This kernel is for recognizing handrwritten digits prsesent in the MNIST dataset uses KNN implementation from the scipy toolkit to classify the digits given in form of a 28*28 pixel image. The kernel has validation accuracy of 97.166% on the trainning set of 42000 images & accuracy of 96.80% on the test data of 28000 images
bhaskarsen98/Handwritten-digit-recognizer
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|