Aplicación de modelo transformer para opciones binarias usando análisis técnico en el mercado Forex

Machine learning techniques have become great tools for predicting patterns, we have decided to use said techniques and apply them as a tool for predicting the behaviour of the forex market; for this purpose we use the relatively recent Transformer machine learning model to analyse the data obtained from the IQoption API to predict the EUR/USD currency behaviour though binary options. Through the implementation of the Transformer model, we obtained an accuracy rate of around 65% from candle to candle; although this can't be considered a success in terms of machine learning model implementations, considering the complexity of the data analysed and the context in which it's used, the execution of the model is a success on practical levels of binary option investments as long as the market is paying more than 110% of the invested capital, getting profits with a large operations volume of 0.15% net earnings. In conclusion, the Transformer model proved to be an effective mean to demonstrate that machine learning models can be of great use in the context of financial marketing.

• Autores: Méndez Macea, Johan Andrés; Roa Pereira, Sergio Alejandro; Espinoza Mejia, Steven Eduardo; García Yepes, Gabriel Jesús
• Fecha de publicación: 2022
• Institución: Universidad del Norte
• Repositorio: Repositorio Uninorte
• Idioma: spa
• OAI Identifier: oai:manglar.uninorte.edu.co:10584/11223
• Acceso en línea: http://hdl.handle.net/10584/11223
• Palabra clave: Machine Learning; Technical Analysis; Transformer Model; Binary Options; Forex; Aprendizaje de Máquina; Análisis técnico; Forex; Modelo Transformer; Opciones Binarias
• License: Universidad del Norte