Modeling and Analysis of Psychological Change and Adaptability of College Students Based on Machine Learning as an Infrastructure to a Smart City

In order to improve the mental health education of college students, reduce the intermediate links, and improve work efficiency, this paper analyzes the psychological change characteristics and adaptability of college students based on machine learning. We should introduce speech recognition technology, carry out natural language processing, vectorize the speech text, construct the semantic perception model of college students’ mental health states, and track the sensitive words similar to the evaluation standard of college students’ mental health states; we should introduce multifeature fusion technology through measuring the description ability of different features, learn the complementary state of different sensitive words of different features, and perceive the psychological change characteristics of college students and serialize them; we should, based on the decision tree algorithm in machine learning, construct the analysis model of psychological adaptability of college students, analyze the sensitive words and the frequency and level of sensitive words in the process of college students’ mental health conversations, determine their adaptability, and complete the modeling and analysis of psychological change characteristics and adaptability of college students based on machine learning. The experimental results show that the method has no change to the original dependency relation, the time cost of feature acquisition is still small, and the sensing effect of sensitive words is close to the ideal value.