Estimation of Probability Density Function by Dependence Tree Methods for Pattern Recognition Systems

Estimation of probability density function is inevitable in some engineering systems, specially in statistical pattern recognition systems. One category of methods applied for estimation is tree dependence methods, which could be classified under nonparametric estimation approaches. In this paper, we surveyed important tree dependence methods. To do this, we pursued a mathematically rigorous manner and in this respect, many mathematical details is added to what is available in literature 1 . Furthermore, connection of studied tree dependence methods is addressed completely, which can be a source of valuable information for future works in this area.