Design And Implementation Of Massive Atmosphere Particulates Components Data Mining Analysis System

Composition of atmospheric particulate has a big impact on human health and the environment, visibility, urban transportation and global atmospheric environmental issues, especially in recent years with the growing air pollution in our country,more and more people have payed more attention on this problem.The traditional means of atmosphere particulate matter analysis relies mainly on technology and artificial methods to identify the types and sources of particulate matter overall analysis, but these tools is time-consuming, high labor costs, low accuracy, there is no way to meet the current needs of the people. The system relies on a spectrometer atmospheric particulate matter can be online real-time analysis and come up with a set of automatic and efficient processing and analysis system of atmospheric particulate matter data.The spectrometer has a powerful data collection capability,it could collect vast amounts of atmospheric particulate matter everyday.Through analysis and research,we use a clustering algorithm with similar properties to a particle aggregation packet,then the subsequent processing can be performed in units of packets,achieve real-time requirements of the system.In order to achieve the system automatically requirement,we used logistic regression algorithm to automatically name the particulate matter packet.By artificial naming we trained the sample training model parameters,making the model can classify the test samples,reduced the manual intervention in the system.After the particulates naming procedure,we could do the analysis of the atmosphere components,understand the quality detection time, detection of location comprehensively.