The more the better

Hi,

you cannot find out the minimum range of dataset. Ex: i took 30 samples of each class and I had 9 classes. Accuracy of this system was 11.11% , in another word, system did not train because 11.11% is the probability of to recognition anything in 9 classes.

in another experiment, i took 30 samples of each class and I had 170 classes, where i achieved 92.65% accuracy.

So, according to my understanding, either you will take help from the literature or hit and trail.

Thanks

So as we increase the data-set more chances to get better results. As I see many people take small data-set and get better results. Our proposed model will be more efficient if we take large amount data-set. My concern is that when we app;y deep learning everyone saying large amount data required. Neha Gautam

Hi,

Yes, if you increase the number of samples of each class so, there will be high changes to increase the accuracy.

The more the better

Thanks sir Sourabh Shastri

Thanks Neha Gautam