Problem: how to cluster non-stationary multivariate time series.
What are stationary time series
How to consider both global features and local features?
Supplementary Knowledge:
1. 聚类评价指标的好坏:
首先必须明确一点,下述的聚类效果好坏的评价指标只与聚类后的类别标签有关。换句话说,就是判断聚类算法给每个数据的类别标签与真实标签的一致性程度。虽说聚类是无监督学习,但评价指标均需要用到类别标签。
下面分别介绍目前常用的评价指标:
- Purity (纯度)
- Normalized Mutual Information (NMI,归一化互信息)
- Entropy(熵值)
- True Positive (TP)、False Positive (FP)、False Positive (FP)、True Negative (TN)
- Rand index (RI,兰德指数)
- F measure