The Application Of Rules Generator On Air Quality Prediction Based On Cloud Model

Since the20th century, and uncertainty of scientific value and practical significance gradually accepted by people. Though the application of probability theory, fuzzy set theory and rough set theory which developed in recent years by the scholars, making uncertain problems research made a lot of progress, its research results were also widely used in various fields. However, with the development of research, the limitation of these methods also gradually emerges. Chinese scholars LiDeYi creatively put forward a new model--cloud model to solve uncertainty on the basis of traditional probability statistics and fuzzy set theory. Cloud model is used to solve the traditional method of existing incompleteness, and effectively realized the uncertainties of qualitative description to quantitative description of the conversion. Since cloud model was proposed it is widely used in the qualitative knowledge, express, uncertainty classification, association knowledge digging, time series prediction and qualitative control and so on.It produced the method of using cloud model to predict air quality based on the previous research achievements, using the air quality daily data of chuzhou city in2008. Due to the influence factors on the quality of the air with uncertainty, this uncertainty comes from each aspect, such as weather conditions, human activities will influence on air quality Meanwhile, the monitoring data has also has a certain randomness, including monitoring accuracy and statistical method of selection, etc. Therefore, in the air quality forecast, we must take into account the forecast of appeared in the course of uncertain factors. This study will try to using cloud model to predict air quality, hope a penetration both in theory and method.The author has made certain progress during the last half year. First analyzed the original data to get primary pollutants then used stepwise regression to make sure the correctness of the factors. Establish rules according to China’s urban air quality classification standard, constructing rules generator to forecast. To test the effect of prediction of cloud model, the article selected randomly the historical data of primary pollutants as the input value of the generator rules, compared the output results with the actual results, it found that the prediction level conformed to the actual level, and API values are also the same, the forecast results are ideal.For air quality forecast, it had many research methods, such as neural networks, grey prediction method, linear regression method, etc. Overall, these methods have obtained results in a certain degree, the prediction effect can be reliable, but due to the limitations of theories themselves, research with these methods can hardly get breakthroughs. The author through extensive reading literature, based on previous studies to find another angle to forecast air quality, seizes an important and easiest uncared point: uncertainty as the point cut. Using cloud model in this field has achieved satisfactory results. This paper only discuss the field of air quality forecast, cloud model can also be used for other areas according to various industries, such as agriculture, industry, medicine industry etc.The research of this article in air quality prediction based on cloud model is made some achievements, but this research is preliminary, cloud model application value remains to be more in-depth mining.