Title: Some classical pattern recognition methods: a review

Authors: Zhifen Wang; Weikuan Jia

Addresses: School of Information Science and Engineering, Shandong Normal University, Jinan 250358, China ' School of Information Science and Engineering, Shandong Normal University, Jinan 250358, China

Abstract: Pattern recognition, a fundamental of artificial intelligence, has been successfully applied in many fields. It has broad and far-reaching significance on the development of economy and the progress of science and technology. Firstly, the basic concepts, principles and methods related to pattern recognition are introduced, which contribute to further understanding of pattern recognition. Feature extraction and selection is the key step of pattern recognition, which aims mainly at low loss dimensionality reduction and provides good premise for recognition. Currently, feature reduction has become a hot and difficult issue of pattern recognition, machine learning and data mining. Then, the common methods of pattern recognition are discussed in detail. The key techniques such as algorithms of each method, parameter settings and assessment standards are discussed at a greatest extent. In addition, the newest development of all the methods is also discussed and primary evaluation of effect of the application is made. At last, the development trend and application prospects of pattern recognition are proposed.

Keywords: modern pattern recognition; feature extraction; statistical pattern recognition; fuzzy pattern recognition; bionic pattern recognition; neural network pattern recognition; synergetic pattern recognition.

DOI: 10.1504/IJCI.2021.122704

International Journal of Collaborative Intelligence, 2021 Vol.2 No.3, pp.159 - 190

Received: 11 May 2020
Accepted: 05 Aug 2020
Published online: 06 May 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article