• 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 : 3 / 3

ÇѱÛÁ¦¸ñ(Korean Title) MMORPG »ç¿ëÀÚ À¯Çü ºÐ·ù¸¦ ÅëÇÑ ÀÌÅ» ¿¹Ãø ¸ðµ¨ »ý¼º ¹× Æò°¡
¿µ¹®Á¦¸ñ(English Title) Constructing and Evaluating a Churn Prediction Model using Classification of User Types in MMORPG
ÀúÀÚ(Author) ¿À¼¼ÁØ   ÀÌÀºÁ¶   ¿ìÁö¿µ   ±èÈÖ°­   Sejoon Oh   Eunjo Lee   Jiyoung Woo   Huy Kang Kim  
¿ø¹®¼ö·Ïó(Citation) VOL 24 NO. 05 PP. 0220 ~ 0226 (2018. 05)
Çѱ۳»¿ë
(Korean Abstract)
´ë±Ô¸ð ´ÙÁß »ç¿ëÀÚ ¿Â¶óÀÎ ·ÑÇ÷¹À× °ÔÀÓ(MMORPG)Àº ¼¼°èÀûÀ¸·Î ¸¹Àº »ç¿ëÀÚµéÀÌ Áñ±â´Â °ÔÀÓÀÇ À帣·Î »ç¿ëÀÚµéÀÌ Áñ±æ ¼ö ÀÖ´Â ´Ù¾çÇÑ ÄÁÅÙÃ÷¸¦ Á¦°øÇØÁØ´Ù. ÇÏÁö¸¸ ´Ù¾çÇÑ ÄÁÅÙÃ÷¸¦ Á¦°øÇÔ¿¡µµ ºÒ±¸ÇÏ°í ÀϺΠ»ç¿ëÀÚµéÀº °ÔÀÓ¿¡¼­ ÀÌÅ»ÇÑ´Ù. Áö¼ÓÀûÀ¸·Î °ÔÀÓÀ» Áñ±â´Â »ç¿ëÀÚ´Â °ÔÀÓȸ»çÀÇ ¼öÀÍ°ú ¹ÐÁ¢ÇÑ °ü°è°¡ Àֱ⠶§¹®¿¡, »ç¿ëÀÚ ÀÌÅ» ¿¹ÃøÀº ¸Å¿ì Áß¿äÇÑ ¹®Á¦ÀÌ´Ù. º» ¿¬±¸¿¡¼­´Â °ÔÀÓ »ç¿ëÀÚ À¯ÇüÀ» ±â¹ÝÀ¸·Î ÀÌÅ»À» ¿¹ÃøÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ½ÇÁ¦ MMORPG µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÏ¿© ¡®ÀÌÅ»ÀÚ¡¯¸¦ Á¤ÀÇÇÏ°í, ¼­·Î ´Ù¸¥ Ư¡À» º¸ÀÌ´Â »ç¿ëÀÚµéÀ» ±ºÁýÈ­ ¾Ë°í¸®ÁòÀ» ÅëÇØ 5°¡Áö À¯ÇüÀ¸·Î ³ª´©¾î ºÐ·ùÇÏ¿´´Ù. ½ÇÇè °á°ú·Î ¾à 98.3%ÀÇ ÀÌÅ»ÀÚ°¡ ¶óÀÌÆ® »ç¿ëÀÚÇü¿¡ ¼ÓÇÏ´Â °ÍÀ» È®ÀÎÇÏ¿´´Ù. ºÐ·ùÇÑ À¯ÇüÀ» ÀÌÅ» ¿¹Ãø ¸ðµ¨ »ý¼ºÀÇ ÇÇó·Î »ç¿ëÇÏ°í, ±â°èÇнÀ ¾Ë°í¸®ÁòÀ» ÀÌ¿ëÇÏ¿© ÀÌÅ» ¿¹Ãø ¸ðµ¨À» ±¸ÃàÇÏ¿´´Ù. ÀÌÅ» ¿¹Ãø ¸ðµ¨Àº ÃÖ´ë 85.7%ÀÇ accuracy¿Í 72.3%ÀÇ F-measureÀÇ ¼º´ÉÀ» º¸¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Massively Multiplayer Online Role Playing Games (MMORPG) have been the most enjoyed genre of games by users all over the world, because of the many enjoyable contents they provide. However, there still are some users leaving MMORPG, in spite of the various contents. Because the users, who are constantly engaged in a game, are directly connected to the profit of the game company, it is necessary to study churn prediction. In this study, we propose a churn prediction method based on the user types. Using actual MMORPG data, we define ¡°Churn Users¡± and categorize users into 5 types by means of the clustering algorithm. The experiment shows that about 98.3% of the churn users are the ¡°Light user type¡±. We then use user types as a feature and construct a churn prediction model with various machine learning algorithms. The churn prediction model shows a maximum accuracy of 85.7% and an F-measure of 72.3%.
Å°¿öµå(Keyword) »ç¿ëÀÚ ÀÌÅ» ¿¹Ãø   »ç¿ëÀÚ À¯Çü ±â¹Ý ºÐ·ù   ±â°èÇнÀ ¾Ë°í¸®Áò   ÀÌÅ» ¿¹Ãø ¸ðµ¨ Æò°¡   churn prediction   user type classification   machine learning algorithm   evaluating churn prediction model  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå