-
There are python implementations of dynamic time warping (dtw) from individual GitHub users, but none have the backing of a large open source project. It feels like the dtw algorithm would be at home …
-
Would you have any suggestion regarding a more efficient clustering method?
I asked Mistral AI and was informed about:
Distance-based methods: These methods measure the similarity between time …
-
Thanks for sharing of your work .
In the paper(View-Invariant Probabilistic Embedding forHuman Pose),the embeddings for action recogni-
tion is evaluated by using nearest neighbor search with the se…
-
First of all let me thank you for this package, and say: great work!
**Is your feature request related to a problem? Please describe.**
I am working with timeseries data. For now I have applied K-…
-
Hi, Paul
Very excellent implementation of dtw.
I am trying to integrate the DTW sub-module into well-known tensorflow as an gpu operator. For being compatible to the tf APIs, it is better to h…
-
Hi there. I'm using Kmeans with softDTW for my thesis on a dataset made of 115 time series that are up to 10000 long.
I'm running my script on a Ryzen 3600X (pretty much last gen mainstream processo…
-
The aim of this issue is to collect some high-level requirements or use-cases where pixelflow could be helpful. Open additional issues based on the replies here. Please leave a comment about your us…
ots22 updated
4 months ago
-
There seems to be no structure to the placement of individual package base classes (or if there is, it is very broad). Most are situated in the main package folder i.e. `sktime.classification.base.py`…
-
It seems that `metrics.lcss` is non-conformant with generic interface expectations towards time series distances when passing 3D `np.ndarrays`.
It would seem that `lcss` fails on some, or all 3D `n…
-
### Describe the bug
Some of the distances have broken logic with the bounding matrix. The reason is due to the cost matrices being initialised to 0s instead of infinity like in DTW. Here is a list…