The knee joint is the largest and most complex flexion and extension joint of the human body. It supports most of the weight of the human body during the whole body during standing or exercise. Because the knee joint has the characteristics of complex structure and large load, it is also vulnerable to damage. An effective diagnosis in the early stage of injury or lesion of the knee joint is of great help to the later treatment. At present, the commonly used knee joint examination methods have the problems of large trauma and high cost. Therefore, this paper uses machine learning technology to study the classification algorithm of knee joint vibration signal. The research results of this paper were verified by selecting the subjects to form a healthy group and a disease injury group. The experimental results show that the proposed signal denoising algorithm is superior to the traditional denoising algorithm. After analyzing several classification algorithms, the multi-classifier fusion algorithm has excellent performance in signal classification. The experimental results show that the research results can be applied to the classification of knee joint vibration signals, and then applied to the clinical diagnosis of knee joint diseases.

膝关节是人体最大，最复杂的屈伸关节。在站立或运动期间，它支撑了人体的大部分体重。由于膝关节具有结构复杂，负荷大的特点，因此也容易受到损坏。在膝关节损伤或病变的早期进行有效诊断对后期治疗有很大帮助。目前，常用的膝关节检查方法存在创伤大，费用高的问题。因此，本文采用机器学习技术研究了膝关节振动信号的分类算法。通过选择受试者分为健康组和疾病损伤组，验证了本文的研究结果。实验结果表明，所提出的信号去噪算法优于传统的去噪算法。在分析了几种分类算法之后，多分类器融合算法在信号分类中具有优异的性能。实验结果表明，该研究结果可应用于膝关节振动信号的分类，进而可用于膝关节疾病的临床诊断。