Outliers are those data points that are distant from most of the similar points – in other words, we can say the outliers are entities that are different from the crowd. Outliers cause problems when it comes to building predictive models, such as long model training times, poor accuracy, an increase in error variance, a decrease in normality, and a reduction in the power of statistical tests.
There are two types of outliers: univariate and multivariate. Univariate outliers can be found in single variable distributions, while multivariates can be found in n-dimensional spaces. We can detect and handle outliers in the following ways:
- Box Plot: We can use a box plot to create a bunch of data points through quartiles. It groups the data points between the first and third quartile into a rectangular box. The box plot also displays the outliers as individual points using the interquartile range.
- Scatter Plot: A scatter plot displays the points (or two variables) on...