Prediction of career tendency and personalized suggestions for college students under cluster analysis

Abstract

This paper aims to solve the problem of insufficient personalization and foresight in college students’ career planning, and to provide more accurate career advice through comprehensive analysis of student data. The study first collected students’ questionnaires and academic records, and used the k-means clustering algorithm to cluster students’ personal characteristics to identify key factors affecting career development, such as major choice, academic performance, and academic activities. In order to deeply understand the dynamic changes of students’ career tendencies, the study introduced the LSTM model, which was analyzed based on the students’ four-year long-term time series data to predict their career development trends. Experimental data show that this method has improved prediction accuracy and correlation compared with traditional multiple linear regression and convolutional neural networks, with prediction accuracy and correlation reaching 0.847 and 0.945, respectively. By predicting the career tendencies of specific students, the study also provides personalized career development suggestions, such as improving interpersonal communication skills and logical analysis skills, to help students clarify their employment direction and make up for potential career path shortcomings.