DeepSupp: Attention-Driven Correlation Pattern Analysis for Dynamic Time Series Support and Resistance Levels Identification

ArXiv ID: 2507.01971 “View on arXiv”

Authors: Boris Kriuk, Logic Ng, Zarif Al Hossain

Abstract

Support and resistance (SR) levels are central to technical analysis, guiding traders in entry, exit, and risk management. Despite widespread use, traditional SR identification methods often fail to adapt to the complexities of modern, volatile markets. Recent research has introduced machine learning techniques to address the following challenges, yet most focus on price prediction rather than structural level identification. This paper presents DeepSupp, a new deep learning approach for detecting financial support levels using multi-head attention mechanisms to analyze spatial correlations and market microstructure relationships. DeepSupp integrates advanced feature engineering, constructing dynamic correlation matrices that capture evolving market relationships, and employs an attention-based autoencoder for robust representation learning. The final support levels are extracted through unsupervised clustering, leveraging DBSCAN to identify significant price thresholds. Comprehensive evaluations on S&P 500 tickers demonstrate that DeepSupp outperforms six baseline methods, achieving state-of-the-art performance across six financial metrics, including essential support accuracy and market regime sensitivity. With consistent results across diverse market conditions, DeepSupp addresses critical gaps in SR level detection, offering a scalable and reliable solution for modern financial analysis. Our approach highlights the potential of attention-based architectures to uncover nuanced market patterns and improve technical trading strategies.

Keywords: Multi-head Attention, Technical Analysis, Support and Resistance, Unsupervised Clustering, Autoencoders, Equities

Complexity vs Empirical Score

• Math Complexity: 6.5/10
• Empirical Rigor: 7.0/10
• Quadrant: Holy Grail
• Why: The paper employs advanced mathematical concepts including multi-head attention mechanisms, Spearman rank correlation, and unsupervised clustering (DBSCAN), indicating high mathematical density. It also demonstrates high empirical rigor by performing comprehensive evaluations on S&P 500 tickers over a specific timeframe, comparing against six baseline methods, and utilizing multiple financial metrics for validation.

  flowchart TD
    A["Research Goal: Dynamic Support Level<br>Identification via Deep Learning"] --> B["Data: S&P 500 Tickers<br>Multi-dimensional Financial Data"]
    B --> C["Computational Process: Correlation Matrix<br>Construction & Attention Mechanism"]
    C --> D["Deep Learning Architecture:<br>Attention-Based Autoencoder"]
    D --> E["Unsupervised Clustering:<br>DBSCAN for Level Extraction"]
    E --> F["Key Outcomes: State-of-the-Art Performance<br>vs. 6 Baselines, 6 Metrics"]
    F --> G["Impact: Scalable SR Detection<br>Improved Market Regime Sensitivity"]