When working on regression-based machine learning problems, sometimes we find that several independent features correlate not only with the dependent features but also with each other. This is what multicollinearity means. In this article, I’ll take you through an introduction to multicollinearity in machine learning.

## What is Multicollinearity?

When there is a linear relationship between independent features in a dataset, it is nothing more than multicollinearity. While working on a regression-based problem, it is good to have a strong correlation between *independent and dependent features*, but if the independent features are not only correlated with the dependent features but also with the independent features, it may affect the performance of a machine learning model.

To detect if there is multicollinearity in your dataset, you can look at the correlation with respect to the independent features. Some data visualization techniques such as a scatter plot or a heatmap can also be used to detect if there is a correlation between independent features.

## How Multicollinearity Affects a Machine Learning Model?

One of the ** assumptions** of the linear regression algorithm is that there should be very little or no multicollinearity in the data set. This means that you should only use a linear regression algorithm when there is very little or no correlation between the independent features of a dataset. So if you use a linear regression model in such a dataset, your model will fail to predict labels.

So when working on a regression-based problem where the dataset has a higher degree of multicollinearity in your dataset, it is not better to use a linear regression algorithm. In such cases, you can use the Decision Trees algorithm which is not affected by multicollinearity in the dataset.

### Summary

When there is a linear relationship between independent features in a dataset, it is nothing more than multicollinearity. To detect if there is any correlation between the independent features in your dataset, you can look at the correlation with respect to the independent features. I hope you liked this article on what is multi-collinearity in Machine Learning. Feel free to ask your valuable questions in the comments section below.