Introduction
There are three main advanced categories in machine learning
- Supervised Learning
- Unsupervised Learning
- Reinforcement Learning
Supervised learning involves using labeled data to train algorithms to identify patterns and make predictions for new unlabeled data. The algorithms study the labeled data to learn how different patterns correspond to different labels and then use this knowledge to predict new data based on the identified patterns.
Supervised learning has two categories.
- Classification
- Regression
This article explains the basics of Linear Regression, a type of Regression Analysis.
Linear Regression
Regression models are used to predict numerical values, such as home prices and fraud detection. One specific type of regression analysis is linear regression, which predicts the value of a variable based on another variable. Linear regression is a powerful tool for modeling and analyzing relationships between variables, and it can be used in various applications.
Let’s gain a deeper understanding of this concept, when we observe a pattern or relationship in a scatterplot, we can use a straight line to summarize the relationship and make predictions based on it. This method of summarizing and predicting is known as linear regression.
The value we aim to predict in predictive modeling is known as the dependent variable. In contrast, the variable that we use to make predictions is known as the independent variable. This terminology reflects the fact that the value of the dependent variable depends on the value of the independent variable.
Equation
This article aims to provide a simplified explanation of linear regression. If the concept is still unclear, one approach is to look into the equation used to fit the line, which can help to clarify the relationship between the variables.
Y = w1x + w2
‘Y’ is the response, the dependent variable, and the value we want to predict.
‘X’ is the predictor, the independent variable.
w1 = Slope
w2 = y-intercept, where the line intersects with Y-axis.
Moving a Line
When using linear regression, it is essential to fit a line to the data that optimizes the accuracy of the model. This process involves finding the best-fit line that minimizes the difference between the predicted and actual output values in the training data. There are four different ways in which the parameters of a linear regression model can be adjusted to change the shape and position of the line that is used to fit the data.
- Increasing the slope (w1)
- Decreasing the slope (w1)
- Increasing the y-intercept (w2)
- Decreasing the y-intercept (w2)
Increasing the slope (w1)
If we increase the value of w1, the line goes upwards.
Decreasing the slope (w1)
If we decrease the slope (w1), the line goes downwards.
Increasing the y-intercept (w2)
Increasing the value of w2 causes the line to shift upwards while remaining parallel to its original position.
Decreasing the y-intercept (w2)
Decreasing w2 causes the line to shift downwards towards the bottom of the graph.
Conclusion
The article provides an introduction to the fundamentals of linear regression, including an explanation of the equation used to fit a line to the data and examples of how moving the line can be visualized using simple graphs to aid in comprehension.