estimates a univariate density of a random variable that is convoluted with a Gaussian random variable. The estimation is based on deconvolution by inverting the characteristic function.
When the program's counter comes inside a switch construct to the keyword default, the following commands are processed in any case. A default statement can be finished by the keyword break, but does not have to.
This is a variant of denxbwcrit and denbwcrit using linear binning for fast computation. All kernel estimates of the integrated squared densities or density derivatives are computed approximately by Rdenbest which uses linear binning.
determines from a range of bandwidths the optimal one using one of the following bandwidth selection criteria: Least Squares Cross Validation (lscv), Biased Cross Validation (bcv), Smoothed Cross Validation (scv), Jones, Marron and Park Cross Validation (jmp), Park and Marron Plug-in (pm), Sheather
transforms density data to regression data using variance stabilizing transform. Divides the sample space into bins, calculates the counts y_i of observations from every bin, and gives the values 2*sqrt(y_i+3/8) as a regression variable.
given a binary tree produced by cartsplit, normalizes the mean values of the leaves so that the function represented by the binary tree integrates to one.
provides extensive descriptive statistics for the columns of a data matrix. An additional vector of name strings can be given to identify columns by names.
Reduces a matrix to its distinct rows and gives the number of replications of each row in the original dataset. An optional second matrix y can be given, the rows of y are summed up accordingly.
plots a coxcomb graph, a special pie chart, for which the frequency is proportional to the area of the corresponding segment and the angles of the segments are all equal. Consequently, the frequency is proportional to the square of the radius of the segment.
estimates a dynamic semiparametric factor model from the form: yt = m0(u) + bt1*m1(u) + bt2*m2(u)... btL*mL(u), where m0 to mL are 2-dimensional invariant basis functions on the grid u and bt0=1. bt1 to btL are scalar weights depending on time T. After estimation, the functions m are orthogonalized
estimates a dynamic semiparametric factor model from the form: yt = m0(u) + bt1*m1(u) + bt2*m2(u)... btL*mL(u), where m0 to mL are 2-dimensional invariant basis functions on the grid u and bt0=1. bt1 to btL are scalar weights depending on time T. After estimation, the functions m are orthogonalized