Using explainable artificial intelligence in buy and sell signals in the cryptocurrency market

This master's thesis investigates the use of explainable artificial intelligence (AI) to predict buy and sell signals in the cryptocurrency market. Within the scope of the study, machine and deep learning models are examined to predict the price movements of cryptocurrencies and generate buy and sell signals. For this purpose, the Long Short-Term Memory (LSTM) model was used. As a result of the experiments, it was determined that the LSTM model can predict future price movements with high accuracy. In order to further improve the performance of the LSTM model, hyperparameter optimization was performed and the results were improved. After the optimization process, the LSTM model produced the best results. In order to make the predictions of the LSTM model more understandable and to explain the model's decisions, the Locally Interpretable Model-Agnostic Explanations (LIME) method was used. LIME is used to explain certain predictions of the model and allow users to understand how the model works. In this way, both a powerful model and a method that can explain the model's decisions have been developed to more effectively predict buy and sell signals in the cryptocurrency market and to make decision processes transparent. As a result, this thesis has shown that explainable artificial intelligence can be used effectively in predicting buy and sell signals in the cryptocurrency market. The combination of the LSTM model and the LIME method provided successful results in terms of both high accuracy and model explainability. These findings aim to provide an important contribution to investors and financial analysts to make more informed and reliable trading decisions in the cryptocurrency market.