Title :
Neural CMOS-Integrated Circuit and Its Application to Data Classification
Author :
Goknar, I.C. ; Yildiz, M. ; Minaei, S. ; Deniz, E.
Author_Institution :
Dept. of Electron. & Commun. Eng., Dogus Univ., Istanbul, Turkey
fDate :
5/1/2012 12:00:00 AM
Abstract :
Implementation and new applications of a tunable complementary metal-oxide-semiconductor-integrated circuit (CMOS-IC) of a recently proposed classifier core-cell (CC) are presented and tested with two different datasets. With two algorithms-one based on Fisher´s linear discriminant analysis and the other based on perceptron learning, used to obtain CCs´ tunable parameters-the Haberman and Iris datasets are classified. The parameters so obtained are used for hard-classification of datasets with a neural network structured circuit. Classification performance and coefficient calculation times for both algorithms are given. The CC has 6-ns response time and 1.8-mW power consumption. The fabrication parameters used for the IC are taken from CMOS AMS 0.35-μm technology.
Keywords :
CMOS integrated circuits; data analysis; electronic engineering computing; learning (artificial intelligence); pattern classification; perceptrons; power consumption; CC tunable parameters; CMOS AMS technology; CMOS-IC; Fisher´s linear discriminant analysis; Haberman dataset; Iris dataset; classification performance; classifier core-cell; coefficient calculation times; data classification; fabrication parameters; hard-classification; neural CMOS-integrated circuit; neural network structured circuit; perceptron learning; power consumption; tunable complementary metal-oxide-semiconductor-integrated circuit; Artificial neural networks; CMOS integrated circuits; Classification algorithms; Iris; Learning systems; Programmable logic arrays; Classifier; Fisher; Haberman; Iris; complementary metal–oxide–semiconductor (CMOS);
Journal_Title :
Neural Networks and Learning Systems, IEEE Transactions on
DOI :
10.1109/TNNLS.2012.2188541
