µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)
ÇѱÛÁ¦¸ñ(Korean Title) |
±ÔÄ¢±â¹ÝÀÇ ÇÊÅ͸µ ¹× BiLSTM ±â¹Ý Àΰø ½Å°æ¸Á ±â¹ýÀ» °áÇÕÇÑ ½ºÆÔ¸ÞÀÏ ÇÊÅ͸µ ±¸Çö |
¿µ¹®Á¦¸ñ(English Title) |
Implementation of Spam Mail Filtering Combining Rule Based Filtering and BiLSTM Based ANN |
ÀúÀÚ(Author) |
¼ÕÇѱâ
¹®Ã¶ÇÑ
ÃÖ¼ºÁØ
¹ÎÁرâ
Han-ki Son
Cheol-han Moon
Sung-jun Choe
Jun-Ki
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 38 NO. 01 PP. 0038 ~ 0050 (2022. 04) |
Çѱ۳»¿ë (Korean Abstract) |
ÀüÀÚ¿ìÆí(e-mail)Àº »ó¿ëÈµÈ ÀÌ·¡ ½ºÆÔ ¸ÞÀÏÀ» ¿Ã¹Ù¸£°í Á¤È®ÇÏ°Ô ºÐ·ùÇÏ´Â ¹®Á¦´Â ¿À·§µ¿¾È ÇÐ°è ¹× º¸¾È¾÷°èÀÇ Å« ¹®Á¦¿´´Ù. ±×µ¿¾È ¼ö¸¹Àº ¹æ¹ýµéÀÌ ³íÀǵǾúÀ¸¸ç, ÇÊÅÍ ±â¹ÝÀÇ ÇüÅÂ¼Ò ºÐ¼®°ú ÀÇ¹Ì ºÐ¼®¿¡ ´ëÇÑ ÀÀ¿ë¿¬±¸°¡ È°¹ßÈ÷ ¼öÇàµÇ¸é¼ ½ºÆÔ ÇÊÅ͸µ ¹æ¹ýÀÌ »ó´ç ºÎºÐ ÁøÀüµÇ¾ú´Ù. ÇÏÁö¸¸ ¿©ÀüÈ÷ ÇØ´ç ¹æ¹ýÀº »ç¶÷°ú µ¿ÀÏÇÑ ¼öÁØÀÇ ºÐ·ù¸¦ ¼öÇàÇÏ´Â µ¥¿¡´Â µµ´ÞÇÏÁö ¸øÇÏ°í ÀÖ´Ù. º» ³í¹®¿¡¼´Â ±ÔÄ¢±â¹ÝÀÇ ½ºÆÔ ÇÊÅÍ ¹æ¹ýÀ» °áÇÕÇÏ¿© ¾ÈÀü¼ºÀ» À¯ÁöÇÏ¸é¼ ±ÔÄ¢±â¹Ý ½ºÆÔ ÇÊÅÍ¿¡¼ Â÷´ÜÇÏÁö ¸øÇÏ´Â »õ·Î¿î ÆÐÅÏÀÇ ½ºÆÔ¸ÞÀÏ¿¡ ´ëÇÏ¿© Ãß°¡ÀûÀÎ BiLSTM(Bidirectional Long Short-Term Memory) ÇÊÅ͸¦ Àû¿ëÇÏ¿© ½ºÆÔ¸ÞÀÏ ÇÊÅ͸µÀÇ Á¤È®µµ¸¦ ³ôÀÌ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. |
¿µ¹®³»¿ë (English Abstract) |
Since e-mail has been commercialized, the problem of correctly and accurately classifying spam has long been a concern in academia and the security industry. In the meantime, many methods have been discussed, and significant progress has been made in spam filtering as filter-based, especially, applied research on stemming analyzers and semantic analysis have been actively conducted. However, it still has not reached the same level of classification as humans. In this paper, we propose a text-based spam filtering method based on BiLSTM (Bidirectional Long Short-Term Memory) combining the existing rule-based spam filtering method. |
Å°¿öµå(Keyword) |
¾ç¹æÇâ Àå´Ü±â ¸Þ¸ð¸® ½Å°æ¸Á
½ºÆÔ
ÇÜ
µö·¯´×
BiLSTM
SPAM
HAM
Deep Learning
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|