What is ML?

Basic idea is to use data from the past to predict unknown outcomes.

Steps involved

Training

Training data has past observations. This includes: features and label. Feature is basically what you are observing (x). And label is the thing you want to train for (y).

Algorithm

An algorithm is applied to the training data to determine relation between features and label.

Result

Result is a model. The so called f in y=f(x).

Inferencing

After training is completed, it can be used to get output given x.

---

references: https://learn.microsoft.com/en-in/training/modules/fundamentals-machine-learning/2-what-is-machine-learning