Fuzzy perceptron with pocket algorithm in postoperative patient data set

• Main Authors: Phitakwinai,S., Auephanwiriyakul,S., Theera-Umpon,N.
• Format: Conference or Workshop Item
• Published: Institute of Electrical and Electronics Engineers Inc. 2015
• Subjects: Artificial Intelligence; Applied Mathematics; Theoretical Computer Science; Software
• Online Access: http://www.scopus.com/inward/record.url?partnerID=HzOxMe3b&scp=84912573346&origin=inward; http://cmuir.cmu.ac.th/handle/6653943832/39078
• Institution: Chiang Mai University

© 2014 IEEE. Classification is one of the problems in pattern recognition. Most of the time this problem will deal with data sets that are in numeric form and represented by vectors of numbers. Since there might be uncertainties embedded in a data set, it is more natural to represent the data set as fuzzy vectors. Hence, in this paper, we develop a fuzzy perceptron with pocket algorithm for fuzzy vectors. This algorithm is based on the extension principle and the decomposition theorem. We implement this algorithm on both synthetic and a real-world data set, i.e., the postoperative patient data. We also compare the result from the fuzzy perceptron with pocket algorithm with that from the regular perceptron with pocket algorithm. The comparison is done on the fuzzy perceptron with and without pocket as well.