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

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ÇÐȸÁö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ÇÐȸÁö > µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) ·ÎÁö½ºÆ½ ȸ±ÍºÐ¼®°ú Àΰø½Å°æ¸ÁÀ» ÀÌ¿ëÇÑ À¯¹æ¾Ï ºÐ·ù ¸ðµ¨ ºñ±³¿¬±¸
¿µ¹®Á¦¸ñ(English Title) A Comparative Study on the Classification Model of Breast Cancer using Logistic Regression and Artificial Neural Networks
ÀúÀÚ(Author) ±è³«ÀÏ   ÀÌö±â   ÀÌ¿ì±â   Nak-Il Kim   Charles Cheolgi Lee   Wookey Lee  
¿ø¹®¼ö·Ïó(Citation) VOL 35 NO. 02 PP. 0139 ~ 0153 (2019. 08)
Çѱ۳»¿ë
(Korean Abstract)
Àΰø½Å°æ¸ÁÀº À¯¿ëÇÑ ¸Ó½Å·¯´× ±â¹ýÀ¸·Î È°¿ëµÇ°í ÀÖÁö¸¸ ±âÁ¸ Åë°èºÐ¼®±â¹ý°ú ºñ±³ÇßÀ» ¶§, ÇÑ°èÁ¡µµ ÀÖ´Ù. ºÐ¼®°á°ú¸¦ Çؼ®ÇϱⰡ ¾î·Æ±â ¶§¹®ÀÌ´Ù. ƯÈ÷, º¯¼ö °£ÀÇ °ü°è ÆľÇÀÌ Áß¿äÇÑ º¸°ÇÀÇ·áÁ¤º¸ ¿¬±¸¿¡¼­´Â ÀÌ·¯ÇÑ ´ÜÁ¡ÀÌ µÎµå·¯Áø´Ù. º» ¿¬±¸ÀÇ ¸ñÀûÀº ÀüÅëÀûÀÎ Åë°è±â¹ý°ú ¸Ó½Å·¯´×±â¹ýÀ» ºñ±³ÇÏ¿© °ü·Ã ÀÖ´Â ÀÎÀÚ¸¦ Á¶»çÇÏ°í ¾î¶² ºÐ¼®±â¹ýÀÌ ´õ ÀûÀýÇÑÁö ÆÇ´ÜÇÏ´Â µ¥ ÀÖ´Ù. ±âÁ¸ À¯¹æ¾ÏÀÇ ¹ßº´ ¿äÀÎÀº ½Ä½À°ü, ºñ¸¸, À½ÁÖ, ¹æ»ç¼± µîÀ¸·Î ¾Ë·ÁÁ® ÀÖÀ¸¸ç, µ¥ÀÌÅÍ ºÐ¼®±â¹ýÀ¸·Î´Â ·ÎÁö½ºÆ½ ȸ±ÍºÐ¼®ÀÌ ÀÚÁÖ »ç¿ëµÇ¾ú´Ù. ·ÎÁö½ºÆ½ ȸ±ÍºÐ¼®Àº °á°ú¸¦ Çؼ®ÇϱⰡ ¿ëÀÌÇÑ °ÍÀÌ ÀåÁ¡ÀÌÁö¸¸ ¼³Á¤ÇÑ ¸ðµ¨ÀÇ ¿¹Ãø·ÂÀÌ ³ôÁö ¾ÊÀ» ¶§µµ ÀÖ´Ù. À̸¦ º¸¿ÏÇÏ°íÀÚ, º» ¿¬±¸¿¡¼­´Â Àΰø½Å°æ¸ÁÀ¸·Î Ãß°¡ºÐ¼®À» ÁøÇàÇß´Ù. ±× °á°ú À¯¹æ¾ÏÀÇ »ýüÁöÇ¥·Î¼­ Age, BMI, Glucose, Insulin, ResistinÀÌ °ü·Ã ÀÖ´Ù´Â °á·ÐÀ» µµÃâÇß´Ù ¶ÇÇÑ, k-fold Cross Validation À» ÀÌ¿ëÇØ ¸ðµ¨ÀÇ ¼º´ÉÀ» ºñ±³ÇßÀ» ¶§ ·ÎÁö½ºÆ½ ȸ±ÍºÐ¼®°ú Àΰø½Å°æ¸Á ¸ðµ¨ÀÇ ¼º´É¿¡ Å« Â÷ÀÌ´Â ¾ø¾úÀ¸¸ç °á°úÇؼ®ÀÌ ¿ëÀÌÇÑ ·ÎÁö½ºÆ½ ȸ±ÍºÐ¼®ÀÌ ´õ ÀûÇÕÇÑ ºÐ¼®±â¹ýÀÏ ¼ö ÀÖ´Ù´Â °ÍÀ» È®ÀÎÇß´Ù.
¿µ¹®³»¿ë
(English Abstract)
Artificial neural networks are used as a useful machine-learning method, but they also have limitations compared to conventional statistical analysis. This is because it is difficult to interpret the results. In particular, this disadvantage is apparent in healthcare information studies where identifying relationships among variables is important. The purpose of this study is to compare conventional statistical analysis and machine-learning method to investigate significant factors and determine which analyzing method is more appropriate. Previously the identified risk factors for breast cancer include fat-centered nutrition, obesity, drinking, radiation, and age. Moreover, logistic regression was used in related studies. Applying to logistic regression model, it is easy to interpret the results, but sometimes its predictive power is not enough high. In order to overcome this problem, Artificial neural network was additionally used. As a result, we concluded that Age, BMI, Glucose, Insulin, and Resistin are variables related to breast cancer. In addition, we compared the performance of the model using k-fold Cross Validation and there was no significant difference. so we found that logistic regression with ease of interpretation may be a more suitable analyzing method.
Å°¿öµå(Keyword) logistic regression   artificial neural network   breast cancer   biomarker   k-fold   ·ÎÁö½ºÆ½ ȸ±ÍºÐ¼®   Àΰø½Å°æ¸Á À¯¹æ¾Ï »ýüÁöÇ¥  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå