Feature-Based Complexity Measure for Multinomial Classification Datasets

被引:0
|
作者
Erwin, Kyle [1 ]
Engelbrecht, Andries [1 ,2 ,3 ]
机构
[1] Stellenbosh Univ, Comp Sci Div, ZA-7600 Stellenbosch, South Africa
[2] Stellenbosh Univ, Dept Ind Engn, ZA-7600 Stellenbosch, South Africa
[3] Gulf Univ Sci & Technol, Ctr Appl Math & Bioinformat, Mubarak Al Abdullah 32093, Kuwait
关键词
multinomial classification datasets; classification problem complexity; feature-based complexity measures; synthetic datasets;
D O I
10.3390/e25071000
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
Machine learning algorithms are frequently used for classification problems on tabular datasets. In order to make informed decisions about model selection and design, it is crucial to gain meaningful insights into the complexity of these datasets. Feature-based complexity measures are a set of complexity measures that evaluates how useful features are at discriminating instances of different classes. This paper, however, shows that existing feature-based measures are inadequate in accurately measuring the complexity of various synthetic classification datasets, particularly those with multiple classes. This paper proposes a new feature-based complexity measure called the F5 measure, which evaluates the discriminative power of features for each class by identifying long sequences of uninterrupted instances of the same class. It is shown that the F5 measure better represents the feature complexity of a dataset.
引用
收藏
页数:18
相关论文
共 50 条
  • [1] A PROPOSAL FOR FEATURE CLASSIFICATION IN FEATURE-BASED DESIGN
    OVTCHAROVA, J
    PAHL, G
    RIX, J
    [J]. COMPUTERS & GRAPHICS, 1992, 16 (02) : 187 - 195
  • [2] Feature-Based Lung Nodule Classification
    Farag, Amal
    Ali, Asem
    Graham, James
    Elhabian, Shireen
    Farag, Aly
    Falk, Robert
    [J]. ADVANCES IN VISUAL COMPUTING, PT III, 2010, 6455 : 79 - +
  • [3] Feature-Based Dissimilarity Space Classification
    Duin, Robert P. W.
    Loog, Marco
    Pekalska, Elzbieta
    Tax, David M. J.
    [J]. RECOGNIZING PATTERNS IN SIGNALS, SPEECH, IMAGES, AND VIDEOS, 2010, 6388 : 46 - +
  • [4] STATISTICAL FEATURE-BASED CRAQUELURE CLASSIFICATION
    Crisologo, Irene
    Monterola, Christopher
    Soriano, Maricor
    [J]. INTERNATIONAL JOURNAL OF MODERN PHYSICS C, 2011, 22 (11): : 1191 - 1209
  • [5] Feature-Based Terrain Classification For LittleDog
    Filitchkin, Paul
    Byl, Katie
    [J]. 2012 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2012, : 1387 - 1392
  • [6] EndNote: Feature-based classification of networks
    Barnett, Ian
    Malik, Nishant
    Kuijjer, Marieke L.
    Mucha, Peter J.
    Onnela, Jukka-Pekka
    [J]. NETWORK SCIENCE, 2019, 7 (03) : 438 - 444
  • [7] A feature-based trust sequence classification algorithm
    Yahyaoui, Hamdi
    Al-Mutairi, Aisha
    [J]. INFORMATION SCIENCES, 2016, 328 : 455 - 484
  • [8] A FEATURE-BASED ULTRASONIC SYSTEM FOR IGSCC CLASSIFICATION
    ROSE, J
    AVIOLI, M
    LAPIDES, M
    [J]. MATERIALS EVALUATION, 1982, 40 (03) : A13 - A13
  • [9] Multimodal Feature-Based Surface Material Classification
    Strese, Matti
    Schuwerk, Clemens
    Iepure, Albert
    Steinbach, Eckehard
    [J]. IEEE TRANSACTIONS ON HAPTICS, 2017, 10 (02) : 226 - 239
  • [10] Feature-based augmentation and classification for tabular data
    Sathianarayanan, Balachander
    Samant, Yogesh Chandra Singh
    Guruprasad, Prahalad S. Conjeepuram
    Hariharan, Varshin B.
    Manickam, Nirmala Devi
    [J]. CAAI TRANSACTIONS ON INTELLIGENCE TECHNOLOGY, 2022, 7 (03) : 481 - 491