Title :
Modified Rosenblatt´s perceptron algorithm and Novikoff´s theorem
Author :
Chen, Huang-Chi ; Lin, Yih-Lon ; Sun, Yeong-Jeu ; Hsieh, Jer-Guang
Author_Institution :
Dept. of Electr. Eng., Nat. Sun Yat-Sen Univ., Kaohsiung, Taiwan
Abstract :
A modified Rosenblatt´s perceptron algorithm (1959) and Novikoff´s theorem (1962) for binary classification problem is presented in this paper. For a linearly separable training set, the modified Rosenblatt´s perceptron algorithm could get better convergence result. An example is provided to illustrate our main result.
Keywords :
convergence; pattern classification; perceptrons; Novikoff theorem; binary classification problem; convergence; linearly separable training set; modified Rosenblatt perceptron algorithm; Algorithm design and analysis; Classification algorithms; Convergence; Educational institutions; Management information systems; Management training; Training data;
Conference_Titel :
Industrial Technology, 2002. IEEE ICIT '02. 2002 IEEE International Conference on
Print_ISBN :
0-7803-7657-9
DOI :
10.1109/ICIT.2002.1189361
