top of page

What is Machine Learning and Explain its working ?

Updated: Apr 23



Machine Learning (ML) is a branch of artificial intelligence (AI) and computer science that focuses on using data and algorithms to enable AI to mimic the way humans learn and increasingly improve its accuracy.

Let's compare this to traditional computing, which is based on deterministic systems where we explicitly tell the computer a set of rules to perform a particular task. This method of starting computer programs is called rule-based. Instead of machine learning being discrete and expendable, rules-based programming can choose those rules itself.

Let's say you are a bank manager and you want to know whether the loan applicant will be able to repay the loan. In a rules-based system, the bank manager (or other professional) directly informs the computer that the application will be rejected if the applicant's credit score is below a certain level.



 Machine Learning
Machine Learning


How does machine learning work?


Understanding Machine Learning

Let's start our journey by highlighting the basic concepts:


1. Supervised Learning:

In simple terms, you can think of this as teaching a computer using examples for which we already know the answers.

Example: Estimating house prices based on factors such as size, location and number of rooms.

2. Unsupervised Learning:

In simple terms, we show the computer patterns in the data without telling it what to look for.

Example: Group customers with similar purchasing history without predefined categories.

3. Reinforcement Learning:

Simply put: it is like training animals; A computer learns by trying different actions and observing the results.

Example: Teaching a computer program to play a game and to get better at it over time.



UC Berkeley divides machine learning algorithms into three main areas.


Decision Process: Machine learning algorithms are often used for predictions or classification. Given some input data, labeled or unlabeled, your algorithm generates a prediction that matches the pattern in the data.

Error function: An error function that evaluates predictions. If the samples are known, the error function can be used to estimate the accuracy of the model.

Model deployment method: If the model can correctly fit the data in the training set, the parameters are adjusted to reduce the difference between the known model and the model. The algorithm repeats this “check and fix” process, automatically updating the parameters until all parameters are met.




1 view0 comments

Comments


bottom of page