Determination of dependence structure by graphical tools

For bivariate data set, global dependence measures, such as correlation, can not reflect the complex dependence structure of two variables. In order to understand the complex dependence structure of two variables, local dependence measures and particularly graphical methods are being used. In this study, determination of local dependence structure of bivariate distribution is discussed. Three local dependence graphs, called chi-plot, local dependence map and Kendall-plot are mentioned in this study. After the introduction of these graphs; chi-plot and local dependence map are applied to different data sets. In order to draw plots, free software, GNU is used. These data sets contain simulated and real data. With these examples the feasibility of the local dependence graphs is exhibited.