Improvement Analysis of Hand Written DigitRecognition using Convolutional Neural Network byVarying Values of CNN Parameters

Hand written digit recognition is the demand of this era. Everybody sharing and storing documents in digital format. Hand written digital documents needs to be processed to make it understandable for everyone. Researchers are writing CNN based codes to achieve perfection in HDR (Handwritten digit recognition).Although CNN has achieved excellent perfection but still there is chance of improvements in terms of speed and complexity. In order make more improvements in the results of HDR, improvement analysis has been done on CNN parameters. There are many positive points which came out of this analysis, like the optimized value of epoch, minibatch and alpha. It is also seen that every handwritten number digit has better recognition at certain values of the parameters, it concludes that if those values are used while verifying the results then HDR final results can be improved in terms of speed and accuracy.