In the heart of the Supervised learning, it involves mathematical functions. From our math class, when we provide an input into a function, it will produces an output. The output is based on the mathematical characteristics of the function. The output can be known in advance or unknown to us.
What is Supervised learning? We as the supervisor, provide inputs to a machine, which performs some mathematical functions. Also, we know precisely what the outputs are.
The machine has a working relationship between the inputs and known outputs. The goal is to get the calculated output to be as close as possible to the known output. Our job is then to optimise the functions such that the calculated outputs match very closely to the actual outputs.
This means after tweaking the existing mathematical function such that each calculated output is as close as possible to the known output.
Or in a simplified version according to Professor Andrew Ng´s words; “we are going to teach the computer on how to do something”.
Supervised learning is categorised according to the output type:
- Regression, this is used in calculating continuous output.
- Classification, this is for discrete output.
Regression is to separate the data based on a best fit method. One of the regression methods is Linear regression. Another is Polynomial regression.
Classification is to organise the data into its common characteristics. k-nearest neighbors (k-NN) is an example of a classification method. Other method is Decision trees.
I appreciate any feedback that you may have after reading this post. Also if you think this will be beneficial to someone you know, do hit the share button or just the like button. Views expressed here are my own.
Jumpa lagi / See you later / Hasta la próxima / Bis bald
Global Professional 