The Evaluation of the Innovation Teaching Ability of Physical Education in Colleges and Universities under the Regional Difference

With the development of economy in our country, the curriculum reform of basic education also in advancing smoothly, China's college sports teaching mode and teaching idea of experiencing a profound change, in constantly improve teaching ability in Colleges and universities in China, policy measures to improve at the same time, the reform encountered a variety of problems are solved, how to solve the contradiction between theory and practice, how to improve our country colleges and universities sports teaching innovation ability, especially in relation to the students' subjective initiative play selection mode of innovation capability and its evaluation and need to more deeper research. In this paper the data analysis basis, application of machine learning algorithm, each regional college physical education course innovation ability evaluation model based on SOM neural network algorithm was established, after the implementation of the course innovation of our country sports teaching to make objective and fair evaluation, develop a good robustness, high precision of teaching ability evaluation model.