Detecting Multielement Algorithmically Generated Domain Names Based on Adaptive Embedding Model

Table 3

Datasets in the experiments.

Dataset type	DGA description	Quantity
DGA	20 types of DGAs are collected, namely: Gameover, Dircrypt, Tinba, Necurs, Ramdo, Cryptolocker, Emotet, Corebot, Banjori, Qakbot, Rovnix, Kraken, Ramnit, Locky, Pykspa, Simda, Symmi, Virut, Matsnu, Suppobox	200,000
SDGA	8 types of SDGAs are collected, namely: DNL1, DNL2, DNL3, DNL4, 9ML1, 500KL1, 500KL2, 500KL3	88,000
ME	ME-DGA: generating domain names based on characters, numbers, special characters, two-character combinations, three-character combinations, and words, using a variety of DGAs ME-SDG: generating domain names based on characters, numbers, special characters, two-character combinations, three-character combinations, and words, using a variety of SDGAs	60000
Legitimate	Top 1 million domain names updated daily by Alexa from January 2009 to February 2019	Prepared in 4:1 quantity with dynamic domain names