03746ntm a22005777i 4500 000725084 CZ-PrVSE 20250824103944.0 m d cr n|||||||||| 250824s2025 xr fsbm 000 0 eng d NEZPRACOVANÝ IMPORT ABA006 cze ABA006 ABA006 rda Kohout, Vít ISIS:162535 dis Implementace obchodovacího algoritmu se zaměření na objemové profilování eng Implementation of a trading algorithm with a focus on volume profiles / Vít Kohout 2025 ?? stran : digital, PDF soubor Vedoucí práce: Václav Zeman Bakalářská práce (Bc.)—Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky, 2025 Obsahuje bibliografii Textový (vysokoškolská kvalifikační práce) Rok obhajoby 2025 Thesis investigates the incremental value of integrating advanced analytics into classical intraday trading strategies. To systematically assess the impact of increasing analytical complexity, a tiered framework of nine strategies was developed spanning baseline (price-only), volume-enhanced, and deep learning-enhanced archetypes. Temporal Fusion Transformer model is implemented to provide 15-minute ahead volume forecasts for the most advanced strategies and everything is rigorously backtested on NASDAQ-100 securities. The empirical results demonstrate a key trade-off: while incorporating volume-based parameters did not significantly increase profitability, it offered a powerful, statistically significant advantage in risk management by consistently reducing maximum drawdowns. Conversely, the hypothesis that a sophisticated transformer-based forecast would further improve performance was rejected, as the strategies operating on predicted volume failed to outperform their simpler, heuristic-based counterparts. Analysis also validates that strategy performance is highly regime-dependent, particularly for mean-reversion and momentum approaches. It is concluded that the primary contribution of adding volume-based complexity in this context is defensive, serving to preserve capital rather than amplify return, and highlight a critical trade-off between the benefits and implementation costs of deploying advanced predictive models in intraday trading strategies on higher frequency data. Způsob přístupu: Internet data analytics [obor bakal. práce] bakalářské práce fd132403 czenas bachelor's theses eczenas intraday trading strategies market risk management quantitative finance time series analysis transformer models algorithmic trading machine learning volume forecasting deep learning Zeman, Václav ISIS:96325 ths Chudán, David js20100609002 opn Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky kn20010709399 dgg https://insis.vse.cz/zp/88190/podrobnosti VŠKP v InSIS https://insis.vse.cz/zp/88190 Hlavní práce https://insis.vse.cz/zp/88190/posudek/vedouci Hodnocení vedoucího https://insis.vse.cz/zp/88190/posudek/oponent/85730 Oponentura https://insis.vse.cz/zp/88190/priloha/32670 Přiloha k práci https://insis.vse.cz/zp/88190/podrobnosti dc:identifier NEPOSILAT VSKP vse88190 250822 88190