Implementation of K-Means Clustering algorithm using python
CIFAR-10, contains 60000 32x32 images, where 50000 are used for training
To find out more about the dataset: https://www.cs.toronto.edu/~kriz/cifar.html
Directory of data-set
/cifar-10-batches-py
- Run main.py
- Enter the value of k
- Enter the number of iterations
- Wait for the program to complete
Ouptut of the run is generated in
/output/iterations= xx, k= xx/run_id=x
It contains:
- Centroids and samples from each cluster
- Distortion measure graph
- Excel sheet containing accuracy fit data
To obtain this accuracy measure results are compared to ground truth
