Method of Profanity Detection Using Word Embedding and LSTM

Table 2

FastText model vector training result [25].

Character vector substring	n = 2	<ieung = [0.121, −0.187, 0.002, …, −0.028] ieung ah = [−0.071, −0.158, 0.101, …, −0.004]…
	n = 3	<ieung ah = [−0.095, 0.189, 0.082, …, 0.113] ieung ah rieul = [0.108, −0.158, 0.462, …, −0.054]…
	n = 4	<ieung ah rieul = [0.003, −0.139, 0.372, …, 0.319] ieung ah rieul rieul = [0.091, 0.133, 0.421, …, 0.104]…
Character vector string	alladin = <ieung + ieung ah + … + ieung ah rieul + … + digeut ee nieun> = [0.234, −0.322, 0.401, …, 0.159]