When you think of machine learning for a particular application, you will come across a number of learning algorithms. You may be confused regarding which one to consider. Before you delve deeper into confusion, it is important to keep in mind the basics. All the machine learning algorithms consist of three fundamental components - representation, evaluation and optimization. The following is a short discussion about these three components.
Representation - A classifier is to be represented in a formal language. You have to be really meticulous about selecting the classifiers and the representation for the learning algorithm. Classifiers need to be chosen aptly so that they can be learned easily which will in turn impact the selection of the representation.
Evaluation - How do you differentiate the efficient classifiers from the inefficient ones? It is by evaluation function. The algorithm uses the evaluation function internally to distinguish the classifiers and this may be unlike the one run externally.
Optimization - You need to identify the best classifiers in the language and for that, optimization is necessary. Optimization is the technique to spot the most efficient classifier and it is essential to choose the most viable method of optimization as this will impact the efficiency of the learner.
So, these are the three basic components of machine learning. You are required to keep these basics very much clear for using machine learning for applications.
I will come up with more write-ups about machine learning and its components. Machine learning is a vast field and in the last few years, it has gained big momentum. It is interesting, helpful and viable along with being cost-effective.
Share your thoughts and comments till Saby Uphadhyay returns with the next post. Have a great day!
