Merge pull request #69 from zStupan/update-paper

Updated paper

paper/paper.bib

@@ -181,3 +181,14 @@ @misc{fister_uarmsolver_2020
 	note = {arXiv:2010.10884 [cs]},
 	keywords = {Computer Science - Databases, Computer Science - Neural and Evolutionary Computing},
 }
+
+@article{alcala2009keel,
+  title={KEEL: a software tool to assess evolutionary algorithms for data mining problems},
+  author={Alcal{\'a}-Fdez, Jes{\'u}s and Sanchez, Luciano and Garcia, Salvador and del Jesus, Maria Jose and Ventura, Sebastian and Garrell, Josep Maria and Otero, Jos{\'e} and Romero, Crist{\'o}bal and Bacardit, Jaume and Rivas, Victor M and others},
+  journal={Soft Computing},
+  volume={13},
+  number={3},
+  pages={307--318},
+  year={2009},
+  publisher={Springer}
+}

paper/paper.md

@@ -71,6 +71,6 @@ of the NiaARM framework are:
 
 5. A simple command-line interface for easier handling with the proposed tool.
 
-To the authors' knowledge, NiaARM is one of only two publicly available software solutions that implement any form of numerical association rule mining, the other being uARMSolver [@fister_uarmsolver_2020]. The uARMSolver framework, written in C++, also implements the ARM-DE algorithm. Comparatively, NiaARM offers better ease of use, the ability to use arbitrary nature-inspired algorithms from the NiaPy framework (uARMSolver only implements DE or PSO), and the ability to optimize using more interest measures.
+To the authors' knowledge, NiaARM is one of only three publicly available software solutions that implement any form of numerical association rule mining, the other two being KEEL [@alcala2009keel] and uARMSolver [@fister_uarmsolver_2020]. KEEL is a software tool used to assess evolutionary algorithms for machine learning problems of various kinds such as regression, classification, unsupervised learning, etc. It's a GUI application written in Java primarily intended for research and educational purposes. Although its scope is much wider it also includes some popular algorithms for numerical association rule mining including GAR, GENAR and MODENAR. The uARMSolver framework, written in C++, also implements the ARM-DE algorithm. Comparatively, NiaARM offers better ease of use, the ability to use arbitrary nature-inspired algorithms from the NiaPy framework (uARMSolver only implements DE and PSO), and the ability to optimize using more interest measures.
 
 # References