Abstract Book of the 8th International Conference on New Trends in Management, Business and Economics
Year: 2026
[PDF]
Applying Artificial Neural Networks on the Study of Trading Patterns of Investors in the Stock Exchange of Thailand
Aekkachai Nittayagasetwat, Jiroj Buranasiri
ABSTRACT:
This paper applies artificial neural networks (ANNs) to study the effect of market variables including return and trading value of various investor types on their trading decision in the Stock Exchange of Thailand (SET) from 2009 to 2025. The SET has divided investors into four groups: retail, institutional, foreign, and proprietary trading investors. Prior to ANNs studies, this paper follows Bae, Yamada, and Ito (2006) that uses the generalized method of moments (GMM) analysis to study investors’ trading patterns. Consistent to the above paper, this paper finds that foreign, institutional, and proprietary trading investors are momentum traders who demand liquidity while retail investors are contrarian traders who supply liquidity. The ANN models are developed and used to compare with GMM models. The results show that ANNs outperform GMMs and have a higher coefficient of determination (R-squared) than GMMs. This research complements the existing body of knowledge on investors’ trading on their strategies and their respective impact on market variation.
Keywords: artificial neural network; generalized method of moments; the Stock Exchange of Thailand; momentum trader; trading pattern