Computer-based techniques that aim to “find the pattern, apply the pattern.”
Machine learning algorithms typically have to go through a learning phase where it is given inputs of source training data, and may be given outputs of target training data. The algorithm is designed to learn how to model the output data based on the input data, or to learn how to detect and recognise patterns in the input data. The machine learning process typically requires vast amounts of data.
There are 2 types of learning. Supervised learning and unsupervised learning.
Like human beings, machines can be overtrained or undertrained. In machine learning terms, it is called overfitting and underfitting.« Back to Index