These seven .csv files contain the logarithms of the stock return volatitilies, as well as the MDA sections of the FIN10K dataset used to replicate the experiment done in Tsai and Wang (2017), but for FETILDA. They are directly used by the Python scripts for training, validation, and testing. Among them, train-results.csv is used for training and testing, and the rest of the six .csv files are used for testing for the years 2001-2006, respectively.