Pattern Discovery in Hydrological Time Series Data Mining during the Monsoon Period of the High Flood Years in Brahmaputra River Basin

paper, based on data mining techniques, the analysis is carried out in hydrological daily discharge time series of the Panchratna station in the river Brahmaputra under Brahmaputra and Barak Basin Organization in India. The data has been selected for the high flood years 1988, 1991,1998, 2004, and 2007. The whole year is divided into three periods known as Pre-monsoon, Monsoon and Post Monsoon. In this paper, only monsoon period data have been used. For standardization of data, statistical analysis such as mean monthly discharge, monthly Maximum Discharge, monthly amplitude and monthly standard deviation have been carried out. K-means clustering is segmented for the monsoon period process of daily discharge. Dynamic Time Warping (DTW) is used to look for similarities in the discharge process under the same climatic condition. Similarity matrix helped in the mining of discharge process in similar time period in the different years. The agglomerative hierarchical clustering is used to cluster and discover the discharge patterns in terms of the autoregressive model. A forecast model has been predicted on the discharge process. In the field of hydrological forecasting, with the help of data mining techniques, earlier various researches have been carried out. Some of them are : Similarity search and pattern discovery in hydrological time series data mining (14), Flood pattern detection using sliding window technique (17), Applications of Data Mining in Hydrology (11), Runoff forecasting using fuzzy support vector regression (23), Forecasting monthly runoff using wavelet neural network model (2), Neural network model for hydrological forecasting based on multivariate phase space reconstruction (22), Mid-short term daily runoff forecasting by ANNs and multiple process based hydrological models (7), Research and application of data mining for runoff forecasting (10), River flow time series using least squares support vector machines(18), a novel approach to the similarity analysis of multivariate time series and its application in hydrological data mining (25), Computational methods for temporal pattern discovery in biomedical genomic databases (16), an efficient k-Means clustering algorithm: analysis and implementation(9), The prediction algorithm based on fuzzy logic using time series data mining methods (3) , A forecast Model of Hydrologic Single, Element Medium and Long-Period Based on Rough Set Theory (19), Application of ANN in Forecast of surface runoff (4) are important contributions in the knowledge discovery from hydrological databases using time series data mining. In this paper, Time Series Data Mining has been used for hydrological study (14). Time series data mining combines the fields of time series and data mining techniques (3). Here, the object is to develop a data mining application using modern information technology to discover the hidden information or patterns behind the historical hydrological data of the river Brahmaputra under the hydrological process, and also to look for a new way to meet the requirements of hydrological time series analysis using TSDM algorithms.