Modelo da estratégia Inner Circle Trade usando Inteligência Artificial

Abstract

The application of Artificial Intelligence (AI) in the financial sector, especially in asset management and finance, has become increasingly common due to the vast availability of data and access to computational power. However, lack of practice or knowledge in this field can result in losses for those involved. The objective of this study is to develop a model called TradeClassifier capable of making decisions about the appropriate time to operate in Long or Short on a Forex market asset, using the Inner Circle Trade (ICT) strategy and Artificial Intelligence techniques. The model aims to assist both inexperienced and experienced traders. The Machine Learning techniques used for classification include Support Vector Machine (SVM), Decision Tree, Random Forest, and Neural Networks, which are widely employed in academic works for prediction and classification. Python programming language was chosen due to the available libraries and its popularity in the finance sector and AI project development. A supervised learning approach was adopted for classification, as the model was given a database of the EUR/USD asset (Euro vs. US Dollar) and classifies the moments to operate. The model accuracy results reached approximately 55% accuracy using the GridSearchCV() technique for parameter optimization. The confusion matrix was used to analyze the classifications made by the models. Additionally, trading simulations were conducted based on the predictions generated by models that showed profitable results, namely the SVM and Neural Networks models. The SVM model achieved a profit of 79,235.23%, and the Neural Network model accumulated 63,450.64% profit.