Application of the Fuzzy Logic to Evaluation and Selection of Attribute Ranges in Machine Learning

被引：2

作者：

Paja, Wieslaw ^{[1
]}

Pancerz, Krzysztof ^{[2
]}

Pekala, Barbara ^{[3
]}

Sarzynski, Jaromir ^{[1
]}

机构：

[1] Univ Rzeszow, Rzeszow, Poland

[2] Szymon Szymonow State Sch Higher Educ, Zamosc, Poland

[3] Univ Rzeszow, Univ Informat Technol & Management, Rzeszow, Poland

来源：

IEEE CIS INTERNATIONAL CONFERENCE ON FUZZY SYSTEMS 2021 (FUZZ-IEEE) | 2021年

关键词：

Attribute selection; Attribute evaluation; Fuzzification; Boruta algorithm; Discretization; SYSTEM; BORUTA;

D O I：

10.1109/FUZZ45933.2021.9494515

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In the paper, we show how the importance of individual ranges of values of attributes describing cases can be determined using the attribute fuzzification process. The importance is determined on the basis of classification capabilities. The described approach is based mainly on fuzzy set theory and the rough set based discretization method. Moreover, an experimental study of the computer-aided classification task is presented.

引用

页数：6

共 18 条

[1]

[Anonymous], 2021, IEEE Trans. Broadcast.

[2]

Bazan JG, 2000, STUD FUZZ SOFT COMP, V56, P49

[3]

Goodfellow I, 2016, ADAPT COMPUT MACH LE, P1

[4] ANALYSIS OF HIDDEN UNITS IN A LAYERED NETWORK TRAINED TO CLASSIFY SONAR TARGETS [J].