Application Effect Evaluation and Influencing Factors of Music Therapy in Psychological Rehabilitation Based on Big Data Analysis

As mental health problems become increasingly serious in modern society, how to effectively promote psychological rehabilitation has become an urgent issue to be solved. To this end, this study explores the application effect of music therapy in psychological rehabilitation based on big data analysis and evaluates its key influencing factors. This study uses XGBoost as the main prediction model. By collecting and integrating student psychological rehabilitation case data from the education field, it analyzes the effect of music therapy on emotional changes and mental health improvements in student groups. The XGBoost model can effectively process high-dimensional data, capture the complex nonlinear relationship between music therapy and students' psychological state, and provide accurate efficacy prediction. To further explore the impact of different types of music therapy on students' mental health, the study also uses heat map analysis methods to reveal the role of factors such as music type, treatment length, and individual differences in treatment, and to provide data support for personalized treatment plans. The results of the experiment showed that music therapy can significantly improve students' anxiety and depression symptoms. The anxiety scores of students who received classical music therapy decreased by 18%, and the depression scores decreased by 15%. In general, music therapy, as an effective means of psychological rehabilitation, has a positive impact on improving students' mental health, and its effect is mediated by multiple factors.