数据挖掘在河道洪水准确预测中的应用研究

Research the Application of Data Mining in the River Flood Accurate Prediction

研究河道洪水的准确预测问题。由于水文数据是河道过去某些较短时间段内获取的数据,不能完整包含河道特性,使得较短的水文数据中的预报因子较难提取。传统的预测方法是直接提取水文数据中的预报因子,不能保证预报因子的准确度而造成预测准确性不高。为此,提出数据挖掘技术应用在河道洪水预测中。对水文数据进行分组后根据模糊算法对数据进行优化,采用数据挖掘算法找到水文数据隐藏的深层规律,并据此提取出预报因子,避免直接从较短的水文数据中提取预报因子而不准确的问题,最终根据关联规则构建洪水预测模型,并输入预报因子和降水值完成洪水预测。实验表明,这种方法能够从较短水文数据中有效提取预报因子,准确完成河道的洪水预测。

Research the accurate prediction of the river flood. Because the hydrological data is obtained from some short river data within the past time period, short hydrological data can＇t complete include river characteristics, making the relatively the forecast factor is hard to extract from short hydrological data. The traditional prediction method is direct extraction of the hydrological data predictor, cannot guarantee the accuracy of prediction factors, causing the prediction accuracy is not high. In order to solve this problem, this paper put forward the data mining technology to apply in river flood forecast. The hydrological data is grouped and then optimized according to fuzzy algorithm, using data mining algorithm to find the hydrological data hidden deep rule which are used to extract the forecast factor, avoiding the low accurate predictor problem of direct extraction from a relatively short period of the hydrological data. Finally, according to the flood forecasting model construction association rules, and input forecast factor and precipitation value to complete the flood forecast. Experimental results show that this method can from short of the hydrological data extracted effectively forecast factor, the flood forecast accurately completing river.