한국 프로농구선수의 경기 실적과 연봉간의 관계분석

김세중(Kim Sae-Joong) 한국체육과학회 2009 한국체육과학회지 Vol.18 No.3

The present study strives to provide a database for an even more reasonable yearly salary contract model to the KBL, which is the arbitration committee for determining the yearly salary, and the professional baseball teams by verifying the rationality between the present contracts concerning the yearly salaries and the performance of the players. 95 players with closed contracts for the 2008~2009 season of the SK telecom pro basketball league were chosen as the subject for the present study. The game records of the players and the data on contract negotiations were collected from the korean basketball league website, www.kbl.or.kr, and with the help of the service bureau of the Incheon J professional basketball team. The data was processed with SPSS VER. 16.0. Frequency analysis, correlation analysis, and a multiplex regression analysis were carried out. According to the results of the analysis, time of participation in the game, scores, and rebounds (in order) were the most influential factors of the pay of the basketball players. Only participation in games allow the accumulation of play performance. Therefore, the time of participation in games is an important factor in the determining the yearly salary. It was possible to explain the current yearly salary determining system for the professional basketball players with time of participation in games, two-points, rebounds, and three-points. It is expected that the data of the present study will be used as a grounds for a systematic and reasonable negotiation of the yearly salary, which is held annually, between the players and professional basketball teams.

Hidden Markov Model을 이용한 스포츠 경기분석

김세중(Kim, Sae-Joong),정영한(Jung, Young-Han),허종관(Heo, Chong-Kwan) 한국체육과학회 2017 한국체육과학회지 Vol.26 No.3

To analyze and forecast location and direction of stroke is important by means of one of factor to be able to improve performance. For this reason, we need to analyse these factor through systematic and multilateral method. Consequently, purpose of this study were to extract a variable rally pattern and evaluate efficient of each rally pattern with check on location and direction of stroke in tennis match situation and analyzing by hidden markov chain model. For this purpose, we checked on rally pattern of Djokovic and Federer from champion match held australian open 2016. And These data analyzed by hidden Markov chain model. results were follow as: 1. We are able to estimate transition probability on stroke/serve direction of each athlete. 2. We find out transition probability on rally direction under the promise performing first/second serve. 3. When number of continuing rally increase, virtually, miss shot increased as increasing number of rally, but another direction of stroke decreased after 2 stroke, consistently.

운동이 노인의 근기능에 미치는 영향에 대한 메타분석

김세중(Kim, Sae-Joong),도재현(Do, Jae-Hyun) 한국체육과학회 2016 한국체육과학회지 Vol.25 No.4

The purpose of this study were to confirm effect size of whole exercise on elderly muscle function using meta analysis. For this purpose, we were extracted 32 studies related elderly muscle function from RISS, and selected 8 studies measured muscle function using isokinetic method. Therefore, 114 results of 8 studies were used in the study. Effect size of whole and exercise type were analyzed by random effect model. Estimating effect size and heterogeneity, all data were analyzed by R. 3.3.0. All significant level of statistics were set up 95%. The results is follow as : 1. Exercise presented to have seriously effect on elderly muscle function. And The study appeared to be Heterogeneity among them. 2. Exercise type as moderator of elderly muscle function appeared to be significant difference. It was shown highest effect size that complex exercise but, core, isotonic, resistance and exercise appeared to be not significant effect. In conclusion, Exercise can be effect on elderly muscle function but, difference depending on exercise type can be difference of effect size.

KBL 경기운영규정 변화에 따른 경기자료 내용분석

김세중(Kim, Sae-Joong),이우경(Lee, Woo-Kyung) 한국체육과학회 2016 한국체육과학회지 Vol.25 No.2

To analyze the operational system before/after changes. The decision tree analysis was conducted by SPSS clementine 11.1 and selected algorithms of CART(Classification and Regression Tree). Prior to decision tree, the variable selection methodology was chosen as Feature Selection of Clementine the rate of maximum missing value over 70%, importance value over .95 The launch interactive session mode of CART algorithm was used in the study. Maximal branch depth was set as 8, minimal change for impurity as .0001 for surrogate split and stopping rules, and Gini impurity measure was used. Default values provided by Clementine were used for the rest of setting. Below is the results. Before changes in the operational system in KBL, defensive-rebounds and fouls with free-throws with assists and free-throw attempts were the most important factor associated with winning. After the system has changed, just both defensive-rebounds and assists are the most important factors.

한국 프로농구 경기 2000~2009시즌 경기력 요인과 관련된 주요 승ㆍ패 패턴과 요인 분석

김세중(Kim Sae-Joong),이기광(Lee Ki-Kwang),허종관(Heo Chong-Kwan) 한국체육과학회 2011 한국체육과학회지 Vol.20 No.3

The propose of this study was to investigate factors and patterns related victory and defeat in basketball play using data mining. For this study, records on 810 games of 2005-2006 ∼ 2008-2009 Korean basketball league were collected and used for developing and determining model. Method of analysis on data selected CART algorism. The Decision tree model created by CART algorism formed by 10 factors and 15 patterns, finally. In results, Main factors and patterns related victory were ; 1. The total 5310 games during 10 seasons in KBL are analyzed by CART. The key factor of victory consists of defensive-rebounds, assists, fouls followed by less free-throw attempts, and more free-throw attempts. As for defeats, it is turned out that lower points of defensive-rebounds, assists, 2 and 3-point shots, and free-throws are defeat factors.

한국 프로농구리그 구단별 경기력의 군집화를 통한 상대적 전력평가

김세중(Kim, Sae-Joong),도재현(Do, Jae-Hyun) 한국체육과학회 2015 한국체육과학회지 Vol.24 No.3

This study aims to provide a quantified scientific analysis index by subdividing groups according to game results, relevant factors affecting player performances and clustering of groups with similar level of performances. In order to conduct this study, 2970 games in 11 regular seasons of a total of 10 teams from 2003 to 2013 were analyzed using the Decision Tree of Data Mining based on the relevant factors affecting game. A Two-Step Cluster Analysis was performed to cluster the target teams. The results are as follows: First, the most significant factors determining game results are defensive rebound and assist. Prediction on total game results showed 69.4% while winning prediction rate showed 74.1% with a 64.8% of losing prediction rate. Second, as a result of the cluster analysis based on the factors extracted from decision tree, 3 clusters were formed. Cluster 1 was composed of Team 1(Jeonju KCC), Team 5(Seoul Samsung), Team 8 (Changwon LG) and Team 10(Ulsan Mobis). Cluster 4 was composed of Team 4(Busan KT) and Team 7(Wonju Dongbu).

WKBL 선수 경기기록을 활용한 경기력 요인과 관련된 주요 경기 패턴 분석

김세중(Kim, Sae-Joong),도재현(Do, Jae-Hyun) 한국체육과학회 2015 한국체육과학회지 Vol.24 No.4

This study utilizes the competition record of six teams of players that belong to the WKBL, to examine the pattern of the game result calculation process. To extract the factors in previous studies associated with game result represented by intuitive indicators specific conditions and game factors that are not derived, I want to know the trend of the game by the extraction process. The subject of this research are 3402 events from players of six of WKBL teams played during KB kookmin bank 2013-2014, 2014-2015, and with this date we applied it to CART algorithm of decision tree analysis, one of date mining method. Gathered player’s date was coded with MS excel then were refined that affect errors associated with missing values. Based on refined data, we analyzed it with SPSS modeler 14.2. As a result of this research, we produced two main causes(PZP : Paint zone point, FZP : Fast Break Point) and seven pattern models. Main node results were node 13(none%) 2.89, gain 5.80, response 100, and index 200.71. Patterns were paint zone gain under 37.5 → paint zone gain over 36.5 → Number of swift attack 2.5 over each game → paint zone game over 33.5.

한국프로농구 국내선수와 귀화혼혈선수의 경기력 특성 비교분석

김세중(Kim, Sae-Joong),도재현(Do, Jae-Hyun),허종관(Heo, Chong-Kwan) 한국체육과학회 2014 한국체육과학회지 Vol.23 No.3

The purpose of this study is comparison of Korean players and half-Korean players who belong to korean basketball league team`s through an actual data and effective analysis of player performance factors and The ability of these individuals from an intuitive way to objectively evaluate. From this method we can discuss how korean players and half-korean`s performance can make win or lose their team`s and how much they are interactive and effective with teams in depth,. Method of analysis on data selected CART algorism. The Decision tree created by CART algorism formed by 10 factors and 15 patterns, finally. In results, Main factors and patterns related victory were ; 1. Korean player`s game has 8 factors(AST, FTA, P2, WFT, WOFT, FT, FTA, BS) and 15 patterns. On the other hand half korean player`s game has 7 factors(TL, WFT, OFF, FT, BS, P2, P3) and 9 patterns. 2. Korean players match patterns and factor analysis of the game. The node pattern is AST 2.5 over→ P2 0.5 over→ WFT 0.5 over→ FTA 2.5 over. 3. Game pattern and factor analysis of Half korean`s games. The node pattern is STL 1.5 over→ WFT 1.5 under→ FT 5.5 over→ FTA 2.5over.

중ㆍ고교 농구 선수들의 의식구조와 생활패턴에 관한 질적 연구

김세중(Kim Sae-Joong),김도연(Kim Do-Yeon) 한국체육과학회 2009 한국체육과학회지 Vol.18 No.1

The purpose of this study was to help to a decision on a policy in school authorities and offer coaches and managers a basic material about efficient training method and systemic athletes' management method through investigating and analyzing their lifestyle, awareness and occurring the present state in social environment by their view point about middle and high school basketball players. For this study, the researcher directly had an interview with middle and high school basketball player. And collected data could concretely categorize sports, society and so on by analyzing their internal and environmental factors. In conclusion, lifestyle and awareness of general present stats occurring middle and high school students were categorized by training, injury, coach, schoolwork, course after graduation, and relationship with schoolmates and teachers. They were aware of anxiety about injury and the pressure according to stress. The feeling of alienation that they had perceived appeared to reduce their schoolwork and relationship with schoolmates and teachers.

데이터 마이닝 기법을 이용한 한국 프로농구 경기 분석

김세중(Kim Sae-Joong),허종관(Heo Chong-Kwan),도재현(Do Jae-Hyun) 한국체육과학회 2010 한국체육과학회지 Vol.19 No.2

The propose of this study was to investigate factors and patterns related victory and defeat in basketball play using data mining. For this study, records on 810 games of 2005-2006~2008-2009 Korean basketball league were collected and used for developing and determining model. Method of analysis on data selected CART algorism, The Decision tree model created by CART algorism formed by 10 factors and 15 patterns, finally. In results, Main factors and patterns related victory were ; 1. Defence rebound, 2point, 3point, 2point attempt and 3point attempt factor and pattern related these. 2. Defence rebound and 2point factor and pattern related these. 3. Defence rebound, assist, 2point and block shot and pattern related these. And Main factors and patterns related defeat were ; 1. Defence rebound, assist, steal and turn over factor and pattern related these. 2. Depence rebound, assist, steal, turn over and free throw factor and pattern related these.