[IGNITE-9283] [ML] Add Discrete Cosine preprocessor

• Type: Sub-task
• Status: In Progress
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 3.0
• Component/s: ml
• Labels: newbie
• Ignite Flags: Docs Required
• Assignee: Ilya Lantukh
• Reporter: Alexey Zinoviev
• Created: 16/Aug/18 04:01
• Updated: 31/Oct/19 13:03

Description

Add https://en.wikipedia.org/wiki/Discrete_cosine_transform

Please look at the MinMaxScaler or Normalization packages in preprocessing package.

Add classes if required

1) Preprocessor

2) Trainer

3) custom PartitionData if shuffling is a step of algorithm

 

Requirements for successful PR:

1. PartitionedDataset usage
2. Trainer-Model paradigm support
3. Tests for Model and for Trainer (and other stuff)
4. Example of usage with small, but famous dataset like IRIS, Titanic or House Prices
5. Javadocs/codestyle according guidelines

 

 

Issue Links

• links to: GitHub Pull Request #6735