Dietary supplements for use in sport activities, for improve body composition and reducing body weight is highly consumed in Brazil, and caffeine is a frequent ingredient of these products. Knowledge discovery in databases (KDD) is a technique which may be used to detect labeling non-conformities more efficiently than traditional methods. KDD involves the selection of data variables and databases, data preprocessing, data mining and data interpretation. Data mining encompasses a number of statistical techniques including cluster analysis, link analysis and feature selection. This methodology was utilized to evaluate the labeling information in thermogenic dietary supplements for the first time. Label parameters were evaluated according to Brazilian Regulatory Agency (ANVISA) recommendations. A total of 70 products were analyzed. The information contained in the product labels was paired with 43 labeling parameters for a total of 3,010 data. The results showed different types of mismatches to the regulations. This methodology allowed to correlate a large volume of data and to evaluate in a more comprehensive way the information contained in the labels for consumer safety. The results suggest that the consumption of products according to label instructions may pose health risks.