A Weights Direct Determination Neural Network for International Standard Classification of Occupations

被引：1

作者：

Lagios, Dimitris ^{[1
]}

Mourtas, Spyridon D. ^{[2
,3
]}

Zervas, Panagiotis ^{[1
]}

Tzimas, Giannis ^{[1
]}

机构：

[1] Univ Peloponnese, Dept Elect & Comp Engn, Data & Media Lab, Patras 26334, Greece

[2] Natl & Kapodistrian Univ Athens, Dept Econ Math Informat & Stat Econometr, Sofokleous 1 St, Athens 10559, Greece

[3] Siberian Fed Univ, Lab Hybrid Methods Modelling & Optimizat Complex S, Prosp Svobodny 79, Krasnoyarsk 660041, Russia

来源：

MATHEMATICS | 2023年 / 11卷 / 03期

关键词：

neural networks; weights and structure determination; multiclass classification; international standard classification of occupations; machine learning; ALGORITHM; SKILLS; MODEL;

D O I：

10.3390/math11030629

中图分类号：

O1 [数学];

学科分类号：

0701 ; 070101 ;

摘要：

Multiclass classification is one of the most popular machine learning tasks. The main focus of this paper is to classify occupations according to the International Standard Classification of Occupations (ISCO) using a weights and structure determination (WASD)-based neural network. In general, WASD-trained neural networks are known to overcome the drawbacks of conventional back-propagation trained neural networks, such as slow training speed and local minimum. However, WASD-based neural networks have not yet been applied to address the challenges of multiclass classification. As a result, a novel WASD for multiclass classification (WASDMC)-based neural network is introduced in this paper. When applied to two publicly accessible ISCO datasets, the WASDMC-based neural network displayed superior performance across all measures, compared to some of the best-performing classification models that the MATLAB classification learner app has to offer.

引用

页数：14