Behavioral Finance Model: Google Trends and LSTM in Forecasting Retail Assets on the Brazilian Stock Exchange

• Lucas D. M. Mendonça UFPA
• Diego L. Cardoso UFPA

DOI: https://doi.org/10.5753/eniac.2026.29960

Resumo

Social data mining, such as Google searches, has been increasingly explored in financial time series forecasting. However, the literature frequently ignores the semantic noise inherent to retail search queries. This paper investigates the impact of different Google Trends syntactic categories on neural network-based predictive models for Brazilian Stock Exchange (B3) retail assets, aiming to mitigate this noise. Analyzing error and volatility metrics across nine assets, our results suggest that the predictive value of online attention is not universal but instead depends on the type of query used. In 77.8% of the analyzed assets, the incorporation of Google Trends data outperformed the pure OHLC model, most notably for AMER3, where searching for the commercial name reduced the MAPE from 643.18% to 99.57%. Based on these observations, we propose an exploratory taxonomy of four categories and argue that financial predictive architectures benefit from semantic parameterization adapted to each asset’s characteristics.

Referências

Andrade, C. F. O., Herrera, P. A. L., and Bonilla, D. A. G. (2025). Google trends as an early signal in international flower trade. Ornamental Horticulture, 31:e312959.

Andrei, D. and Hasler, M. (2014). Investor attention and stock market volatility. The Review of Financial Studies, 28(1):33–72.

Bantis, E., Clements, M. P., and Urquhart, A. (2023). Forecasting gdp growth rates in the united states and brazil using google trends. International Journal of Forecasting, 39(4):1909–1924.

Barber, B. M. and Odean, T. (2007). All that glitters: The effect of attention and news on the buying behavior of individual and institutional investors. The Review of Financial Studies, 21(2):785–818.

Bolfarine, H. and de Oliveira Bussab, W. (2005). Elementos de amostragem. Editora Blucher.

Cebrián, E. and Domenech, J. (2023). Is google trends a quality data source? Applied Economics Letters, 30(6):811–815.

Churi, A., Chakraborty, D., Khatwani, R., Pinto, G., Shah, P., and Sekhar, R. (2023). Stock price prediction using deep learning and sentiment analysis. In 2023 2nd International Conference on Futuristic Technologies (INCOFT), pages 1–6. IEEE.

Dhanawansa, V., Samarasinghe, P., Yogarajah, P., Gardiner, B., and Karunasena, A. (2022). Comparative study of deep learning parameter selection for multi-output regression on head pose estimation. In 2022 IEEE International Conference on Industrial Technology (ICIT), pages 1–6.

Fahrudin, T., Asniar, and Faizul Ula, M. (2022). Comparison the new computer sciences study programs in indonesia using pytrends. In 2022 International Conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS), pages 379–383.

Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The journal of Finance, 25(2):383–417.

Gopali, S., Siami-Namini, S., Abri, F., and Namin, A. S. (2024). A comparative multivariate analysis of var and deep learning-based models for forecasting volatile time series data. IEEE Access, 12:155423–155436.

Hamid, A. and Heiden, M. (2015). Forecasting volatility with empirical similarity and google trends. Journal of Economic Behavior & Organization, 117:62–81.

Leippold, M., Wang, Q., and Zhou, W. (2022). Machine learning in the chinese stock market. Journal of Financial Economics, 145(2):64–82.

Mesquita, L. A., Ranieri, C. M., and Ueyama, J. (2023). Artificial neural networks applied to time series for flood prediction. In Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), pages 712–725. SBC.

Petropoulos, A., Siakoulis, V., Stavroulakis, E., Lazaris, P., and Vlachogiannakis, N. (2022). Employing google trends and deep learning in forecasting financial market turbulence. Journal of Behavioral Finance, 23(3):353–365.

Shiller, R. J. (2003). From efficient markets theory to behavioral finance. Journal of Economic Perspectives, 17(1):83–104.

Suthiponpisal, V. and Tancharoen, D. (2024). A comparative evaluation of noise reduction versus data normalization techniques in stock market prediction using transformer models. In 2024 8th International Conference on Information Technology (InCIT), pages 775–780.

Taleb, N. N. (2010). The Black Swan: The Impact of the Highly Improbable, with a New Section on Robustness and Fragility, volume 2. Random house trade paperbacks.

Wang, C.-H. (2022). Considering economic indicators and dynamic channel interactions to conduct sales forecasting for retail sectors. Computers & Industrial Engineering, 165:107965.