Among the neglected tropical diseases (NTDs), arboviruses have a significant number of cases worldwide. In addition, the effects of the lockdown caused by COVID-19 contributed to the increase in cases of this type of virus. Its correct classification is a complex process due to the great similarity of symptoms between arboviruses. In addition, the lack of laboratory tests, especially in the interior of the country, is an additional obstacle to this problem. Given this context, this work proposes a machine learning model to assist health professionals in the clinical diagnosis of patients suspected of the most common arboviruses, Dengue and Chikungunya. For this, the model will make a multi-class classification between DENGUE, CHIKUNGUNYA and INCONCLUSIVE, to identify patients who do not have any of this two diseases. Eight models were initially tested and optimized through Grid Search technique, feature selection techniques were also performed to select the best attributes (symptoms and patient history) from the dataset. Finally, an evaluation of the selected attributes was also carried out with experts in the field to create a model that is more interpretable for health professionals. This work developed the GBM-Specialist, a Gradient Boosting model validated by experts, which achieved 76% sensitivity in the CHIKUNGUNYA class. Finally, a prototype, VALERIA, was developed so that the model can be used by healthcare professionals in real-world application.