• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) ÀüÆÄ ³×Æ®¿öÅ© Ư¼º ÃßÃâÀ» ÅëÇÑ ¼Ò¼È ¹Ìµð¾î ¼Ó ·ç¸Ó ºÐ·ù
¿µ¹®Á¦¸ñ(English Title) Feature Extraction from a Diffusion Network for Rumor Classification in Social Media
ÀúÀÚ(Author) ÃÖÁöÈ£   Á¤½ÃÇö   ±èÁ¾±Ç   Jiho Choi   Sihyun Jeong   Chong-kwon Kim  
¿ø¹®¼ö·Ïó(Citation) VOL 26 NO. 02 PP. 0110 ~ 0115 (2020. 02)
Çѱ۳»¿ë
(Korean Abstract)
¼Ò¼È ¹Ìµð¾î´Â Çö´ë »çȸ¿¡¼­ °­·ÂÇÑ Á¤º¸ Àü´Þ·ÂÀ» °¡Áø ¸Åü Áß Çϳª·Î À̸¦ ÅëÇØ ·ç¸Ó(rumor)°¡ ÀüÆÄµÉ °æ¿ì Å« »çȸÀû ¿µÇâÀ» ³¢Ä¥ ¼ö ÀÖ´Ù. ¼Ò¼È ¹Ìµð¾î¸¦ ÅëÇØ È®»êµÈ ·ç¸Ó´Â ¶§·Ð Çö½Ç¿¡ Å« ÇÇÇظ¦ ¾ß±âÇϹǷΠ·ç¸Ó¸¦ ºÐ·ùÇÏ¿© °ÅÁþ Á¤º¸ÀÇ ÀüÆĸ¦ ¸·´Â ³ë·ÂÀº Áß¿ä½ÃµÇ°í ÀÖ´Ù. º» ³í¹®¿¡¼­´Â ¼Ò¼È ¹Ìµð¾î ¼Ó ·ç¸ÓÀÇ ÀÚÃë(trace)¸¦ ÀÌ¿ëÇÏ¿© ÀüÆÄ ³×Æ®¿öÅ©(diffusion network)¸¦ ±¸¼ºÇÏ°í À̷κÎÅÍ ±¸Á¶Àû (structural), ½Ã°£Àû(temporal), »çȸÀû(social) Ư¼ºÀ» ÃßÃâÇÏ¿© ·ç¸Ó ºÐ·ù¸¦ À§ÇÑ ±â°èÇнÀ ºÐ·ù±â ÇнÀ¹ýÀ» Á¦¾ÈÇÑ´Ù. ½ÇÁ¦ Æ®À§ÅÍ µ¥ÀÌÅÍ¿¡¼­ ÃßÃâµÈ Ư¼ºÀ» ¹ÙÅÁÀ¸·Î ¾Ó»óºí ºÐ·ù±â¸¦ ÇнÀ½ÃŲ °á°ú Áø½ÇÀÎ ·ç¸Ó, °ÅÁþÀÎ ·ç¸Ó, Áõ¸íµÇÁö ¾ÊÀº ¸Þ½ÃÁö, ·ç¸Ó°¡ ¾Æ´Ñ ¸Þ½ÃÁö±îÁö 4°³ÀÇ ·¹À̺í·Î ÀÌ·ç¾îÁø ³×Æ®¿öÅ© ºÐ·ù ¹®Á¦¿¡¼­ 69.3% ÀÌ»óÀÇ Á¤È®µµ¿Í 0.686 F1 Á¡¼öÀÇ ¼º´ÉÀ» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Social media has become one of the most powerful mediums for information dissemination in modern society. However, it can cause great social impact if it facilitates the spread of rumors. Efforts to classify rumors and prevent the spread of false information are imperative because rumors can cause great damage in reality. In this paper, we proposed a method of training machine learning classifiers for rumor classification tasks by modeling the diffusion network from tracing rumors in social media. We extracted structural, temporal, and social characteristics from the diffusion network and used them as features to classify rumors. Based on the characteristics extracted from the real-world Twitter data, the study showed that the ensemble classifier had 69.3% accuracy and a 0.686 F1 score in network classification tasks with four labels: true rumors, false rumors, unverified messages, and non-rumors
Å°¿öµå(Keyword) ¼Ò¼È ¹Ìµð¾î ºÐ¼®   ·ç¸Ó ºÐ·ù   ÀüÆÄ ³×Æ®¿öÅ©   ³×Æ®¿öÅ© ºÐ·ù   Ư¼º ÃßÃâ   social network analysis   rumor classification   diffusion network   network classification   feature extraction  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå