{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,30]],"date-time":"2026-01-30T06:59:17Z","timestamp":1769756357811,"version":"3.49.0"},"reference-count":49,"publisher":"Emerald","issue":"2","license":[{"start":{"date-parts":[[2024,12,24]],"date-time":"2024-12-24T00:00:00Z","timestamp":1734998400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.emerald.com\/insight\/site-policies"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["DTA"],"published-print":{"date-parts":[[2025,4,11]]},"abstract":"<jats:sec><jats:title content-type=\"abstract-subheading\">Purpose<\/jats:title><jats:p>International trade transactions, extracted from customs declarations, include several fields, among which the product description and the product category are the most important. The product category, also referred to as the Harmonised System Code (HS code), serves as a pivotal component for determining tax rates and administrative purposes. A predictive tool designed for product categories or HS codes becomes an important resource aiding traders in their decision to choose a suitable code. This tool is instrumental in preventing misclassification arising from the ambiguities present in product nomenclature, thus mitigating the challenges associated with code interpretation. Moreover, deploying this tool would streamline the validation process for government officers dealing with extensive transactions, optimising their workload and enhancing tax revenue collection within this domain.<\/jats:p><\/jats:sec><jats:sec><jats:title content-type=\"abstract-subheading\">Design\/methodology\/approach<\/jats:title><jats:p>This study introduces a methodology focused on the generation of sentence embeddings for trade transactions, employing Sentence BERT (SBERT) framework in conjunction with the Multiple Negative Ranking (MNR) Loss function following a contrastive learning paradigm. The procedure involves the construction of pairwise samples, including anchors and positive transactions. The proposed method is evaluated using two publicly available real-world datasets, specifically the India Import 2016 and United States Import 2018 datasets, to fine-tune the SBERT model. Several configurations involving pooling strategies, loss functions, and training parameters are explored within the experimental setup. The acquired representations serve as inputs for traditional machine learning algorithms employed in predicting the product categories within trade transactions.<\/jats:p><\/jats:sec><jats:sec><jats:title content-type=\"abstract-subheading\">Findings<\/jats:title><jats:p>Encoding trade transactions utilising SBERT with MNR loss facilitates the creation of enhanced embeddings that exhibit improved representational capacity. These fixed-length embeddings serve as adaptable inputs for training machine learning models, including support vector machine (SVM) and random forest, intended for downstream tasks of HS code classification. Empirical evidence supports the superior performance of our proposed approach compared to fine-tuning transformer-based models in the domain of trade transaction classification.<\/jats:p><\/jats:sec><jats:sec><jats:title content-type=\"abstract-subheading\">Originality\/value<\/jats:title><jats:p>Our approach generates more representative sentence embeddings by creating the network architectures from scratch with the SBERT framework. Instead of exploiting a data augmentation method generally used in contrastive learning for measuring the similarity between the samples, we arranged positive samples following a supervised paradigm and determined loss through distance learning metrics. This process involves continuous updating of the Siamese or bi-encoder network to produce embeddings derived from commodity transactions. This strategy aims to ensure that similar concepts of transactions within the same class converge closer within the feature embedding space, thereby improving the performance of downstream tasks.<\/jats:p><\/jats:sec>","DOI":"10.1108\/dta-01-2024-0052","type":"journal-article","created":{"date-parts":[[2024,12,24]],"date-time":"2024-12-24T00:02:16Z","timestamp":1734998536000},"page":"276-301","source":"Crossref","is-referenced-by-count":1,"title":["Harmonized system code classification using supervised contrastive learning with sentence BERT and multiple negative ranking loss"],"prefix":"10.1108","volume":"59","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-3175-1444","authenticated-orcid":false,"given":"Angga Wahyu","family":"Anggoro","sequence":"first","affiliation":[]},{"given":"Padraig","family":"Corcoran","sequence":"additional","affiliation":[]},{"given":"Dennis","family":"De Widt","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2913-4478","authenticated-orcid":false,"given":"Yuhua","family":"Li","sequence":"additional","affiliation":[]}],"member":"140","published-online":{"date-parts":[[2024,12,24]]},"reference":[{"key":"key2025041114531597100_ref001","doi-asserted-by":"crossref","unstructured":"Altaheri, F. and Shaalan, K. (2020), \u201cExploring machine learning models to predict harmonized system code\u201d, in Themistocleous, M. and Papadaki, M. (Eds), Information Systems, Springer International Publishing, Cham, pp.\u00a0291-303, doi: 10.1007\/978-3-030-44322-1_22.","DOI":"10.1007\/978-3-030-44322-1_22"},{"key":"key2025041114531597100_ref002","doi-asserted-by":"publisher","first-page":"3","DOI":"10.18523\/2617-7080320203-10","article-title":"Generalization of cross-entropy loss function for image classification","volume":"3","year":"2021","journal-title":"Mohyla Mathematical Journal"},{"key":"key2025041114531597100_ref003","article-title":"Using DistilBERT to assign HS codes to international trading transactions","year":"2023"},{"key":"key2025041114531597100_ref004","doi-asserted-by":"publisher","first-page":"632","DOI":"10.18653\/v1\/D15-1075","article-title":"A large annotated corpus for learning Natural Language inference","year":"2015"},{"issue":"1","key":"key2025041114531597100_ref005","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","year":"2001","journal-title":"Machine Learning"},{"key":"key2025041114531597100_ref006","first-page":"1597","article-title":"A simple framework for contrastive learning of visual representations","year":"2020"},{"key":"key2025041114531597100_ref007","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/DSAA53316.2021.9564203","article-title":"The use of machine learning to identify the correctness of HS Code for the customs import declarations","year":"2021"},{"key":"key2025041114531597100_ref008","doi-asserted-by":"publisher","article-title":"AugGPT: leveraging ChatGPT for text data augmentation","year":"2023","DOI":"10.48550\/ARXIV.2302.13007"},{"issue":"1","key":"key2025041114531597100_ref009","doi-asserted-by":"publisher","first-page":"8","DOI":"10.3390\/bdcc6010008","article-title":"An empirical comparison of Portuguese and multilingual BERT models for auto-classification of NCM codes in international trade","volume":"6","year":"2022","journal-title":"Big Data and Cognitive Computing"},{"key":"key2025041114531597100_ref010","doi-asserted-by":"publisher","first-page":"4171","DOI":"10.18653\/v1\/N19-1423","article-title":"BERT: pre-training of deep bidirectional transformers for language understanding","year":"2019"},{"key":"key2025041114531597100_ref011","doi-asserted-by":"publisher","first-page":"676","DOI":"10.1007\/978-3-030-75765-6_54","article-title":"HScodeNet: combining hierarchical sequential and global spatial information of text for commodity HS code classification","volume":"12713","year":"2021","journal-title":"Advances in Knowledge Discovery and Data Mining"},{"key":"key2025041114531597100_ref012","unstructured":"Enigma (2018), \u201cUS imports - automated manifest system (AMS) shipments 2018\u201d, available at: https:\/\/aws.amazon.com\/ (accessed 1 August 2023)."},{"key":"key2025041114531597100_ref013","doi-asserted-by":"publisher","article-title":"CERT: contrastive self-supervised learning for language understanding","year":"2020","DOI":"10.48550\/arXiv.2005.12766"},{"issue":"90","key":"key2025041114531597100_ref014","first-page":"3133","article-title":"Do we need hundreds of classifiers to solve real world classification problems?","volume":"15","year":"2014","journal-title":"Journal of Machine Learning Research"},{"key":"key2025041114531597100_ref015","doi-asserted-by":"publisher","first-page":"6894","DOI":"10.18653\/v1\/2021.emnlp-main.552","article-title":"SimCSE: simple contrastive learning of sentence embeddings","year":"2021"},{"issue":"6","key":"key2025041114531597100_ref016","doi-asserted-by":"publisher","first-page":"964","DOI":"10.3390\/sym13060964","article-title":"A commodity classification framework based on machine learning for analysis of trade declaration","volume":"13","year":"2021","journal-title":"Symmetry"},{"key":"key2025041114531597100_ref017","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1705.00652","article-title":"Efficient Natural Language response suggestion for smart reply","year":"2017","journal-title":"arXiv preprint"},{"key":"key2025041114531597100_ref018","doi-asserted-by":"crossref","unstructured":"Henderson, M., Casanueva, I., Mrk\u0161i\u0107, N., Su, P.-H., Wen, T.-H. and Vuli\u0107, I. (2020), \u201cConveRT: efficient and accurate conversational representations from transformers\u201d, in Cohn, T., He, Y. and Liu, Y. (Eds), Findings of the Association for Computational Linguistics: EMNLP 2020, pp.\u00a02161-2174, findings-emnlp.196, doi: 10.18653\/v1\/2020.","DOI":"10.18653\/v1\/2020.findings-emnlp.196"},{"key":"key2025041114531597100_ref019","doi-asserted-by":"publisher","first-page":"1061","DOI":"10.1145\/3357384.3357891","article-title":"A Semantics aware random forest for text classification","year":"2019"},{"issue":"March","key":"key2025041114531597100_ref020","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2004.11362","article-title":"Supervised contrastive learning","volume":"10","year":"2021","journal-title":"arXiv"},{"issue":"2","key":"key2025041114531597100_ref021","doi-asserted-by":"publisher","first-page":"280","DOI":"10.1007\/s10559-022-00460-3","article-title":"Justification for the use of Cohen's Kappa statistic in experimental studies of NLP and text mining","volume":"58","year":"2022","journal-title":"Cybernetics and Systems Analysis"},{"issue":"4","key":"key2025041114531597100_ref022","doi-asserted-by":"publisher","first-page":"150","DOI":"10.3390\/info10040150","article-title":"Text classification algorithms: a survey","volume":"10","year":"2019","journal-title":"Information-an International Interdisciplinary Journal"},{"key":"key2025041114531597100_ref023","doi-asserted-by":"publisher","first-page":"1188","DOI":"10.48550\/arXiv.1405.4053","article-title":"Distributed representations of sentences and documents","year":"2014"},{"issue":"2","key":"key2025041114531597100_ref024","doi-asserted-by":"publisher","first-page":"25:1","DOI":"10.1145\/3635158","article-title":"Explainable product classification for customs","volume":"15","year":"2024","journal-title":"ACM Transactions on Intelligent Systems and Technology"},{"key":"key2025041114531597100_ref025","doi-asserted-by":"publisher","first-page":"7871","DOI":"10.18653\/v1\/2020.acl-main.703","article-title":"BART: denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension","year":"2020"},{"issue":"1","key":"key2025041114531597100_ref026","doi-asserted-by":"publisher","first-page":"127","DOI":"10.3905\/jfds.2020.1.047","article-title":"The best way to select features?","volume":"3","year":"2020","journal-title":"The Journal of Financial Data Science Winter 2021"},{"key":"key2025041114531597100_ref027","doi-asserted-by":"crossref","unstructured":"Moukafih, Y., Ghanem, A., Abidi, K., Sbihi, N., Ghogho, M. and Smaili, K. (2022), \u201cSimSCL: a simple fully-supervised contrastive learning framework for text representation\u201d, in Long, G., Yu, X. and Wang, S. (Eds), AI 2021: Advances in Artificial Intelligence, pp.\u00a0728-738, doi: 10.1007\/978-3-030-97546-3_59.","DOI":"10.1007\/978-3-030-97546-3_59"},{"issue":"13","key":"key2025041114531597100_ref028","doi-asserted-by":"publisher","first-page":"140","DOI":"10.59503\/29538009-2024.1.13-140","article-title":"Interpretable and generalizable HTS code classification framework","volume":"1","year":"2024","journal-title":"Economics, Finance and Accounting"},{"key":"key2025041114531597100_ref029","volume-title":"Harmonized System Code Classification Using Transfer Learning with Pre-trained Weights","year":"2021"},{"key":"key2025041114531597100_ref030","doi-asserted-by":"publisher","first-page":"1","DOI":"10.48550\/arXiv.1905.10626","article-title":"Rethinking softmax cross-entropy loss for adversarial robustness","year":"2020"},{"key":"key2025041114531597100_ref031","first-page":"8748","article-title":"Learning transferable visual models from Natural Language supervision","volume-title":"International Conference on Machine Learning","year":"2021"},{"issue":"December","key":"key2025041114531597100_ref032","article-title":"Sentence transformers NLI training readme, GitHub","volume":"19","year":"2022","journal-title":"GitHub"},{"key":"key2025041114531597100_ref033","doi-asserted-by":"publisher","first-page":"3982","DOI":"10.18653\/v1\/d19-1410","article-title":"Sentence-BERT: sentence embeddings using siamese BERT-networks","year":"2019"},{"key":"key2025041114531597100_ref034","doi-asserted-by":"publisher","first-page":"1135","DOI":"10.48550\/arXiv.1602.04938","article-title":"Why should I trust you?\u2019: explaining the predictions of any classifier","year":"2016"},{"key":"key2025041114531597100_ref035","doi-asserted-by":"publisher","article-title":"DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter","year":"2020","DOI":"10.48550\/arXiv.1910.01108"},{"key":"key2025041114531597100_ref036","article-title":"Decision support system for HS classification of commodities","year":"2004"},{"key":"key2025041114531597100_ref037","doi-asserted-by":"crossref","unstructured":"Spichakova, M. and Haav, H.-M. (2020), \u201cUsing machine learning for automated assessment of misclassification of goods for fraud detection\u201d, in Robal, T., Haav, H.-M., Penjam, J. and Matulevi\u010dius, R. (Eds), Databases and Information Systems, pp.\u00a0144-158, doi: 10.1007\/978-3-030-57672-1_12.","DOI":"10.1007\/978-3-030-57672-1_12"},{"key":"key2025041114531597100_ref038","doi-asserted-by":"publisher","article-title":"BEIR: a heterogeneous benchmark for zero-shot evaluation of information retrieval models","year":"2021","DOI":"10.48550\/arXiv.2104.08663"},{"issue":"1","key":"key2025041114531597100_ref039","doi-asserted-by":"publisher","first-page":"104","DOI":"10.1016\/j.ipm.2013.08.006","article-title":"The impact of preprocessing on text classification","volume":"50","year":"2014","journal-title":"Information Processing and Management"},{"key":"key2025041114531597100_ref040","doi-asserted-by":"publisher","first-page":"6000","DOI":"10.48550\/arXiv.1706.03762","article-title":"Attention is all you need","volume-title":"NIPS'17: Proceedings of the 31st International Conference on Neural Information Processing Systems","year":"2017"},{"key":"key2025041114531597100_ref041","doi-asserted-by":"publisher","first-page":"2146","DOI":"10.1109\/TASLP.2020.3008390","article-title":"SBERT-WK: a sentence embedding method by dissecting BERT-based word models","volume":"28","year":"2020","journal-title":"IEEE\/ACM Transactions on Audio, Speech, and Language Processing"},{"issue":"2","key":"key2025041114531597100_ref042","doi-asserted-by":"publisher","first-page":"187","DOI":"10.1007\/s40745-020-00253-5","article-title":"A comprehensive survey of loss functions in machine learning","volume":"9","year":"2022","journal-title":"Annals of Data Science"},{"key":"key2025041114531597100_ref043","doi-asserted-by":"publisher","article-title":"Importance of tuning hyperparameters of machine learning algorithms","year":"2020","DOI":"10.48550\/arXiv.2007.07588"},{"key":"key2025041114531597100_ref044","unstructured":"World Customs Organization (2013), \u201cHS classification handbook\u201d, available at: http:\/\/harmonizedsystem.wcoomdpublications.org\/pdfs\/WCOOMD_MSH_EN.pdf (accessed 1 January 2023)."},{"key":"key2025041114531597100_ref045","volume-title":"WCO Annual Report 2022-2023","author":"World Customs Organization","year":"2022"},{"issue":"4","key":"key2025041114531597100_ref046","doi-asserted-by":"publisher","first-page":"1","DOI":"10.48550\/arXiv.2012.15466","article-title":"CLEAR: contrastive learning for sentence representation","volume":"14","year":"2020","journal-title":"ACM Transactions on Intelligent Systems and Technology"},{"key":"key2025041114531597100_ref047","doi-asserted-by":"publisher","first-page":"4","DOI":"10.1016\/j.patrec.2017.09.018","article-title":"Kernelized support vector machine with deep learning: an efficient approach for extreme multiclass dataset","volume":"115","year":"2018","journal-title":"Pattern Recognition Letters"},{"key":"key2025041114531597100_ref048","unstructured":"Zauba (2016), \u201cIndia HS code data\u201d, available at: https:\/\/www.zauba.com\/ (accessed 1 August 2023)."},{"key":"key2025041114531597100_ref049","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2022.109836","article-title":"Harmonized system code prediction of import and export commodities based on Hybrid Convolutional Neural Network with Auxiliary Network","volume":"256","year":"2022","journal-title":"Knowledge-Based Systems"}],"container-title":["Data Technologies and Applications"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.emerald.com\/insight\/content\/doi\/10.1108\/DTA-01-2024-0052\/full\/xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.emerald.com\/insight\/content\/doi\/10.1108\/DTA-01-2024-0052\/full\/html","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,24]],"date-time":"2025-07-24T23:14:56Z","timestamp":1753398896000},"score":1,"resource":{"primary":{"URL":"http:\/\/www.emerald.com\/dta\/article\/59\/2\/276-301\/1246547"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,24]]},"references-count":49,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2024,12,24]]},"published-print":{"date-parts":[[2025,4,11]]}},"alternative-id":["10.1108\/DTA-01-2024-0052"],"URL":"https:\/\/doi.org\/10.1108\/dta-01-2024-0052","relation":{},"ISSN":["2514-9288","2514-9318"],"issn-type":[{"value":"2514-9288","type":"print"},{"value":"2514-9318","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,12,24]]}}}