Is it better to train one neural network for a dispersed labeled data with large number of classes or first classify data by unsupervised learning then train each part by a separate NN? I mean by unsupervised learning we help each NN to classify in lower dispersed data with lower number of labels. So for test data the class of data is found by unsupervised learning then the final label is found by the network associated with that class. Does this question generally have an answer or it depends on data and needs to be answered in practice?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.