giotto-ai / giotto-tda

A high-performance topological machine learning toolbox in Python
https://giotto-ai.github.io/gtda-docs
Other
849 stars 174 forks source link

Theory page issues #250

Closed ulupo closed 4 years ago

ulupo commented 4 years ago

Description

This thread is for reports on issues/typos with the theory page, uploaded as a tex file by https://github.com/giotto-ai/giotto-tda/commit/7b4ead4bb04260eb1c6408195832e5e2bf3bb505. All welcome to participate!

marek094 commented 4 years ago

There is a one-letter typo in the definition of Persistence module morphism to commute with simplicial maps: F(t) \circ f{st} = f{st} \circ F(s). (moreover, assuming that the f_{xy} are generic for both modules which may not always be the case :/) Persistence entropy -- missing L_{D} index in the where clause

ammedmar commented 4 years ago

Preparing a PR to fix those typos. Thanks @marek094 !

ammedmar commented 4 years ago

From gtda.time_series.TakensEmbedding:

"Based on a time-delay embedding technique named after F. Takens [1]. Given a discrete time series X and a sequence of evenly sampled times , one extracts a set of -dimensional vectors of the form for X. This set is called the Takens embedding of the time series and can be interpreted as a point cloud."

I think we have a lot of things called Takens embedding. At the most granular level we have the "embedding" technique which sometimes is called time delay. The smooth and discrete contexts. The theorem that states that time delay defines an embedding in the smooth context when done appropriately. And the image of the "embedding".

I am thinking of having a single glossary entry called Takens Embedding clarifying this terminology. At the moment the entry mostly focuses on the Theorem.

ammedmar commented 4 years ago

This system to report issues has been deprecated. Current method is based on a single issue per requested upgrade.

ulupo commented 4 years ago

Thanks @ammedmar !