Innovation and entrepreneurship education (IEE) is a powerful driving force to promote the transformation and development of college education and an effective way to broaden the employment channels. However, IEE has been used to train students in classrooms by college managers who do not systematically think of practical training for innovation and entrepreneurship talents. There are still many difficulties in the process of development of IEE that managers cannot ignore. The relevant models that could deeply analyze the connotation of IEE were established through the game theory that explained the game balance among government, universities, and students in the IEE process. A game theory model explained the external game behavior of universities and students and the internal game behavior between students. Another model analyzed the strategies of players pursuing maximum payoffs from the perspective of both sides of the game and the influences of these strategies on the players. The problems in the process of IEE development are concluded including imperfect management system, unscientific curriculum design, nonideal talent training and unprofessional teachers. Based on it, the development path of IEE is explored through the game concept. A scientific and leading system of management and evaluation is needed to put forward by strengthening the top-level design at the national level, and a win-win game situation between colleges and students will be achieved by the optimized education system.

An adaptive learning model for English vocabulary through a machine learning is proposed in this paper. The four main types of user information, including basic student information, quiz information, course video viewing information, and forum interaction information, are processed through feature engineering, and a better model on sparse data is proposed through comparison on different models, and the prediction accuracy of the model is improved through natural language processing techniques, to achieve feedback on user learning efficiency through user data and provide teachers and students with the corresponding teaching and learning suggestions for teachers and students. It is found that the quiz information has more influence than the course video viewing information, and the accuracy is improved by about 3% compared with TF-IDF after introducing word embedding. The use of mobile for English learners to learn to read in a fragmented learning context enables targeted training in weak areas of English reading, thus improving different aspects of learners' reading skills.