Classification is a supervised machine learning technique. It is the process of finding patterns in given data with the help of mapping functions. Let’s understand Mapping Function.

Mapping functions are the mathematical expressions which converts input data to any one of the possible outputs or categories/labels. Mapping function is also known as classifier and the underlying algorithm is known as classification algorithm.

Depending upon complexity of application, Classifiers can be simple or complex. Some of the well known classifiers are Decision Tree, Logistic Regression, Artificial Neural Networks (ANN), K-Nearest Neighbour (KNN) and Naive-Bayes (NB) model. None of these classifiers is fit for all problems but suits in specific requirements.

The Naive-Bayes and Logistic regression classifiers have low variance. The prediction models built using these classifiers are more tolerant with the underlying change in input training data. These algorithms require only a few parameters to tune, which is ideal for quick learning and annotation. While, decision tree and KNN classifiers are more flexible. These classifiers are very sensitive to high variance which may cause major fluctuations in the prediction model with the changes in input training data.

Classifier such as Artificial Neural Networks (ANN) are complex systems which requires high computational power as well as time to tune several parameters. These classifiers uses back propagation algorithm and multi-layer feed forward networks to optimize the classification outcome. ANNs are widely used in Image classification, Character Recognition and sales prediction etc.