At present, colleges and universities attach great importance to the assessment of teachers, strengthen the assessment of college teachers, and create a team of teachers with high quality, professionalism, and vitality, which can not only fully develop and allocate human resources in colleges and universities, but also promote colleges and universities. The reform of the personnel system will continuously optimize the proportion and structural allocation of teachers’ resources in colleges and universities, improve the quality and efficiency of running schools, enhance the competitiveness and attractiveness of schools, and better achieve the school’s school-running goals. In order to obtain higher signal reconstruction accuracy and satisfy random matrix with a relatively small compression ratio, this paper combines QR decomposition, which is used to increase the column independence of random matrix, and gradient descent method, which can reduce the cross-correlation between random matrix and sparse matrix. Based on the actual situation of a certain university, this paper studies and designs a set of perfect and reasonable teacher performance appraisal system that meets the requirements of the college, analyzes and researches the current situation of teacher performance appraisal in the college, and finds out the problem. Combining the advanced concept of being good at performance appraisal with the actual situation of the college, it analyzes the problems in the teacher performance appraisal system, summarizes the reasons for the formation of the problem, further redesigns the teacher performance appraisal index system, and gives the relevant implementation of the system. The basic characteristics presented in the organization of colleges and universities will not change. With the continuous adjustment of the organization of colleges and universities, the goals of colleges and universities are also being adjusted. Colleges and universities should continuously improve the performance appraisal system of teachers to make the goals of teachers consistent with the development goals of colleges and universities, so as to continuously improve the basic quality and realization of college teachers.

Teachers’ teaching psychological behavior and classroom development are the current research hotspots in the field of educational psychology. How to realize the data analysis of teachers’ teaching psychological behavior and classroom development is a problem that researchers urgently need to solve. Based on the theory of data correlation analysis, this paper uses modern Internet technology and big data analysis teacher teaching system to quantitatively and qualitatively analyze the potential of students, and build a corresponding model. Through rule correlation technology, the article studies various internal correlations between teachers’ teaching psychological behavior, extracts valuable information from various daily data of students through big data analysis technology, and the WEB teacher’s teaching psychological behavior analysis system based on B/S structure solves the problem that the traditional model cannot measure. In the simulation process, the system is implemented by MVC three-tier architecture, the database uses MYSQL 5.0, the prediction questionnaire is formulated on the basis of the literature method and interviews, and the scale is compiled and tested after repeated revisions. Project analysis and factor analysis are performed on the data obtained from the table test to construct and screen indicators. The experimental results show that the teacher’s classroom teaching behavior index system adopted by the system is practical and feasible, including three first-level indicators, 10 s-level indicators, and 21 third-level indicators. The system has 87.1% completeness, which effectively improves teachers’ teaching psychology.