Adaptive Multidimensional Fusion Network with Dynamic Decision Trees for Financial Market Trend Forecasting
Abstract
This study proposes an Adaptive Multidimensional Fusion Network (AMFN) for financial market trend forecasting. The model integrates heterogeneous data sources through a multidimensional data fusion module, combining historical price and volume data with external information such as macroeconomic indicators and sentiment indices. An adaptive temporal processing module is employed to model time-varying dependencies and regime shifts in market behavior, while a dynamic decision-tree prediction module captures nonlinear patterns in the fused representations. Experiments are conducted on multiple financial datasets, including the S&P 500 Index, China A-share market, and Gold Futures, using a rolling time-window evaluation to avoid information leakage. The AMFN model achieves lower MSE and MAE and higher R² than traditional SVM and LSTM baselines, with up to 24.4% relative improvement in forecasting accuracy. These results demonstrate that AMFN provides interpretable, stable, and robust trend predictions across diverse market environments.References
Kumar A, Alsadoon A, Prasad PWC, Abdullah S, Rashid TA, Pham DTH, et al. Generative adversarial network (GAN) and enhanced root mean square error (ERMSE): deep learning for stock price movement prediction. Multimedia Tools and Applications. 2022; 81(3):3995-4013. DOI: 10.1007/s11042-021-11670-w
Huang Y, Wan JS. Hierarchical analysis of Chinese financial market based on manifold structure. Annals of Operations Research. 2022; 315(2):1135-50. DOI: 10.1007/s10479-021-03959-8
Schmidhuber C. Financial markets and the phase transition between water and steam. Physica a-Statistical Mechanics and Its Applications. 2022; 592. DOI: 10.1016/j.physa.2022.126873
Shih KH, Wang YH, Kao IC, Lai FM. Forecasting ETF performance: A comparative study of deep learning models and the fama-french three-factor model. Mathematics. 2024; 12(19). DOI: 10.3390/math12193158
Ziolo M, Bak I, Spoz A. The role of financial markets in energy transitions. Energies. 2024; 17(24). DOI: 10.3390/en17246315
Popov E, Veretennikova A, Fedoreev S. The model of OTC securities market transformation in the context of asset tokenization. Mathematics. 2022; 10(19). DOI: 10.3390/math10193441
Lei WN, Li Z, Mei DZ. Financial crisis, labor market frictions, and economic volatility. Plos One. 2023; 18(9). DOI: 10.1371/journal.pone.0291106
Ni YS. Navigating energy and financial markets: A review of technical analysis used and further investigation from various perspectives. Energies. 2024; 17(12). DOI: 10.3390/en17122942
Zhang X. RETRACTION: Analysis of financial market trend based on autoregressive conditional heteroscedastic model and BP neural network prediction. (Retraction of Vol 39, Pg 5845, 2020). Journal of Intelligent & Fuzzy Systems. 2021; 41(5):5771-.
Hao YP, Gao Q. Predicting the trend of stock market index using the hybrid neural network based on multiple time scale feature learning. Applied Sciences-Basel. 2020; 10(11). DOI: 10.3390/app10113961
Ju CB, Chen AP. Identifying Financial market trend reversal behavior with structures of price activities based on deep learning methods. IEEE Access. 2022; 10:12853-65. DOI: 10.1109/access.2022.3146371
Liu H, Long ZH. An improved deep learning model for predicting stock market price time series. Digital Signal Processing. 2020; 102. DOI: 10.1016/j.dsp.2020.102741
Long JW, Chen ZP, He WB, Wu TY, Ren JT. An integrated framework of deep learning and knowledge graph for prediction of stock price trend: An application in Chinese stock exchange market. Applied Soft Computing. 2020; 91. DOI: 10.1016 /j.asoc.2020.106205
Wu DM, Wang XL, Su JY, Tang BZ, Wu SC. A labeling method for financial time series prediction based on trends. Entropy. 2020; 22(10). DOI: 10.3390/e22101162
Fritz M, Gries T, Wiechers L. An early indicator for anomalous stock market performance. Quantitative Finance. 2024; 24(1):105-18. DOI: 10.1080/14697688.2023.2281529
Jin ZB, Jin YX, Chen ZY. Empirical mode decomposition using deep learning model for financial market forecasting. Peerj Computer Science. 2022; 8. DOI: 10.7717/peerj-cs.1076
Lee MC, Chang JW, Hung JSC, Chen BL. Exploring the effectiveness of deep neural networks with technical analysis applied to stock market prediction. Computer Science and Information Systems. 2021; 18(2):401-18. DOI: 10.2298/csis200301002l
Shen JY, Shafiq MO. Short-term stock market price trend prediction using a comprehensive deep learning system. Journal of Big Data. 2020; 7(1). DOI: 10.1186/s40537-020-00333-6
Neuman Y, Cohen Y. Unveiling herd behavior in financial markets. Journal of Statistical Mechanics-Theory and Experiment. 2023; 2023(8). DOI: 10.1088/1742-5468/aceef0
Huang WC, Chen CT, Lee C, Kuo FH, Huang SH. Attentive gated graph sequence neural network-based time-series information fusion for financial trading. Information Fusion. 2023; 91:261-76. DOI: 10.1016/j.inffus.2022.10.006
Afilipoaei A, Carrero G. A mathematical model of financial bubbles: a behavioral approach. Mathematics. 2023; 11(19). DOI: 10.3390/math11194102
Fang Z, Wang SY. Boosting financial market prediction accuracy with deep learning and big data: Introducing the CCL model. Journal of Organizational and End User Computing. 2024; 36(1). DOI: 10.4018/joeuc.358454
Wang RY, Xie YJ, Chen H, Jia GZ. Analyzing the impact of COVID-19 on the cross-correlations between financial search engine data and movie box office. Fluctuation and Noise Letters. 2021; 20(05). DOI: 10.1142/ s0219477521500218
Wang CH, Liang H, Wang B, Cui XX, Xu YW. MG-Conv: A spatiotemporal multi-graph convolutional neural network for stock market index trend prediction. Computers & Electrical Engineering. 2022; 103. DOI: 10.1016/j.compeleceng.2022.108285
Liang MX, Wang XL, Wu SC. Improving stock trend prediction through financial time series classification and temporal correlation analysis based on aligning change point. Soft Computing. 2023; 27(7):3655-72. DOI: 10.1007/s00500-022- 07630-7
DOI:
https://doi.org/10.31449/inf.v50i5.10720Downloads
Published
How to Cite
Issue
Section
License
Authors retain copyright in their work. By submitting to and publishing with Informatica, authors grant the publisher (Slovene Society Informatika) the non-exclusive right to publish, reproduce, and distribute the article and to identify itself as the original publisher.
All articles are published under the Creative Commons Attribution license CC BY 3.0. Under this license, others may share and adapt the work for any purpose, provided appropriate credit is given and changes (if any) are indicated.
Authors may deposit and share the submitted version, accepted manuscript, and published version, provided the original publication in Informatica is properly cited.
