| EUSFLAT- LFA 2005 Fitting Piecewise Linear Copulas to Data (2009) | |||||||||||||||
Abstract | |||||||||||||||
| Given a set of n data points (xi, yi, zi) ∈ [0, 1] 3, two methods are established for constructing two piecewise linear copulas, respectively denoted CA(x, y) and CB(x, y), in such a way that CA(xi, yi) and CB(xi, yi) both yield an approximation of zi for all i. If the data points are (randomly) sampled from an unknown copula C, i.e. zi = C(xi, yi), then CA and CB provide two piecewise linear estimates of C. The quality of both estimations is illustrated on synthetic randomized data sets. | |||||||||||||||
Details der Publikation | |||||||||||||||
| |||||||||||||||