Study On Data Mining Technology's Application In Water Resources

It is the time that information technologies keep high speed development in 21st century, and the information resources are important strategic resources. The water resources management is an information intensity profession, so the importance of information resources is very crucial. With the wide application of remote sensing, telemetering, network and database technologies, it powerfully promoted the development of gathering and processing technologies of water resources data. There is much important information in the massive precious water resources data. It has been interested more and more that how to analyze these data timely and effectively, to find valuable information from the data which increases sharply, and to provide important support for the decision-making. The data mining technology are an effective method to solve the problem of "Data rich-Information poor". The data mining technology is a high-level treating process to discover and extract unknown, useful and significant information and pattern from the mass data. We can achieve the transformation from the simple data to the information again to the knowledge by the data mining technology.This article takes Beijing as the study region with its seriously lack of water resources, studying on the data mining technology's application in water resources by SAS data mining tool and providing information support for solving water resources problems. Firstly, by the correlation analysis and regression analysis methods the paper analyzes the rules of water use in various department and establishes the function relation between water use quantity and the major effect factors in each kind of department. This example proves that the data mining technology can discover the correlation between a mass of data, build the model which shows the the factual rules and provide important support for water using management. Secondly, using the SAS data description and exhibition tool the example analyzes the tendency of ET (Evapo Transpiration) of different land utilization types and different month, as well as the influence of hydrometeorological factors such as rainfall, temperature, wind speed, relative humidity etc. acting on ET. We can find from the application that the data mining technology excels at dealing with vast data, which can carry out various statistical analyses and exhibit the results to dicision-maker by plenty graphs. Finally, by the self regression model, the index smooth model and the seasonal model this article separately establishes three forecast models of the groundwater level of unconfined water well, and makes the comparison between the simulation and forecast results. As a whole, the precision ofsimulation and forecast of the seasonal model is higher than others. The time sequence model synthetically reflects trends of groundwater level. Because of simple computation and less demand of data, the time sequence model is a better simulation and forecast model and can supply available basis for resonable groundwater development and using.