Profiling Smurfs And Boosters on Dota 2 Using K-Means

Ding, Ying Jih (2021) Profiling Smurfs And Boosters on Dota 2 Using K-Means. Final Year Project, UTAR.

Preview

PDF Download (8Mb) | Preview

Abstract

Dota 2 is one of the most popular Multiplayer Online Battle Arena (MOBA) game and it also holds the grandest e-Sports tournament in the world —— The International. However, the game is experiencing a continuous decline in its player count. This is because the existence of smurfs/boosters in Dota 2 is ruining the game experience for all other Dota 2 players. Hence, this project aims to identify the smurfs/boosters and analyse their skills. The data were collected from OpenDota API and a data set was created after cleaning and pre-processing. To identify the smurfs and boosters in the data set, K-Means was used to divide the players into groups. To identify the high-skill players group, feature values of the data were examined. Interquartile Range (IQR) method was then used on the high skill players group to identify and profile smurfs/boosters. The resulted profile was reviewed by two game experts and one active player. A 95% accuracy score was achieved using majority voting. It is hoped that this work can be furthered for identifying the different skill levels of the smurfs/boosters after identifying them.

Actions (login required)