Hybrid Fuzzy Classification Algorithm with Modifed Initialization and Crossover

Tatiana Pleshkova, Vladimir Stanovov

2022

Abstract

The article proposes two modifications of initialization and crossover operations for the design of a genetic fuzzy system. A fuzzy logic system is used to solve data classification problems and is automatically generated by a genetic algorithm. The paper uses a genetic algorithm to encode of several fuzzy granulations into a single rule, while each individual encodes a rule base. The proposed algorithm uses several training objects of the same class to create a single rule during initialization. The modified crossover creates a new rule base from most efficient rules selected from parents. To evaluate the effectiveness of the modification, the computational experiments were carried out on several datasets, followed by verification using Mann-Whitney U test. The proposed initialization modification allows reducing the number of rules in a fuzzy rule base and increasing the accuracy and F-score on some datasets. The crossover modification shows higher efficiency only on one dataset.

Paper Citation

in Harvard Style

Pleshkova T. and Stanovov V. (2022). Hybrid Fuzzy Classification Algorithm with Modifed Initialization and Crossover. In Proceedings of the 14th International Joint Conference on Computational Intelligence (IJCCI 2022) - Volume 1: FCTA; ISBN 978-989-758-611-8, SciTePress, pages 225-231. DOI: 10.5220/0011587500003332

in Bibtex Style

@conference{fcta22,
author={Tatiana Pleshkova and Vladimir Stanovov},
title={Hybrid Fuzzy Classification Algorithm with Modifed Initialization and Crossover},
booktitle={Proceedings of the 14th International Joint Conference on Computational Intelligence (IJCCI 2022) - Volume 1: FCTA},
year={2022},
pages={225-231},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011587500003332},
isbn={978-989-758-611-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Joint Conference on Computational Intelligence (IJCCI 2022) - Volume 1: FCTA
TI - Hybrid Fuzzy Classification Algorithm with Modifed Initialization and Crossover
SN - 978-989-758-611-8
AU - Pleshkova T.
AU - Stanovov V.
PY - 2022
SP - 225
EP - 231
DO - 10.5220/0011587500003332
PB - SciTePress