{"paper_id":"2403.00777","title":"Combating Financial Crimes with Unsupervised Learning Techniques: Clustering and Dimensionality Reduction for Anti-Money Laundering","authors":["Ahmed N. Bakry","Almohammady S. Alsharkawy","Mohamed S. Farag","Kamal R. Raslan"],"paper_date":"2024-02-14","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2403.00777v1","page_url":"https://thequant.space/flowcharts/combating-financial-crimes-with-unsupervised-learning-techni/","scores":{"empirical_rigor":4,"math_complexity":6,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes combining agglomerative hierarchical clustering with dimensionality reduction techniques, especially Kernel PCA, to improve AML clustering performance.","quote":"In this paper, we investigate the effectiveness of combining clustering method agglomerative hierarchical clustering with four dimensionality reduction techniques -Independent Component Analysis (ICA), and Kernel Principal Component Analysis (KPCA), Singular Value Decomposition (SVD), Locality Preserving Projections (L","summary":"Anti-Money Laundering (AML) is a crucial task in ensuring the integrity of financial systems. One keychallenge in AML is identifying high-risk groups based on their behavior. Unsupervised learning, particularly clustering, is a promising solution for this task. However, the use of hundreds of…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"High-dimensional feature sets negatively impact clustering performance.","type":"model","stated":"explicit","quote":"However, the use of hundreds of features to describe behavior results in a high-dimensional dataset that negatively impacts clustering performance."},{"assumption":"Dimensionality reduction can capture the underlying structure of customer transaction data while minimizing information loss.","type":"data","stated":"implied","quote":"The objective is to identify the best DRTs that can accurately capture the underlying structure of the customer transaction data, while minimizing the loss of information."}]},"datasets":{"status":"available","items":[{"name":"Banking industry transaction dataset","provider":null,"kind":"other","access":"proprietary","period":null,"universe":"4099 companies","quote":"Transaction data from various customers were collected. The data was organized into yearly profiles for each customer, with a total of 4099 companies included in the dataset. Each company profile was characterized by 80 features, resulting in a dataset with 4099 rows and 80 columns."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"Kernel PCA combined with agglomerative hierarchical clustering outperforms other dimensionality reduction techniques across most validation indices.","kind":"performance","quote":"The experimental results demonstrate that KPCA outperforms other dimension reduction techniques when combined with agglomerative hierarchical clustering. This superiority is observed in the majority of situations, as confirmed by three distinct validation indices."},{"finding":"KPCA yields near‑perfect scores on Silhouette, Calinski‑Harabasz, and Davies‑Bouldin metrics, especially with a low number of features.","kind":"performance","quote":"KPCA, especially with a low number of features, consistently yields nearly perfect scores across all evaluation indices when combined with AHC."},{"finding":"Original high‑dimensional data without reduction performs relatively poorly compared to most dimensionality reduction techniques.","kind":"performance","quote":"When the original data, without any dimensionality reduction, is utilized, it exhibits relatively poor performance compared to most Dimensionality Reduction Techniques."}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2407.12044","title":"Credit Risk Assessment Model for UAE Commercial Banks: A Machine Learning Approach","paper_date":"2024-07-02","authors":["Aditya Saxena","Dr Parizad Dungore"],"empirical_rigor":4.5,"page_url":"https://thequant.space/flowcharts/credit-risk-assessment-model-for-uae-commercial-banks-a-machine-learning/"},"question":"Does dimensionality reduction improve model performance?","this_paper_finding":"Original high‑dimensional data without reduction performs relatively poorly compared to most dimensionality reduction techniques.","other_paper_finding":"Applying LDA increased the accuracy of all models, indicating that dimensionality reduction improves performance.","other_paper_quote":"There is a significant difference in accuracy before and after applying LDA on the dataset. This indicates that applying LDA on the dataset increases the performance of the machine learning models in terms of accuracy.","note":"Both findings report that applying dimensionality reduction (e.g., LDA) increases accuracy, while using original high‑dimensional data performs relatively poorly.","confidence":0.763}],"contradicts":[],"qualifies":[],"compared_with":21,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:14:43Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}