Enhanced K-Means Clustering Implementation for Web-Based Suspicious Profiles Detection

• Author(s): R. Prasanth Reddy; Nagavelli Yogender Nath; Gattu Ramya; Syed Abdul Haq
• Paper ID: 1712447
• Page: 548-555
• Published Date: 28-02-2024
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 7 Issue 8 February-2024

Abstract

Fraud detection often requires a hybrid approach combining both human expertise and AI techniques. While AI models can process large amounts of data and detect intricate patterns, human intervention is sometimes necessary to refine rules, especially when dealing with ambiguous or doubtful cases. This is particularly important when distinguishing between fraudulent and legitimate behavior, as subtle differences can exist. Another challenge is the multilingual nature of the web. Fraud can occur across different languages and cultural contexts, making it difficult to build comprehensive fraud detection models that account for all variations. In cases where linguistic resources (such as parsers or language models) are not readily available, ML tools can be adapted to learn from the data itself. However, for such models to function effectively, it is crucial to develop a rich set of features, ensuring that the data is representative and inclusive of the varied aspects of web-based fraud. This paper compares the performance of various clustering algorithms and develops an enhanced k-means algorithm utilizing the Minkowski metric. This modified algorithm, applied to an unlabeled dataset from an online dating site, effectively clusters users into authentic and suspicious categories. Supervised machine learning techniques are subsequently employed to validate the proposed model using another labeled online dating fraud dataset from the US. This methodology underscores the potential of combining traditional clustering methods with machine learning techniques to enhance anomaly detection across various domains, offering a practical tool for administrators and security experts.

Keywords

Deep Learning, Approaches Web-based Fraud, Fraud Detection, k-means algorithm

Citations

IRE Journals:
R. Prasanth Reddy, Nagavelli Yogender Nath, Gattu Ramya, Syed Abdul Haq "Enhanced K-Means Clustering Implementation for Web-Based Suspicious Profiles Detection" Iconic Research And Engineering Journals Volume 7 Issue 8 2024 Page 548-555 https://doi.org/10.64388/IREV7I8-1712447

IEEE:
R. Prasanth Reddy, Nagavelli Yogender Nath, Gattu Ramya, Syed Abdul Haq "Enhanced K-Means Clustering Implementation for Web-Based Suspicious Profiles Detection" Iconic Research And Engineering Journals, 7(8) https://doi.org/10.64388/IREV7I8-1712447