Article: Anomaly detection and pattern recognition methods for high-dimensional data Journal: International Journal of Computer Applications in Technology (IJCAT) 2026 Vol.78 No.2 pp.155 - 166 Abstract: High-dimensional data has become increasingly prevalent in a wide range of fields, including cybersecurity, finance, healthcare and industrial monitoring. However, the sparsity, redundancy and complex inter-feature relationships inherent in such data significantly complicate anomaly detection and pattern recognition tasks. Traditional machine learning methods often suffer from poor scalability and limited generalisation in high-dimensional settings. To address these limitations, this paper proposes a novel deep learning framework specifically designed for high-dimensional anomaly detection and pattern recognition. The proposed model introduces three key innovations. First, a hierarchical representation module is developed to extract multi-level semantic features by integrating adaptive kernel transformations with semantic-preserving aggregation strategies. This design improves the model's ability to capture both global patterns and local anomalies. Second, a dual-branch attention mechanism is introduced to jointly learn feature-level and instance-level relevance, enhancing the model's robustness to noise and irrelevant dimensions. Third, an interpretable anomaly scoring strategy is constructed based on prototype deviation in latent space, offering transparency and actionable insights for decision support. Extensive experiments are conducted on multiple real-world high-dimensional data sets. Results demonstrate that the proposed method consistently outperforms existing approaches in terms of accuracy, robustness and interpretability. Inderscience Publishers - linking academia, business and industry through research

Title: Anomaly detection and pattern recognition methods for high-dimensional data

Authors: Shengqi Wang

Addresses: School of Accounting, Shandong University of Financial and Economics, Jinan, Shandong, China

Abstract: High-dimensional data has become increasingly prevalent in a wide range of fields, including cybersecurity, finance, healthcare and industrial monitoring. However, the sparsity, redundancy and complex inter-feature relationships inherent in such data significantly complicate anomaly detection and pattern recognition tasks. Traditional machine learning methods often suffer from poor scalability and limited generalisation in high-dimensional settings. To address these limitations, this paper proposes a novel deep learning framework specifically designed for high-dimensional anomaly detection and pattern recognition. The proposed model introduces three key innovations. First, a hierarchical representation module is developed to extract multi-level semantic features by integrating adaptive kernel transformations with semantic-preserving aggregation strategies. This design improves the model's ability to capture both global patterns and local anomalies. Second, a dual-branch attention mechanism is introduced to jointly learn feature-level and instance-level relevance, enhancing the model's robustness to noise and irrelevant dimensions. Third, an interpretable anomaly scoring strategy is constructed based on prototype deviation in latent space, offering transparency and actionable insights for decision support. Extensive experiments are conducted on multiple real-world high-dimensional data sets. Results demonstrate that the proposed method consistently outperforms existing approaches in terms of accuracy, robustness and interpretability.

Keywords: high-dimensional data; anomaly detection; hierarchical representation learning; attention mechanism.

DOI: 10.1504/IJCAT.2026.151722

International Journal of Computer Applications in Technology, 2026 Vol.78 No.2, pp.155 - 166

Received: 19 May 2025
Accepted: 18 Sep 2025
Published online: 17 Feb 2026 *

Title: Anomaly detection and pattern recognition methods for high-dimensional data

Keep up-to-date