Machine Learning Algorithms

According to Arthur Samuel (1959), Machine Learning is a field of study that gives computers the ability to learn without being explicitly programmed.

Tom Mitchell (1998) Well-posed Learning Problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.

Machine learning algorithms

  1. Supervised learning
  2. Unsupervised learning
  3. Reinforcement learning

Supervised learning

It is machine learning task of at function that maps an input to on output based on example input-output pair. Basically Supervised learning is learning in which We teach or train the machine using data which is well labelled that means Some data is already tagged with correct answer. We pass data, train it and predict output.

Example 1 – House price prediction : In this data set can be given a which contain locality, size of house, age, no. of rooms, price at which it sell. In this are example locality, size of house are independent variables from which we can we predict house price. In this we can take prices of other houses to train our data. We take real prices map them and can predict price.

Example 2 – If we have different kinds of Fruits. To train the machine with all different Fruits one by one like shape of fruit, colour. Since machine has already learned from previous data this time it will classify fruit with its colour and shape and give output.

  • supervised learning allows collecting data and produce data output from previous experience.
  • It helps to solve various types of real world computation problems.

Unsupervised learning:

It is the training of machine using information that is neither classified nor labeled and allowing the algorithm to act on that without guidance. In this past data is pointless and we need to club similar data together. There is no way to measure similarity before we run the program. It is less accurate.

Example 1 – Google news : In google news, clustering where they use to club similar types of news together. They find some keywords, club similar news and show it on feed.

Example 2 – Feature selection : Assume that we want to predict how capable an applicant is of repaying a loan from the perspective of a bank now, we need to help the bank set up machine learning system so that each loan given to applicant who can repay the loan. So by gathering the Information about applicants average monthly income, debt credit & history we can predict this

Reinforcement learning

It means to establish & encourage a pattern of behavior. It is area of machine learning concerned with how software agent ought to take actions in an environment in order to maximize the notion of cumulative reward.

Example 1 – Chess game : In chess game there are different types of pieces which can move differently. The next move will depend om opponent move or your previous. It is trial and error and decision is dependent.

Example 2 – Web system configuration : there are so many parameters in web system and the process of tuning the parameters requires a skilled operator. This can be automated by using reinforcement machine learning technique to learn from different trial & error phases.