Method and system for realizing change rule analysis of atmospheric components based on machine learning

The invention relates to the field of atmospheric composition change analysis, and discloses an atmospheric composition change rule analysis method and system, and the method comprises the steps: obtaining the historical atmospheric data of a to-be-analyzed region and the historical geographic data of the to-be-analyzed region; querying variable data in the historical geographic data, identifying a variable time node of the variable data, and extracting gas component data in the historical atmosphere data corresponding to the variable time node; inputting the variable data and the gas component data into a pre-constructed atmospheric analysis model, and combining the variable data and the gas component data by using a data combination module in the atmospheric analysis model to obtain combined data; sampling the combined data to obtain sampling data; constructing a data scatter diagram of the sampled data; analyzing scatter correlation coefficients of scatters in the scatter diagram; and constructing an atmos