振动数据集
收藏资源简介:
该数据集由都灵理工大学的研究人员创建,用于评估无监督机器学习算法在新颖性检测中的性能。数据集通过在实验室中使用振动器记录特定频率的振动信号生成,并通过改变输入波信号来引入新颖条件。数据集包含统计特征和小波分解系数,旨在解决复杂数据结构中的新颖模式识别问题,特别适用于预测性维护和异常检测领域。
This dataset was developed by researchers at Politecnico di Torino for evaluating the performance of unsupervised machine learning algorithms in novelty detection. It is generated by recording vibration signals at specific frequencies using a vibrator in a laboratory environment, with novel conditions introduced by altering the input wave signals. The dataset contains statistical features and wavelet decomposition coefficients, and is intended to solve novelty pattern recognition problems in complex data structures, making it particularly suitable for applications in predictive maintenance and anomaly detection.

- 1Unsupervised Novelty Detection Methods Benchmarking with Wavelet Decomposition都灵理工大学电子与电信系 · 2024年



