04 Dec 2022
Back in September
2016 I
implemented the ClusterR
package. One of the algorithms included in ClusterR was the ‘Partition
Around Medoids’
(Cluster_Medoids)
algorithm which was based on the paper “Anja Struyf, Mia Hubert, Peter
J. Rousseeuw, (Feb. 1997), Clustering in an Object-Oriented Environment,
Journal of Statistical Software, Vol 1, Issue 4” (at that time I didn’t
have access to the book of Kaufman and Rousseeuw, Finding Groups in
Data (1990) where the exact algorithm was described), thus I
implemented the code and compared my results with the output of the
cluster::pam()
function,
which was available at that time. Thus, my method was not an exact but
an approximate one. Recently, a user of the ClusterR package opened
an issue mentioning that the results were not
optimal compared to the
cluster::pam() function and this allowed me to go through my code once
again and also to compare my results to new R packages that were not
existent at that time. Most of these R packages include a new version of
the ‘Partition Around Medoids’ algorithm, “Erich Schubert, Peter J.
Rousseeuw,”Faster k-Medoids Clustering: Improving the PAM, CLARA, and
CLARANS Algorithms” 2019, <doi:10.1007/978-3-030-32047-8_16>”.
In this blog-post, I’ll use the following R packages,
to compare between the (as of December 2022) existing ‘Partition Around
Medoids’ implementations in terms of output dissimilarity cost and
elapsed time.
Continue reading...
16 Aug 2022
This blog post is a slight modification of the R package Vignette and
shows how to use the Rcpp Armadillo version of the
fastGLCM R package. The
fastGLCM R package is an RcppArmadillo implementation of the
Python Code for Fast Gray-Level Co-Occurrence Matrix by numpy,
The python version works similarly and is included as an R6 class (see
the documentation of
fastglcm).
However, it requires a python configuration in the user’s operating
system and additionally the installation of the
reticulate R package.
Continue reading...
11 Jun 2022
This blog post is a slight modification of the R package Vignette. The
VMDecomp R package is an RcppArmadillo implementation of the
Matlab Code for Variational Mode Decomposition (1- and 2-dimensional)
based on the papers,
- “Variational Mode Decomposition” by K. Dragomiretskiy and D. Zosso
in IEEE Transactions on Signal Processing, vol. 62, no. 3,
pp. 531-544, Feb.1, 2014, https://doi.org/10.1109/TSP.2013.2288675
- “Two-Dimensional Variational Mode Decomposition” by Dragomiretskiy,
K., Zosso, D. (2015), In: Tai, XC., Bae, E., Chan, T.F., Lysaker, M.
(eds) Energy Minimization Methods in Computer Vision and Pattern
Recognition. EMMCVPR 2015. Lecture Notes in Computer Science,
vol 8932. Springer,
https://dx.doi.org/10.1007/978-3-319-14612-6_15
The Matlab code is available to download in the Author’s website
(https://math.montana.edu/dzosso/code/).
Continue reading...
12 Feb 2022
This blog post (which is a slight modification of both package Vignettes) explains the functionality of the IceSat2R R package and shows how to use the OpenAltimetry API from within R. It consists of three parts,
- OpenAltimetry
- IceSat-2 Mission Orbits
- IceSat-2 Atlas Products
Continue reading...
12 Jun 2021
In this blog post I’ll explain the functionality of the PlanetNICFI R package based on a change detection use case. The official website of NICFI (Norway’s International Climate and Forest Initiative) includes all the details about the initiative against global deforestation. This initiative was also covered extensively on the web especially from the provider of the free Satellite Imagery. Users have the opportunity to download high-resolution imagery of forests globally using a simple sign up form.
To take advantage of the PlanetNICFI R package you will need also an API key which you can receive once you are registered. For more details see the Getting Started with Planet APIs website.
Continue reading...