Abstract
Handwriting recognition finds itself in many domains. Developing improved methods for handwriting recognition is an open research challenge. I will focus on the sub problem of digit recognition. By using the MNIST dataset provided by Kaggle, I will use the Random Forest approach to train and test a digit recognition system. Random Forest is a learning technique that uses an ensemble of decision trees. An ensemble of classifiers has better classification performance than individual classifiers and is more resilient to noise. Thereafter, I shall compare the results with the Random Forest approach with those obtained using Support Vector Machines. Through this research, I hope to build a good understanding of machine learning algorithms and their practical applications.
Faculty Sponsors
Dr. Sumitra Mukherjee
Project Type
Event
Location
Alvin Sherman Library
Start Date
4-8-2016 1:00 PM
End Date
4-8-2016 5:30 PM
Digit Recognition Through Machine Learning
Alvin Sherman Library
Handwriting recognition finds itself in many domains. Developing improved methods for handwriting recognition is an open research challenge. I will focus on the sub problem of digit recognition. By using the MNIST dataset provided by Kaggle, I will use the Random Forest approach to train and test a digit recognition system. Random Forest is a learning technique that uses an ensemble of decision trees. An ensemble of classifiers has better classification performance than individual classifiers and is more resilient to noise. Thereafter, I shall compare the results with the Random Forest approach with those obtained using Support Vector Machines. Through this research, I hope to build a good understanding of machine learning algorithms and their practical applications.
