The World men's Basketball Championship organized by the FIBA is one of the highest level basketball match in the world. Making research on the attack and defense ability of the FIBA World Men's Championship top 16 teams will be very helpful to know the present situation and future trend of the world basketball. The research also can provide academic guidance for the team development. For this purpose, the paper intend to choose 11 technical indicators: two-point shot percentage, three-point shot percentage, free-throw percentage, offensive rebounds, defensive rebounds, assist, fouls, steals, mistake, blocks, and score. By using the factor analysis and Q-type analysis, the authors make a compositive analysis on the offensive and defensive abilities of 2010 FIBA World men's Championship top 16 teams. Meanwhile, the authors try to find out a reasonable quantitative evaluation by study the problems in their performance and offer a basic scientific opinion to improve the competitiveness in each country.
1. RESEARCH OBJECT AND RESEARCH METHODS
1.1 Research Object 2010 World Men's Basketball Championship top 16 teams
1.2 Research Methods
1.2.1 Literature Review
By literature review, the authors find out which technical indicator is useful in the research and a suitable standard in the data filtration. The competition statistics of the top 16 is also collected by this way.
1.2.2 Mathematical Statistics
Using the factor analysis, Q-type analysis, classification analysis, the authors make a study on the 11 technical indicators of the 16 teams with the software SPSS 11.0.
2. DATA PROCESSING
We do not appoint initial cluster centers, so the centers are decided by the system. In the clustering we use the default, the maximum number of iterations is 10, when the convergence parameters is 0.02. It means that the iteration will be stopped when it gets 10 times or when the iteration centers are less than 2%, comparing with the mix cluster centers. In this clustering, the iteration has passed 4 times. After 4 times the iteration center has been almost the same, by the judgement of the 0.02, and the clustering can be stopped. So the class 1-3 can be divided as three levels, A, B and C.
3. RESULTS AND ANALYSIS
3.1 The Q-type Clustering in the Study of TOP 16
The Q-type clustering is a useful way for the sample clustering, making the similar samples get together and separating the difference. The World Men's Basketball Competition is one of the highest level basketball match in the world. In the game, each team treated the champion as the highest honour. In this paper, the top 16 teams are clustered by the definition of Q-type. The raw data is in the Table 1.
The Table 2 show the results of Q-type clustering. There are two teams in the first class. They are the 12.5% of the 16 teamsThey get the most out-standing competition ability in the top 16, own the 1st and the 4th rank in this game. The two teams can get an A level. The 43.75% of the 16 teams are clustered in the second class. The ranks of the 7 teams are the 2nd , the 3rd, the 5th, the 6th, the 9th, the 11th and the 12th. They show a normal competition ability in this game, they can get a B level in the top 16. Other 7 teams are clustered in the third class. The get the 7th, the 8th, the 10th, the 13th, the 14th, the 15th and the 16th rank in this game. Their competition ability is weaker, the can get a C level.
The clustering results show the same phenomenon as the game results. A team which can get the higher shooting percentage, more offensive rebounds, better defensive rebounds, more assists, more steals, a better position it will get. It shows that the occasionality is actually in the basketball games, but a better team always get better rank. A stronger offensive, excellent defensive team always gets a higher winning percentage. Clustering each team into 3 classes by their ranks. …