README.md

Distances between time series of sparse vectors

Various distance functions between time series of sparse vectors: * either directly in the input space (distances_linear.py): dist(bcsc1, bcsc2, **kwargs) * or in the RKHS induced by a kernel between vectors (distances_rkhs.py): dist(K, T1, T2, **kwargs)

Includes the Difference Between Auto-Correlation Operators (DACO) proposed in the paper:

A time series kernel for action recognition
Adrien Gaidon, Zaid Harchaoui, Cordelia Schmid,
BMVC, 2011,
https://hal.inria.fr/inria-00613089v2

Requires the ekovof package.

Author

Adrien Gaidon

License

MIT, except for the global alignment kernel (cf. README-logGAK.txt).