Optimisation of Random Forest Algorithm Based on Data Mining Analysis and Sparrow Search Analysis of the Impact of 'BM Innovative Thinking System' on Innovation and Entrepreneurship Education in Colleges and Universities

Artificial intelligence-generated content based on big language models has been widely used in marketing, customer service, education, entertainment, news and other fields. Business English programmes need to be integrated with innovation and entrepreneurship education to help students hone their international communication skills and innovative thinking to better meet the challenges of the modern business world. The BM Innovative Thinking System developed in this paper is a web-based software based on artificial intelligence methodology designed to help undergraduate students in Chinese higher education institutions train in innovative and entrepreneurial thinking. The system employs advanced algorithms and technologies that can automatically adjust the learning content and difficulty based on user input and feedback to achieve optimal learning results. In this study, we investigated the role of practical projects in business English courses in improving innovation and entrepreneurship-related competencies and collected self-assessment data from a total of 1205 students. One-way ANOVA showed that there was a significant difference between the pre and post data, i.e., students believed that their innovative and entrepreneurial competencies were significantly improved through the course. The paper also develops a random forest algorithm optimised on the basis of an improved sparrow search algorithm that can predict students' experiences after using the innovative thinking system. In this paper, the model was trained and tested using relevant data, and the confusion matrices in the training and test sets show that the accuracy of the four classifications in the training set is 100%, and the accuracy of the four classifications in the test set is 96.26%, which can well predict the students' experience after using the system. This result has some reference value for the development of business English course design and practice programme.