Outlier Detection & Removal | How to Detect & Remove Outliers (Updated 2023)

Introduction

In my previous article, I talked about the theoretical concepts of outliers and tried to find the answer to the question: “When should we drop outliers and when should we keep them?”. In this article, I will focus on outlier detection and the different ways of treating them. It is important for a data scientist to find outliers and remove them from the dataset as part of the feature engineering before training machine learning algorithms for predictive modeling. Outliers present in a classification or regression dataset can lead to lower predictive modeling performance.

I recommend you read this article before proceeding so that you have a clear idea about the outlier analysis in Data Science Projects.

Learning Objectives

1. An Overview of outliers and why it’s important for a data scientist to identify and remove them from data.
2. Undersand different techniques for outlier treatment: trimming, capping, treating as a missing value, and discretization.
3. Understanding different plots and libraries for visualizing and trating ouliers in a dataset.

This article was published as a part of the Data Science Blogathon

How to Treat Outliers?

There are several ways to treat outliers in a dataset, depending on the nature of the outliers and the problem being solved. Here are some of the most common ways of treating outlier values.

Trimming

It excludes the outlier values from our analysis. By applying this technique, our data becomes thin when more outliers are present in the dataset. Its main advantage is its fastest nature.

Capping

In this technique, wecap our outliers data and make the limit i.e, above a particular value or less than that value, all the values will be considered as outliers, and the number of outliers in the dataset gives that capping number.

For example, if you’re working on the income feature, you might find that people above a certain income level behave similarly to those with a lower income. In this case, you can cap the income value at a level that keeps that intact and accordingly treat the outliers.

Treating outliers as a missing value: Byassuming outliers as the missing observations, treat them accordingly, i.e., same as missing values imputation.

You can refer to the missing value article here.

Discretization

In this technique, by making the groups, we include the outliers in a particular group and force them to behave in the same manner as those of other points in that group. This technique is also known as Binning.

You can learn more about discretization here.

How to Detect Outliers?

For Normal Distributions

• Use empirical relations of Normal distribution.
• The data points that fall below mean-3*(sigma) or above mean+3*(sigma) are outliers, where mean and sigma are the average value and standard deviation of a particular column.

Source: sphweb.bumc.bu.edu

For Skewed Distributions

• Use Inter-Quartile Range (IQR) proximity rule.
• The data points that fall below Q1 – 1.5 IQR or above the third quartile Q3 + 1.5 IQR are outliers, where Q1 and Q3 are the 25th and 75th percentile of the dataset, respectively. IQR represents the inter-quartile range and is given by Q3 – Q1.

For Other Distributions

• Use a percentile-based approach.
• For Example, data points that are far from the 99% percentile and less than 1 percentile are considered an outlier.

Source: acutecaretesting.org

How to Detect and Remove Outliners in Python

Z-score Treatment

Output:

count    1000.000000
mean        6.961499
std         0.612688
min         5.113546
25%         6.550000
50%         6.960000
75%         7.370000
max         8.808934
Name: cgpa, dtype: float64

This completes our Z-score-based technique!

IQR Based Filtering

Used when our data distribution is skewed.

Step-1: Import necessary dependencies

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

Step-2: Read and load the dataset

df = pd.read_csv('placement.csv')
df.head()

Step-3: Plot the distribution plot for the features

plt.figure(figsize=(16,5))
plt.subplot(1,2,1)
sns.distplot(df['cgpa'])
plt.subplot(1,2,2)
sns.distplot(df['placement_exam_marks'])
plt.show()

Step-4: Form a box-plot for the skewed feature

sns.boxplot(df['placement_exam_marks'])

Step-5: Finding the IQR

percentile25 = df['placement_exam_marks'].quantile(0.25)
percentile75 = df['placement_exam_marks'].quantile(0.75)

Step-6: Finding the upper and lower limits

upper_limit = percentile75 + 1.5 * iqr
lower_limit = percentile25 - 1.5 * iqr

Step-7: Finding outliers

df[df['placement_exam_marks'] > upper_limit]
df[df['placement_exam_marks'] < lower_limit]

Step-8: Trimming outliers

new_df = df[df['placement_exam_marks'] < upper_limit]
new_df.shape

Step-9: Compare the plots after trimming

plt.figure(figsize=(16,8))
plt.subplot(2,2,1)
sns.distplot(df['placement_exam_marks'])
plt.subplot(2,2,2)
sns.boxplot(df['placement_exam_marks'])
plt.subplot(2,2,3)
sns.distplot(new_df['placement_exam_marks'])
plt.subplot(2,2,4)
sns.boxplot(new_df['placement_exam_marks'])
plt.show()

Step-10: Capping

new_df_cap = df.copy()
new_df_cap['placement_exam_marks'] = np.where(
    new_df_cap['placement_exam_marks'] > upper_limit,
    upper_limit,
    np.where(
        new_df_cap['placement_exam_marks'] < lower_limit,
        lower_limit,
        new_df_cap['placement_exam_marks']

Step-11: Compare the plots after capping

plt.figure(figsize=(16,8))
plt.subplot(2,2,1)
sns.distplot(df['placement_exam_marks'])
plt.subplot(2,2,2)
sns.boxplot(df['placement_exam_marks'])
plt.subplot(2,2,3)
sns.distplot(new_df_cap['placement_exam_marks'])
plt.subplot(2,2,4)
sns.boxplot(new_df_cap['placement_exam_marks'])
plt.show()

This completes our IQR-based technique!

Percentile Method

• This technique works by setting a particular threshold value, which is decided based on our problem statement.
• While we remove the outliers using capping, then that particular method is known as Winsorization.
• Here, we always maintain symmetry on both sides, meaning if we remove 1% from the right, the left will also drop by 1%.

Steps to follow for the percentile method:

Step-1: Import necessary dependencies

import numpy as np
import pandas as pd

Step-2: Read and Load the dataset

df = pd.read_csv('weight-height.csv')
df.sample(5)

Step-3: Plot the distribution plot of the “height” feature

sns.distplot(df['Height'])

Step-4: Plot the box-plot of the “height” feature

sns.boxplot(df['Height'])

Step-5: Finding the upper and lower limits

upper_limit = df['Height'].quantile(0.99)
lower_limit = df['Height'].quantile(0.01)

Step-6: Apply trimming

new_df = df[(df['Height'] <= 74.78) & (df['Height'] >= 58.13)]

Step-7: Compare the distribution and box-plot after trimming

sns.distplot(new_df['Height'])
sns.boxplot(new_df['Height'])

Winsorization

Step-8: Apply Capping (Winsorization)

df['Height'] = np.where(df['Height'] >= upper_limit,
        upper_limit,
        np.where(df['Height'] <= lower_limit,
        lower_limit,
        df['Height']))

Step-9: Compare the distribution and box-plot after capping

sns.distplot(df['Height'])
sns.boxplot(df['Height'])

This completes our percentile-based technique!

Conclusion

Outlier detection and removal is a crucial data analysis step for a machine learning model, as outliers can significantly impact the accuracy of a model if they are not handled properly. The techniques discussed in this article, such as Z-score and Interquartile Range (IQR), are some of the most popular methods used in outlier detection. The technique to be used depends on the specific characteristics of the data, such as the distribution and number of variables, as well as the required outcome.

Key Takeaways

• Outliers can be treated in different ways, such as trimming, capping, discretization, or by treating them as missing values.
• Emperical relations are used to detect outliers in normal distributions, and Inter-Quartile Range (IQR) is used to do so in skewed distributions. For all other distributions, we use the percentile-based approach.
• Z-score treatment is implemented in Python by importing the necessary dependencies, reading and loading the dataset, plotting the distribution plots, finding the boundary values, finding the outliers, trimming, and then capping them.

Frequently Asked Questions