Anmol Srivastava

Implement NLP transforms, KS alarm + detector

3

**As experimental code, this serves as an outline of the NLP framework rather than final versions.** Closes #152 # Added - transforms **(`Representation` removed)** related to `alibi-detect` example - basic...

### Task Migrate `experimental` into `menelaus` proper. - [ ] assure testing/formatting is finished - [ ] rename to `nlp` for now -- we may merge detector, alarm, and transform...

### Task - [x] correct / fill in docstrings for added components - [x] add top-level docstrings - [x] fill out prose for example notebook

### Task - [ ] unit tests for added components - [ ] verify output and results in example notebook, adding plots where relevant

Add example notebook for NLP drift detector(s)

1

Placeholder issue for development of first example added to RTD, demonstrating use of algorithms developed in #153

Design `nlp` submodule and skeleton of drift detection methods

5

### Task NLP-based drift detection algorithms do not always fit into data-drift or concept-drift definitions, so a separate submodule can be made and a basic skeleton of a language or...

Placeholder issue for implementation of the first NLP-based drift detection algorithm added to `menelaus`

Comprehensive Example for Drift Injection Methods

2

We should probably add a notebook which covers at least the more complex drift injection functions implemented from literature. Can use other examples for reference.

http://proceedings.mlr.press/v80/boracchi18a/boracchi18a.pdf