Algorithm 1: TON_IoT Dataset Proprocessing

Input: TON_IoT train-test Dataset (461,043 instances, 45 features) Output: Balanced Dataset with no missing data (461,043 instance, 37 features) $Array\leftarrow Dataset$ In ($Array$) remove $type$ label In ($Array$) remove $weird\_notice,weird\_add,weir{d}_{n}ame,ts$ features In ($Array$) remove $src\_ip,dst\_ip$ features label-encode $nonnumericalfeatures$