DkNN is a version of the well-known supervised classification algorithm k-Nearest Neighbours, with a slight twist that allows for the algorithm to be deployed on devices with scarce hardware resources.
- Updates dilution parameters and class centers.
- Calculates for a data point, distance per class center.
- Monitors overconfidence circles of classes.
- Does DkNN classification on data.
- Drops incomplete batches.
To use CSVLib in your project, follow these steps:
-
Clone this repository to your local machine:
-
Include the 'dknn.h' header file in your C source files.
-
Build your project with 'dknn.c' as part of your source files.
Contributions are welcome! If you encounter a bug or have ideas for improvements, please open an issue or submit a pull request.
DkNN is released under the MIT License. See the LICENSE file for full licensing details.