{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,20]],"date-time":"2026-03-20T09:55:29Z","timestamp":1774000529810,"version":"3.50.1"},"reference-count":39,"publisher":"Wiley","issue":"1","license":[{"start":{"date-parts":[[2026,3,19]],"date-time":"2026-03-19T00:00:00Z","timestamp":1773878400000},"content-version":"vor","delay-in-days":77,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"funder":[{"DOI":"10.13039\/100010046","name":"University of Essex","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100010046","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["International Journal of Intelligent Systems"],"published-print":{"date-parts":[[2026,1]]},"abstract":"<jats:p>Meta\u2010learning techniques aim to predict the most suitable learning algorithm for a given dataset based on its intrinsic structural characteristics. These techniques provide a robust framework for understanding algorithmic behavior across diverse data distributions and attributes. Although these state\u2010of\u2010the\u2010art models (CNNs and transformers) are widely applied in various machine learning tasks, their use on numerical datasets remains underexplored due to the complexity of their internal structures. This study aims not only to predict the performance of two black\u2010box deep learning models on static datasets but also to conduct a behavioral analysis in order to identify which meta\u2010features most strongly influence their outcomes. It seems unclear which specific attributes of a dataset positively or negatively affect the performance of these deep learning models. To bridge this gap, we constructed a meta\u2010dataset consisting of 296 datasets, each characterized by 20 meta\u2010features describing the dataset\u2019s statistical, geometric, and structural properties. The analysis identifies which intrinsic dataset properties influence model accuracy, without relying on raw data or hyperparameter tuning. Results show that both models perform best on datasets with high feature discriminability, as captured by meta\u2010features such as maximum feature efficiency, collective feature efficiency, and directional separability. In contrast, performance declines with increasing class boundary complexity and nonlinearity, reflected in features like class separability measures and the linear classifier nonlinearity metric. While CNNs are more sensitive to local geometric complexity, transformers respond more strongly to global statistical measures such as mutual information and entropy, highlighting their distinct inductive biases. The proposed meta\u2010model accurately predicts the performance of both architectures on unseen datasets (0.96 correlation coefficient, 0.019 MAE, and 0.025 RMSE for CNNs; 0.92 correlation coefficient, 0.027 MAE, and 0.036 RMSE for transformers), enabling performance estimation without costly training. These findings emphasize the importance of aligning model architecture with dataset geometry and structure. Additionally, the framework supports more interpretable, efficient, and sustainable deep learning model selection in structured data settings.<\/jats:p>","DOI":"10.1155\/int\/8573962","type":"journal-article","created":{"date-parts":[[2026,3,20]],"date-time":"2026-03-20T06:51:32Z","timestamp":1773989492000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Meta\u2010Learning Analysis of Deep Neural Network Architectures on Diverse Numeric Datasets via Geometric Complexity Descriptors"],"prefix":"10.1155","volume":"2026","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2960-8725","authenticated-orcid":false,"given":"Faruk","family":"Bulut","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8344-1180","authenticated-orcid":false,"given":"\u0130knur","family":"D\u00f6nmez","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2026,3,19]]},"reference":[{"key":"e_1_2_12_1_2","doi-asserted-by":"publisher","DOI":"10.1109\/tpami.2024.3357847"},{"key":"e_1_2_12_2_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2021.108101"},{"key":"e_1_2_12_3_2","article-title":"Masif: Meta-Learned Algorithm Selection Using Implicit Fidelity Information","author":"Ruhkopf T.","year":"2023","journal-title":"Transactions on Machine Learning Research"},{"key":"e_1_2_12_4_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-43046-5"},{"key":"e_1_2_12_5_2","doi-asserted-by":"crossref","DOI":"10.1109\/TPAMI.2026.3656494","article-title":"First-Order Cross-Domain Meta Learning for Few-Shot Remote Sensing Object Classification","author":"Zhao W.","year":"2026","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"e_1_2_12_6_2","doi-asserted-by":"crossref","unstructured":"SurendroK. RachmatullahM. I. C. andSantosoJ. Improving 1d Convolutional Neural Network (1D Cnn) Performance in Processing Tabular Datasets Using Principal Component Analysis 2022.","DOI":"10.21203\/rs.3.rs-2149472\/v1"},{"key":"e_1_2_12_7_2","article-title":"CTMNR: A Hybrid CNN-Transformer Approach for Enhanced Multi-Nuclide Classification in Radiation Spectroscopy","author":"Zhang Z.","year":"2025","journal-title":"IEEE Transactions on Nuclear Science"},{"key":"e_1_2_12_8_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-024-08328-6"},{"key":"e_1_2_12_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/access.2024.3352634"},{"key":"e_1_2_12_10_2","article-title":"Network Intrusion Detection Based on Feature Selection and Hybrid Metaheuristic Optimization","volume":"74","author":"Alkanhel R.","year":"2023","journal-title":"Computers, Materials and Continua"},{"key":"e_1_2_12_11_2","doi-asserted-by":"publisher","DOI":"10.3390\/math10234421"},{"key":"e_1_2_12_12_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10664-024-10554-5"},{"key":"e_1_2_12_13_2","first-page":"571","article-title":"Data Complexity Measures for Classification of a Multi-Concept Dataset","volume":"84","author":"Gyanchandani M.","year":"2025","journal-title":"Multimedia Tools and Applications"},{"key":"e_1_2_12_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/34.990132"},{"key":"e_1_2_12_15_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2024.110881"},{"key":"e_1_2_12_16_2","doi-asserted-by":"crossref","first-page":"54","DOI":"10.1007\/s00357-024-09481-3","article-title":"Cluster Validation Based on Fisher\u2019s Linear Discriminant Analysis","volume":"42","author":"K\u00e4chele F.","year":"2025","journal-title":"Journal of Classification"},{"key":"e_1_2_12_17_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10044-015-0504-0"},{"key":"e_1_2_12_18_2","doi-asserted-by":"publisher","DOI":"10.1007\/s00500-022-06787-5"},{"key":"e_1_2_12_19_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2024.107865"},{"key":"e_1_2_12_20_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10898-025-01483-8"},{"key":"e_1_2_12_21_2","doi-asserted-by":"crossref","first-page":"144082","DOI":"10.1109\/ACCESS.2024.3433513","article-title":"State-of-the-Art in 1d Convolutional Neural Networks: A Survey","volume":"12","author":"Ige A. O.","year":"2024","journal-title":"IEEE Access"},{"key":"e_1_2_12_22_2","doi-asserted-by":"crossref","first-page":"53","DOI":"10.1007\/978-3-030-42227-1_5","volume-title":"Computational Intelligence: A Methodological Introduction","author":"Kruse R.","year":"2022"},{"key":"e_1_2_12_23_2","article-title":"Parametric Activation Functions for Neural Networks: A Tutorial Survey","author":"Pusztah\u00e1zi L. S.","year":"2024","journal-title":"IEEE Access"},{"key":"e_1_2_12_24_2","doi-asserted-by":"crossref","DOI":"10.1109\/TNNLS.2025.3646122","article-title":"Graph Transformers: A Survey","author":"Shehzad A.","year":"2026","journal-title":"IEEE Transactions on Neural Networks and Learning Systems"},{"key":"e_1_2_12_25_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ins.2024.121607"},{"key":"e_1_2_12_26_2","doi-asserted-by":"publisher","DOI":"10.32604\/cmc.2023.031723"},{"key":"e_1_2_12_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/access.2023.3298955"},{"key":"e_1_2_12_28_2","unstructured":"Kaggle Classification Datasets 2025 https:\/\/www.kaggle.com\/datasets?tags=13302-Classification."},{"key":"e_1_2_12_29_2","volume-title":"UCI Machine Learning Repository","author":"Bache K.","year":"2025"},{"key":"e_1_2_12_30_2","first-page":"255","article-title":"KEEL Data-Mining Software Tool: Data Set Repository, Integration of Algorithms and Experimental Analysis Framework","volume":"17","author":"Alcal\u00b4a-Fdez J.","year":"2011","journal-title":"Journal of Multiple-Valued Logic and Soft Computing"},{"key":"e_1_2_12_31_2","article-title":"Documentation for the Data Complexity Library in C++","volume":"196","author":"Orriols-Puig A.","year":"2010","journal-title":"Universitat Ramon Llull, La Salle"},{"key":"e_1_2_12_32_2","volume-title":"The WEKA Workbench. Online Appendix for Data Mining: Practical Machine Learning Tools and Techniques","author":"Frank E.","year":"2016"},{"key":"e_1_2_12_33_2","doi-asserted-by":"crossref","first-page":"69","DOI":"10.1007\/978-3-031-38747-0_3","volume-title":"An Introduction to Statistical Learning: With Applications in Python","author":"James G.","year":"2023"},{"key":"e_1_2_12_34_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11749-016-0481-7"},{"key":"e_1_2_12_35_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.sigpro.2024.109511"},{"key":"e_1_2_12_36_2","doi-asserted-by":"publisher","DOI":"10.1080\/24725854.2023.2299809"},{"key":"e_1_2_12_37_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11222-016-9646-1"},{"key":"e_1_2_12_38_2","doi-asserted-by":"publisher","DOI":"10.1186\/s13040-023-00322-4"},{"key":"e_1_2_12_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/access.2023.3253430"}],"container-title":["International Journal of Intelligent Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1155\/int\/8573962","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1155\/int\/8573962","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1155\/int\/8573962","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,20]],"date-time":"2026-03-20T06:51:35Z","timestamp":1773989495000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1155\/int\/8573962"}},"subtitle":[],"editor":[{"given":"Richard","family":"Murray","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2026,1]]},"references-count":39,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2026,1]]}},"alternative-id":["10.1155\/int\/8573962"],"URL":"https:\/\/doi.org\/10.1155\/int\/8573962","archive":["Portico"],"relation":{},"ISSN":["0884-8173","1098-111X"],"issn-type":[{"value":"0884-8173","type":"print"},{"value":"1098-111X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1]]},"assertion":[{"value":"2025-10-16","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-02-11","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-03-19","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"8573962"}}