Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{209123,
        author = {Venkata Maheswara Rao Atchula and Dr. Pawan Kumar Pareek},
        title = {Machine Learning-Assisted Validation of Supertrend Trading Signals Across Multiple Timeframes},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {5},
        pages = {778-786},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=209123},
        abstract = {Financial trading systems frequently combine deterministic technical indicators with machine learning in an attempt to reduce low-quality signals while retaining the interpretability of a rule-based strategy. This study evaluates a machine learning-assisted validation framework applied to Supertrend trading signals. The experimental design uses seven years of historical market data, with four years allocated to training/development and three years reserved for unseen machine-learning test evaluation. The experiment output contains rule-based training runs and machine-learning test runs across five timeframes 1-minute, 3-minute, 5-minute, 15-minute and 1-hour and evaluates two confidence thresholds explicitly represented in the supplied results, 0.50 and 0.53. The analysis records total orders, wins, losses, win percentage, total profit, total loss, net realized result and maximum drawdown. Across the training configurations, the descriptive sum of net realized results is -118,653.94, whereas the corresponding machine-learning test configurations sum to +3,992.86 across independent runs. Within the paired threshold tests, the higher 0.53 threshold reduces order count in every tested timeframe and reduces maximum drawdown in four of the five timeframe comparisons, while its effect on net realized performance is mixed. The results indicate that confidence-threshold changes alter signal selectivity and risk characteristics, but do not produce a monotonic relationship with profitability. The study is therefore framed as an empirical validation of threshold sensitivity and out-of-sample behaviour rather than as evidence of a universally optimal Supertrend configuration.},
        keywords = {Supertrend, Machine Learning, Algorithmic Trading, Signal Validation, Technical Analysis, Confidence Threshold, Out-of-Sample Testing, Financial Markets.},
        month = {October},
        }