Paper: arXiv 2507.01971

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.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: 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.

Research Flowchart

  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"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes DeepSupp, a deep learning model using multi-head attention on spatial correlations to detect support levels.

Source sentence
We introduce DeepSupp, a new deep learning approach that utilizes multi-head attention mechanisms on spatial correlations to identify meaningful support levels.

Assumes 2

  • Market correlations are non‑static and change over time, especially during stress periods. market
    Source sentence
    Rather than assuming static relationships, our approach recognizes market correlations change over time, particularly during periods of market stress or rigid changes.
  • Spearman rank correlation adequately captures non‑linear relationships between market variables. statistical
    Source sentence
    We employ Spearman rank correlation to capture non‑linear relationships using sliding windows of length n = 32

Data

  • S&P 500 tickers market-prices unknown
    S&P 500 constituents
    Source sentence
    Comprehensive evaluations on S&P 500 tickers demonstrate that DeepSupp outperforms six baseline methods

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • DeepSupp achieves the highest overall performance score of 0.554 ± 0.039, with the lowest variability across all evaluated metrics. performance
    Source sentence
    score of 0.554 with the lowest variability at ±0.039, demonstrating superior and consistent performance across the weighted combination of all metrics evaluated on tickers.
  • DeepSupp shows balanced performance across multiple metrics, though it does not achieve the top score on any single metric. performance
    Source sentence
    DeepSupp’s competitive advantage lies in its balanced performance across multiple metrics combined with the lowest performance variability.
  • Local Minima detection attains the highest support accuracy (0.603) while HMM achieves the best price proximity (0.826) and support duration stability (0.859). performance
    Source sentence
    Local Minima detection achieved the highest support accuracy at 0.603

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 26 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (23)

  1. The use of technical analysis in the foreign exchange market (1992) — link
  2. Identifying and evaluating horizontal support and resistance levels: an empirical study on US stock markets (2012)
  3. Stock identification methods: an overview (2014)
  4. Testing Horizontal Support and Resistance Zones on Cryptocurrencies (2023)
  5. Systematic identification of signaling pathways with potential to confer anticancer drug resistance (2014)
  6. Support for resistance: technical analysis and intraday exchange rates (2000)
  7. Automated support and resistance (1998)
  8. Evolutionary optimized stock support-resistance line detection for algorithmic trading systems (2019)
and 15 more
  1. Optimal prediction of resistance and support levels (2016) — link
  2. Currency orders and exchange-rate dynamics: explaining the success of technical analysis (2001) — link
  3. Memory effects in stock price dynamics: evidences of technical trading (2014) — link
  4. Simple technical trading rules and the stochastic properties of stock returns (1992) — link
  5. The use of technical analysis by fund managers: International evidence (2010) — link
  6. Technical analysis and individual investors (2014) — link
  7. Technical Analysis of Stock Trends (1948)
  8. Technical Analysis of the Financial Markets: A Comprehensive Guide to Trading Methods and Applications (1999)
  9. Fibonacci Applications and Strategies for Traders (1993)
  10. Trading Chaos: Maximize Profits with Proven Technical Techniques (1995)
  11. Foundations of Technical Analysis: Computational Algorithms, Statistical Inference, and Empirical Implementation (2000)
  12. Quantile Regression (2005)
  13. Stock Market Forecasting Using Hidden Markov Model: A New Approach (2005) — link
  14. Financial Time Series Forecasting – A Deep Learning Approach (2017) — link
  15. Using Deep Learning to Detect Price Change Indications in Financial Markets (2017) — link

0 of 23 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.