solved by technology

[0020] (2) When using the trained model for forecasting, it is necessary to use the data of the last L days to do an error test on the forecasting model of each decision tree, and then determine which tree's forecasting data to choose. The steps are relatively complicated

[0021] To sum up, the existing air quality forecasting methods and systems all have limitations in data interval identification, and do not give full play to the advantages of various statistical algorithms in the identification and capture of air quality change characteristics. The ability to identify and report air quality inflection points is low, which is far from meeting the needs of the public to provide health guidance

Images