Абстрактный

Exploratory data analysis of a large groundwater quality dataset of Bhavnagar city by using Kohonen's self organizing map

J.Jani, Shriji Kurup, Ratna Trivedi, P.M.Chatrabhuji, P.N.Bhatt

In groundwater quality research, researchers are often confronted with large, multidimensional datasets. An exploratory data analysis is often carried out, aiming at summarizing the available data, extracting useful information and formulating hypothesis for further research. The techniques traditionally used in this phase of research are basic statistics (i.e. mean, median, standard deviation) along with graphical techniques (i.e. scatter plots, histograms, time series plots).Alsomultivariate statistical techniques, like principal component analysis or discriminate analysis are used to explore multidimensional datasets. KohonenÂ’s Self-Organizing Map is an artificial neural network technique, developed by Kohonen in the early Â’80s, especially designed for 2Dvisualizing and analyzing large,multidimensional datasets[3].