This is an implementation for the following paper:
@article{ALSHAMMARI201931,
title = {Approximate spectral clustering with eigenvector selection and self-tuned k},
doi = {https://doi.org/10.1016/j.patrec.2019.02.006},
journal = {Pattern Recognition Letters},
volume = {122},
pages = {31-37},
year = {2019},
author = {Mashaan Alshammari and Masahiro Takatsuka}
}
Run BATCH_Points.m
which will execute the following:
PRE_Points.m
to load toy data, csv files are the groundtruth labels.RUN_Points.m
to perform spectral clustering with 4 functions to estimate k:CostEigenGap.m
a conventional method to estimate kCostZelnik.m
uses the method proposed by (Zelnik-manor 2005) to estimate kCostDBIOverLambda.m
uses the method proposed by our paper to estimate kCostDBIOverLambdaPCA.m
a uses the method proposed by our paper to estimate k followed by PCA variance filtering
POST_Points.m
to compute the accuracy of clustering