LOGISTIC REGRESSION AS BRIDGE BETWEEN STATISTICS AND ARTIFICIAL INTELLIGENCE

被引：0

作者：

Quang Van Tran ^{[1
]}

Kukal, Jaromir ^{[1
]}

Kalcevova, Jana ^{[1
]}

Bostik, Josef ^{[1
]}

机构：

[1] Fac Nucl Sci & Phys Engn CTU Prague, Dept Software Engn Econ, Trojanova 13, Prague 12000 2, Czech Republic

来源：

16TH INTERNATIONAL CONFERENCE ON SOFT COMPUTING MENDEL 2010 | 2010年

关键词：

Al; ANN pruning; binary model; logit model; probit model; likelihood ratio;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Well known logistic regression and the other binary response models can be used in the area of AI Beginning with basic properties of proposed model, we use likelihood ratio testing as a tool for designing of ANN. We are focused to ANN pruning as a procedure for size reducing of hidden layer, first. The second aim is related to bottom up method of ANN structure generation. Our approach is based on likelihood maximization and binary optimization.

引用

页码：491 / +

页数：2