This is a Digit Recognization multiclass classfication problem. The data set is a subset on MNSIT dataset, only containing images of 1, 4,3 and 9
The data is given in form of feature vectors
What is it about?
In this task you will see the performance of different classifier. You will train multiple classifiers and check their accuracy in different settings for this problem, some of the mandatory classifiers are
kNN, Decision trees and  Neural Networks
Once you have trained your classifier, label the test data and upload the submissions. You will be able to see you performance of public leaderboard.
The performance of your classifier will be evaluated using accuracy:
[accuracy= number of correctly classified test instance/ total test instances]
You will upload your results, i.e labels of test data, as submissions. The accuracy shown on plublic leader board will be of 50% of the data. The accuracy for rest of the 50% of data will be on private leader board, that will be disclosed at the end of the competition.
You may submit upto 50 submissions each day