{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,9,11]],"date-time":"2025-09-11T19:20:05Z","timestamp":1757618405163,"version":"3.44.0"},"reference-count":39,"publisher":"Springer Science and Business Media LLC","issue":"8","license":[{"start":{"date-parts":[[2025,6,24]],"date-time":"2025-06-24T00:00:00Z","timestamp":1750723200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,6,24]],"date-time":"2025-06-24T00:00:00Z","timestamp":1750723200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100008131","name":"Rheinische Friedrich-Wilhelms-Universit\u00e4t Bonn","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100008131","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Mach Learn"],"published-print":{"date-parts":[[2025,8]]},"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:p>Graph neural networks (GNNs) are among the most widely used methods for node classification in graphs. A common strategy to improve their predictive performance is to enrich nodes with additional features. A weakness of this method is that the set of appropriate features can vary from graph to graph. We address this shortcoming by proposing a novel method. In a preprocessing step, a first GNN is trained on a set of graphs with varying structural properties, using a candidate set of node features fixed in advance. The resulting GNN model is then used to predict the most relevant features from the candidate set for unseen target graphs, which are later processed for node classification. For each target graph, a second GNN is trained on the graph, which is enriched with the node feature vectors calculated for the features selected by the first GNN. A key advantage of the proposed method is that the features are selected without computing the candidate features for the target graph. Our experimental results on synthetic and real-world graphs show that even a few features selected in this way is sufficient to significantly improve the predictive performance of GNNs that use either none or all of the candidate features. Moreover, the time needed to learn the second GNN for the target graph can be reduced by up to two orders of magnitude.<\/jats:p>","DOI":"10.1007\/s10994-025-06815-z","type":"journal-article","created":{"date-parts":[[2025,6,24]],"date-time":"2025-06-24T11:36:40Z","timestamp":1750765000000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Improving graph neural networks through feature importance learning"],"prefix":"10.1007","volume":"114","author":[{"given":"Fouad","family":"Alkhoury","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tam\u00e1s","family":"Horv\u00e1th","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Christian","family":"Bauckhage","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Stefan","family":"Wrobel","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2025,6,24]]},"reference":[{"key":"6815_CR1","doi-asserted-by":"crossref","unstructured":"Abboud, R., Ceylan, \u0130. \u0130., Grohe, M., & Lukasiewicz, T. (2021). The surprising power of graph neural networks with random node initialization. In Proceedings of the 30th International Joint Conference on Artificial Intelligence (pp. 2112\u20132118).","DOI":"10.24963\/ijcai.2021\/291"},{"issue":"7","key":"6815_CR2","doi-asserted-by":"publisher","first-page":"0130140","DOI":"10.1371\/journal.pone.0130140","volume":"10","author":"S Bach","year":"2015","unstructured":"Bach, S., Binder, A., Montavon, G., Klauschen, F., M\u00fcller, K.-R., & Samek, W. (2015). On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation. PloS One, 10(7), 0130140.","journal-title":"PloS One"},{"issue":"5439","key":"6815_CR3","doi-asserted-by":"publisher","first-page":"509","DOI":"10.1126\/science.286.5439.509","volume":"286","author":"A-L Barab\u00e1si","year":"1999","unstructured":"Barab\u00e1si, A.-L., & Albert, R. (1999). Emergence of scaling in random networks. Science, 286(5439), 509\u2013512.","journal-title":"Science"},{"key":"6815_CR4","unstructured":"Barcel\u00f3, P., Geerts, F., Reutter, J. L., & Ryschkov, M. (2021). Graph neural networks with local graph parameters. In 35th annual conference on neural information processing systems (NeurIPS) (pp. 25280\u201325293)."},{"key":"6815_CR5","doi-asserted-by":"publisher","DOI":"10.1093\/oso\/9780198538493.001.0001","volume-title":"Neural networks for pattern recognition","author":"CM Bishop","year":"1995","unstructured":"Bishop, C. M. (1995). Neural networks for pattern recognition. Oxford University Press."},{"key":"6815_CR6","unstructured":"Bojchevski, A., & G\u00fcnnemann, S. (2018). Deep gaussian embedding of graphs: unsupervised inductive learning via ranking. In 6th international conference on learning representations, ICLR. openreview https:\/\/openreview.net\/forum?id=r1ZdKJ-0W."},{"issue":"1","key":"6815_CR7","doi-asserted-by":"publisher","first-page":"657","DOI":"10.1109\/TPAMI.2022.3154319","volume":"45","author":"G Bouritsas","year":"2023","unstructured":"Bouritsas, G., Frasca, F., Zafeiriou, S., & Bronstein, M. M. (2023). Improving graph neural network expressivity via subgraph isomorphism counting. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(1), 657\u2013668.","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"6815_CR8","doi-asserted-by":"crossref","unstructured":"Cui, H., Lu, Z., Li, P., & Yang, C. (2022). On positional and structural node features for graph neural networks on non-attributed graphs. In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management (pp. 3898\u20133902).","DOI":"10.1145\/3511808.3557661"},{"key":"6815_CR9","unstructured":"Duong, C. T., Hoang, T. D., Dang, H. T. H., Nguyen, Q. V. H., & Aberer, K. (2019). On node features for graph neural networks. CoRR arXiv:1911.08795."},{"key":"6815_CR10","doi-asserted-by":"publisher","first-page":"290","DOI":"10.5486\/PMD.1959.6.3-4.12","volume":"6","author":"P Erd\u0151s","year":"1959","unstructured":"Erd\u0151s, P., & R\u00e9nyi, A. (1959). On random graphs I. Publicationes Mathematicae Debrecen, 6, 290\u2013297.","journal-title":"Publicationes Mathematicae Debrecen"},{"key":"6815_CR11","unstructured":"Fey, M., & Lenssen, J. E. (2019). Fast graph representation learning with pytorch geometric. CoRR arXiv:1903.02428."},{"key":"6815_CR12","doi-asserted-by":"crossref","unstructured":"Grover, A., & Leskovec, J. (2016). node2vec: Scalable feature learning for networks. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (pp. 855\u2013864). New York: ACM Press.","DOI":"10.1145\/2939672.2939754"},{"key":"6815_CR13","first-page":"219","volume-title":"Innovative methods in computer science and computational applications in the era of industry 5.0","author":"A G\u00fcrb\u00fcz","year":"2024","unstructured":"G\u00fcrb\u00fcz, A., & U\u011fur, A. (2024). Feature processing on artificial graph node features for classification with graph neural networks. In D. J. Hemanth, U. Kose, B. Patrut, & M. Ersoy (Eds.), Innovative methods in computer science and computational applications in the era of industry 5.0 (pp. 219\u2013231). Springer."},{"key":"6815_CR14","unstructured":"Hamilton, W.L., Ying, Z., & Leskovec, J. (2017). Inductive representation learning on large graphs. In Advances in neural information processing systems 30: Annual conference on neural information processing systems 2017 (pp. 1024\u20131034)."},{"key":"6815_CR15","unstructured":"Hamilton, W. L., Ying, Z., & Leskovec, J. (2017). Inductive representation learning on large graphs. In: Guyon, I., Luxburg, U., Bengio, S., Wallach, H. M., Fergus, R., Vishwanathan, S. V. N., Garnett, R. (Eds.), Advances in neural information processing systems 30: Annual conference on neural information processing systems (pp. 1024\u20131034)."},{"key":"6815_CR16","volume-title":"Graph representation learning synthesis. Lectures on Artificial Intelligence and Machine Learning","author":"WL Hamilton","year":"2020","unstructured":"Hamilton, W. L. (2020). Graph representation learning synthesis. Lectures on Artificial Intelligence and Machine Learning. Morgan &amp; Claypool Publishers."},{"key":"6815_CR17","unstructured":"He, X., Cai, D., & Niyogi, P. (2005). Laplacian score for feature selection. Advances in Neural Information Processing Systems, 18, 507\u2013514."},{"key":"6815_CR18","unstructured":"Kipf, T. N., & Welling, M. (2016). Semi-supervised classification with graph convolutional networks. CoRR arXiv:1609.02907."},{"key":"6815_CR19","doi-asserted-by":"crossref","unstructured":"Liu, X., Pan, H., He, M., Song, Y., Jiang, X., & Shang, L. (2020). Neural subgraph isomorphism counting. In The 26th ACM SIGKDD conference on knowledge discovery and data mining (pp. 1959\u20131969). New York: ACM Press.","DOI":"10.1145\/3394486.3403247"},{"key":"6815_CR20","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1007\/BF02280291","volume":"18","author":"LM Lov\u00e1sz","year":"1967","unstructured":"Lov\u00e1sz, L. M. (1967). Operations with structures. Acta Mathematica Academiae Scientiarum Hungarica, 18, 321\u2013328.","journal-title":"Acta Mathematica Academiae Scientiarum Hungarica"},{"issue":"1","key":"6815_CR21","doi-asserted-by":"publisher","first-page":"14","DOI":"10.1049\/cit2.12166","volume":"8","author":"SK Maurya","year":"2023","unstructured":"Maurya, S. K., Liu, X., & Murata, T. (2023). Feature selection: Key to enhance node classification with graph neural networks. CAAI Transactions on Intelligence Technology, 8(1), 14\u201328.","journal-title":"CAAI Transactions on Intelligence Technology"},{"key":"6815_CR22","doi-asserted-by":"crossref","unstructured":"McAuley, J. J., Targett, C., Shi, Q., & Hengel, A. (2015). Image-based recommendations on styles and substitutes. In: Proceedings of the 38th International ACM SIGIR conference on research and development in information retrieval (pp. 43\u201352). New York: ACM Press.","DOI":"10.1145\/2766462.2767755"},{"key":"6815_CR23","first-page":"333","volume":"24","author":"C Morris","year":"2023","unstructured":"Morris, C., Lipman, Y., Maron, H., Rieck, B., Kriege, N. M., Grohe, M., Fey, M., & Borgwardt, K. M. (2023). Weisfeiler and leman go machine learning: The story so far. Journal of Machine Learning Research, 24, 333\u2013133359.","journal-title":"Journal of Machine Learning Research"},{"key":"6815_CR24","doi-asserted-by":"crossref","unstructured":"Morris, C., Ritzert, M., Fey, M., Hamilton, W. L., Lenssen, J. E., Rattan, G., & Grohe, M. (2019). Weisfeiler and Leman go neural: Higher-order graph neural networks. In The 33rd AAAI conference on artificial intelligence (Vol. 33, pp. 4602\u20134609).","DOI":"10.1609\/aaai.v33i01.33014602"},{"key":"6815_CR25","unstructured":"Naik, H., Polster, J., Shekhar, R., Horv\u00e1th, T., & Tur\u00e1n, G. (2024). Iterative graph neural network enhancement via frequent subgraph mining of explanations. CoRR arXiv:2403.07849."},{"key":"6815_CR26","unstructured":"Namata, G., London, B., Getoor, L., Huang, B., & Edu, U. (2012). Query-driven active surveying for collective classification. In 10th international workshop on mining and learning with graphs (Vol. 8, p. 1)."},{"key":"6815_CR27","unstructured":"Nguyen, H., & Maehara, T. (2020). Graph homomorphism convolution. In Proceedings of the 37th international conference on machine learning, ICML 2020. Proceedings of machine learning research (Vol. 119, pp. 7306\u20137316). Vienna, Austria: PMLR."},{"key":"6815_CR28","unstructured":"Pan, S., Wu, J., Zhu, X., Zhang, C., & Wang, Y. (2016). Tri-party deep network representation. In Proceeding of the twenty-fifth international joint conference on artificial intelligence, IJCAI (pp. 1895\u20131901). New York: IJCAI\/AAAI Press."},{"key":"6815_CR29","doi-asserted-by":"crossref","unstructured":"Perozzi, B., Al-Rfou, R., & Skiena, S. (2014). Deepwalk: Online learning of social representations. In The 20th ACM SIGKDD international conference on knowledge discovery and data mining (pp. 701\u2013710). New York: ACM Press.","DOI":"10.1145\/2623330.2623732"},{"key":"6815_CR30","unstructured":"Platonov, O., Kuznedelev, D., Diskin, M., Babenko, A., & Prokhorenkova, L. (2023). A critical look at the evaluation of gnns under heterophily: Are we really making progress? In The 11th international conference on learning representations, ICLR. OpenReview.net, Kigali, Rwanda. openreview https:\/\/openreview.net\/forum?id=tJbbQfw-5wv"},{"key":"6815_CR31","doi-asserted-by":"crossref","unstructured":"Ribeiro, L. F. R., Saverese, P. H. P., & Figueiredo, D. R. (2017). struc2vec: Learning node representations from structural identity. In: Proceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining (pp. 385\u2013394). New York: ACM Press.","DOI":"10.1145\/3097983.3098061"},{"key":"6815_CR32","doi-asserted-by":"crossref","unstructured":"Sato, R., Yamada, M., spsampsps Kashima, H. (2021). Random features strengthen graph neural networks. In Demeniconi, C., spsampsps Davidson, I. (Eds.), Proceedings of the 2021 SIAM international conference on data mining (pp. 333\u2013341).","DOI":"10.1137\/1.9781611976700.38"},{"issue":"3","key":"6815_CR33","doi-asserted-by":"publisher","first-page":"93","DOI":"10.1609\/aimag.v29i3.2157","volume":"29","author":"P Sen","year":"2008","unstructured":"Sen, P., Namata, G., Bilgic, M., Getoor, L., Galligher, B., & Eliassi-Rad, T. (2008). Collective classification in network data. AI Magazine, 29(3), 93\u201393.","journal-title":"AI Magazine"},{"key":"6815_CR34","unstructured":"Shchur, O., Mumme, M., Bojchevski, A., & G\u00fcnnemann, S. (2018). Pitfalls of graph neural network evaluation. CoRR arXiv:1811.05868."},{"key":"6815_CR35","unstructured":"Simonyan, K., Vedaldi, A., & Zisserman, A. (2013). Deep inside convolutional networks: Visualising image classification models and saliency maps. arXiv preprint arXiv:1312.6034."},{"key":"6815_CR36","doi-asserted-by":"publisher","first-page":"46","DOI":"10.1016\/j.patrec.2022.04.036","volume":"159","author":"DC Thang","year":"2022","unstructured":"Thang, D. C., Dat, H. T., Tam, N. T., Jo, J., Hung, N. Q. V., & Aberer, K. (2022). Nature vs. nurture: Feature vs structure for graph neural networks. Pattern Recognition Letters, 159, 46\u201353.","journal-title":"Pattern Recognition Letters"},{"issue":"20","key":"6815_CR37","first-page":"10","volume":"1050","author":"P Velickovic","year":"2017","unstructured":"Velickovic, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., Bengio, Y., et al. (2017). Graph attention networks. stat, 1050(20), 10\u201348550.","journal-title":"stat"},{"issue":"9","key":"6815_CR38","first-page":"12","volume":"2","author":"B Weisfeiler","year":"1968","unstructured":"Weisfeiler, B., & Leman, A. A. (1968). A reduction of a graph to a canonical form and an algebra arising during this reduction. Nauchno-Technicheskaya Informatsiya, 2(9), 12\u201316.","journal-title":"Nauchno-Technicheskaya Informatsiya"},{"issue":"1","key":"6815_CR39","doi-asserted-by":"publisher","first-page":"37","DOI":"10.14778\/3421424.3421430","volume":"14","author":"R Yang","year":"2020","unstructured":"Yang, R., Shi, J., Xiao, X., Yang, Y., Liu, J., & Bhowmick, S. S. (2020). Scaling attributed network embedding to massive graphs. Proceedings of the VLDB Endowment, 14(1), 37\u201349.","journal-title":"Proceedings of the VLDB Endowment"}],"container-title":["Machine Learning"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10994-025-06815-z.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10994-025-06815-z\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10994-025-06815-z.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,6]],"date-time":"2025-09-06T22:17:33Z","timestamp":1757197053000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10994-025-06815-z"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,24]]},"references-count":39,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2025,8]]}},"alternative-id":["6815"],"URL":"https:\/\/doi.org\/10.1007\/s10994-025-06815-z","relation":{},"ISSN":["0885-6125","1573-0565"],"issn-type":[{"type":"print","value":"0885-6125"},{"type":"electronic","value":"1573-0565"}],"subject":[],"published":{"date-parts":[[2025,6,24]]},"assertion":[{"value":"11 April 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"28 May 2025","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"3 June 2025","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 June 2025","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"178"}}