[Supervised learning application] Classification, annotation and regression-notes

Learn

A system can improve its performance by performing a process, this is learning

Supervised learning

Supervised learning refers to learning prediction models from labeled data.

The labeled data indicates the corresponding relationship between input and output, and the prediction model produces corresponding output for a given input. The essence of supervised learning is to learn the statistical law of the mapping of input to output.

 

The application of supervised learning is mainly in three aspects: classification, labeling and regression

classification

• Supervised learning learns a classification model or classification decision function from the data, which is called a classifier. The classifier predicts the output of a new input, which is called classification.
• The possible output is called class
• The classification problem includes two processes of learning and classification

Callout

• It can be considered that the labeling problem is a generalization of the classification problem. But the labeling problem is a simple form of the more complex structure prediction problem.
• In the classification problem, the output variable Y has a finite number of discrete values, and the input X can be discrete or continuous.
• The input of the labeling problem is an observation sequence, and the output is a label sequence or state sequence.
• The goal of the labeling problem is to learn a model, and after observing the sequence, a prediction is obtained, which is a labeled sequence
• The number of possible markers is limited, but the number of marker sequences called by the combination increases exponentially according to the sequence length
• The labeling problem includes two processes of learning and labeling

return

• Regression is used to predict the relationship between input variable X and output variable Y
• The so-called regression model is a function of mapping between input and output, Y = F (X)