Various ways to evaluate a machine learning model's performance
The task of deep learning and neural networks is to learn and predict. You can use them to learn and predict any problem such as boiler problem.
ANN is based on the method of geometric transformations and is therefore suitable for this type of problem. You can read more about R. Tkachenko and I. Izonin, ‘Model and Principles for the Implementation of Neural-Like Structures Based on Geometric Data Transformations’, in Advances in Computer Science for Engineering and Education, vol. 754, Z. Hu, S. Petoukhov, I. Dychka, and M. He, Eds. Cham: Springer International Publishing, 2019, pp. 578–587.
Dear, Irsa Talib!
You should conduct a series of studies on all ML algorithms, and then build an ANN model. And then, comparing the results of the accuracy of the model, make a conclusion about the optimality of the algorithm or model.
Dear Irsa,
In fact, ANN is the general name for neural networks. Of course the deep neural networks (e.g., DNN, CNN, LSTM, etc.) is much accurate than shallow networks (e.g., MLP).
You can use evaluations criteria, confusion matrix or ROC to estimate the model performance. Also, utilise the benchmark classifiers for justification and model control.
Dear Irsa Talib,
A shallow network (Neural Network) has less number of hidden layers as compared to Deep neural Network.
- Badges
- Science topic