1.2. Disadvantages:
(i) Large dimensionality.
(ii) Results are always generated even if the input data are unreasonable.
(iii) They are not scalable i.e. once an ANN is trained to do certain task, it is difficult to extend for other tasks
without retraining the neural network.