Fundamentals of Machine Learning that Every Beginner Should Know.
Hello Learners! This is a basic and important Fundamental of Machine Learning who should know to all beginners.
Table of contents
The fundamentals of Machine Learning are shown in 5 concepts.
- Labels
- Features
- Models
- Regression
- Classification
1. Labels
Labels are the variables whose value we predict.
A label can be any value we predict, like the future price of the stock, the type of animal in the image, or even predict what word the person will type next.
So, any value we predict is represented as a label in machine learning.
While this is not the rule of thumb, most machine learning practitioners represent the labels by Y.
2. Features
Since labels are something that we predict, so we can say that labels are the output values we get after using machine learning.
Features are input values that we use to predict labels.
Again, this is not the rule of thumb, but most practitioners represent features by X.
Now features are of two types dependent and independent.
Independent features are represented on the x-axis and the dependent features are represented on the y-axis.
3. Models
A model defines the relationship between labels and features which are the input variables and the predicted value.
There are two stages in the model process which are training and testing.
Training: Training means that we fit the data on the algorithm so that the model learns from features.
Testing: Training means testing the model on new unseen examples performed on a new set of data.
4. Regression
in Machine Learning, there are two types of models, and regression is one of them.
Regression means predicting continuous values.
The type of problems where regression models are used like predicting the value of stock prices, predicting the value of house prices, etc.
Regression models are used while predicting future values by using historical and present data.
5. Classification
As mentioned above, there are two main types of models in machine learning, classification models are the other.
Classification models are used to predict discrete values.
The types of problems where classification models are used are problems such as classifying email as spam or not spam and classifying whether the person is wearing a mask or not.