Análise comparativa entre algoritmos de predição de preço para o Bitcoin

The financial market has captured attention for centuries, including assets such as rice, gold, and stocks. Price fluctuations can reflect significant events, and the ability to predict these changes often determines the success or failure of businesses. Over the years, techniques like technical and fundamental analysis have been developed. Recently, artificial neural networks, particularly with the integration of Long Short-Term Memory (LSTM) cells, have gained prominence for identifying previously unnoticed patterns in time series data. Simultaneously, volatile markets like cryptocurrencies have emerged, driven by Bitcoin and Blockchain technology introduced in 2008. These markets offer alternatives to counter inflation and centralized control. This research aims to analyze and compare algorithms on a real-world dataset to predict asset behavior in volatile markets, focusing on Bitcoin. Based on recent research, the results reveal that statistical models achieved better performance for data sampled at fifteen-minute intervals.