With the significant growth of the cryptocurrency market, predicting its short-term movements has become a significant challenge, requiring methodologies that combine both on-chain data and traditional financial indicators to improve prediction accuracy. This work proposes an integrated approach to predicting short-term movements in the cryptocurrency market by combining on-chain data with traditional financial indicators within the framework of the Rainbow Bitcoin Chart. The methodology encompasses data preparation, model development, and performance evaluation, aiming to build a robust and accurate tool for classifying market trends. By integrating blockchain structural analysis with financial signals, this approach seeks to enhance forecast accuracy and support more informed decision-making. The results show that the normalized model with 14-day windows achieved the best performance among the tested configurations, reaching an accuracy of over 75%.