Quasi-supervised Learning (QSL)
Quasi-supervised Learning (QSL) computes the posterior probabilities of two classes at the data points in Z. The class labels are stored in the vector Y. k denotes the number of points to be included in the reference set for exhaustive nearest neighbor classifications, and D is the symmetrical matrix of pairwise distances of the points in Z.
library(mvtnorm) n<-1000 x<- rmvnorm(n, c(0,0)) y<- rmvnorm(n, c(0,4)) z <- rbind(x,y) label1 <-rep(0, times = n) label2 <-rep(1, times = n) label<- c(label1,label2) D <- as.matrix(dist(z, method = "euclidean", p = 2, diag = TRUE, upper = TRUE)) N <- 2*n k<- 1 P0<-qsl(z,label,D,N,k)
In order to use qsl function, you need to add kstar.R function into your workspace.
Basak Esin KOKTURK GUZEL
QSL toolbox for MATLAB
If you use the algorithm, please cite:
B. Karacali, "Quasi-supervised learning for biomedical data analysis," Pattern Recognition, Vol. 43, No. 10, p. 3674-3682, 2010