Skip to content

pramodini18/Digit-recognition-using-SVM

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 

Repository files navigation

Digit-recognition-using-SVM

A classic problem in the field of pattern recognition is that of handwritten digit recognition. Suppose that you have images of handwritten digits ranging from 0-9 written by various people in boxes of a specific size - similar to the application forms in banks and universities.

The goal is to develop a model that can correctly identify the digit (between 0-9) written in an image.

Objective: To develop a model using Support Vector Machine which should correctly classify the handwritten digits from 0-9 based on the pixel values given as features. Thus, this is a 10-class classification problem.

Data Description: For this problem, we use the MNIST data which is a large database of handwritten digits. The 'pixel values' of each digit (image) comprise the features, and the actual number between 0-9 is the label.

Since each image is of 28 x 28 pixels, and each pixel forms a feature, there are 784 features. MNIST digit recognition is a well-studied problem in the ML community, and people have trained numerous models (Neural Networks, SVMs, boosted trees etc.) achieving error rates as low as 0.23% (i.e. accuracy = 99.77%, with a convolutional neural network). Before the popularity of neural networks, though, models such as SVMs and boosted trees were the state-of-the-art in such problems.

In this project, trying to experiment with various hyperparameters in SVMs. With a sub-sample of 10-20% of the training data, we should expect to get an accuracy of more than 90%.