I have a few thousand files and each file has multiple "issues" after being deconstructed. They are labelled 0 or 1 for good or malicious. I think these "issues" can be a feature to determine if the file is malicious or not. Any suggestions on how to build the dataset? I am a begginer in ML