Abstract: Studies on competitive sports efficiency can help to evaluate the investment of a country or region in ' competitive sports and figure the related influence factors on the development of competitive sports so as to provide ideas for the further development of sports in this region. Depending on the DEA method and the K-centers functional clustering method, the paper analyzes the situations of economy, population and medals at the National Games in 31 provinces, regions and cities from 1994 to 2013, finding that there exist obvious differences in competitive sports development efficiency between different places. According to the situations of different provinces and regions, the 31 provinces are divided into five different development modes of economic-oriented high efficiency group, population-oriented high efficiency group, low efficiency group with economic advantages, low efficiency group with population advantages, and low efficiency group with both economic and population advantages. The paper further discusses the reasons for the increases and decreases of competitive sports efficiency so as to come up with suggestions for the improvement of competitive sports efficiency in different provinces and regions.

Key words: DEA, k-centers functional clustering method, panel data, competitive sports efficiency