International Open Access Journal Platform

聆听生命的旋律：机器学习在音乐行为对生命意义感预测中的应用

Listening to the Melody of Life: The Application of Machine Learning in Predicting Life Meaning Through Musical Behavior

冯媛媛

Psychology of China / 2026,8(6): 908-916 / 2026-06-16 10 6

View Download PDF

• Information:

  宁夏大学教师教育学院，银川

• Keywords:
  Music behavior; Sense of life meaning; Machine learning; Shapley value
  音乐行为; 生命意义感; 机器学习; Shapley 值
• Abstract: The sense of life meaning is a critical factor in mental health, and individual musical behavior may be closely associated with it. This study employs machine learning methods to investigate the role of musical behavior in predicting life meaning among college students. A total of 5,618 valid questionnaires from college students were collected. Musical behavior (including music usage patterns, musical preferences and tastes, as well as music-related functions and emotional responses) and life meaning were measured through the questionnaires. Multiple machine learning models were developed, and Shapley values were used to identify key predictive factors. The results demonstrated that the Gradient Boosting Tree model achieved the best predictive performance, with an average AUC of 94.45%, indicating a complex relationship between musical behavior and life meaning. Shapley value analysis revealed that musical behavior characteristics such as preference for listening to lyrics with depth and extended daily engagement with music positively correlated with life meaning, whereas listening to music in the early morning or using music as a sleep aid had negative predictive effects. This study validates the significant predictive role of musical behavior in life meaning through machine learning and identifies key predictive factors, providing a theoretical foundation for mental health interventions among college students. It also highlights the broad application prospects of machine learning in psychological research. 生命意义感是心理健康的关键因素，个体的音乐行为可能与之存在密切联系。本研究采用机器学习方法，探讨音乐行为在预测大学生生命意义感方面的作用。共收集5618名大学生有效问卷。通过问卷测量音乐行为（包括音乐使用行为、音乐偏好与品位、音乐功能与情绪）和生命意义感。采用多种机器学习建立预测模型，并利用Shapley值分析关键预测因素。结果显示，梯度提升树的预测效果最佳，AUC均值高达94.45%，表明音乐行为与生命意义感之间确实存在复杂关联。Shapley值分析发现，喜欢聆听有深度歌词、长时间参与音乐活动等音乐行为特征能正向预测生命意义感，而喜欢在凌晨聆听音乐、将音乐作为助眠手段对生命意义感有负向预测作用。本研究通过机器学习方法验证了音乐行为对生命意义感的重要预测作用，并初步找到了预测的关键因素，为大学生心理健康干预提供了理论依据。同时也展示了机器学习在心理学研究中的广泛应用前景。
• DOI: 10.35534/pc.0806134 (registering DOI)
• Cite: 冯媛媛. (2026). 聆听生命的旋律: 机器学习在音乐行为对生命意义感预测中的应用. 中国心理学前沿, 8 (6), 908-916.