https://waffle.io/corlin?tab=billing
- 数据结构 (lambda x: (x - x.min()) / (x.max()-x.min())) (lambda x: (x - x.mean()) / x.std())
- 缺失值计数，
- 缺失值插补，
- https://docs.scipy.org/doc/scipy/reference/interpolate.html#module-scipy.interpolate
- http://pandas.pydata.org/pandas-docs/stable/generated/pandas.Series.interpolate.html
- 'krogh'，'piecewise_polynomial'，'spline'，'pchip'和'akima'
- https://machinelearningmastery.com/resample-interpolate-time-series-data-python/
kpi id | value | value_normalized | label | datastime
046ec29ddf80d62e 07927a9a18fa19ae 54e8a140f6237526 b3b2e6d1a791d63a 8a20c229e9860d0c 769894baefea4e9e 76f4550c43334374
Ahmad, S., Lavin, A., Purdy, S., & Agha, Z. (2017). Unsupervised real-time anomaly detection for streaming data. Neurocomputing, Available online 2 June 2017, ISSN 0925-2312, https://doi.org/10.1016/j.neucom.2017.04.070
一种基于滑动窗口的流数据聚类算法
Supervised Sequence Labelling with Recurrent Neural Networks