Summary

The web content discusses the application of Z-scores in machine learning for data standardization and pre-processing, emphasizing their role in normal distribution and comparison across different data sets.

Abstract

The article "Z-Distribution or Z-Score Application in Machine Learning" elucidates the concept of Z-scores as a method for standardizing data to a uniform scale. It explains how Z-scores quantify the distance of data points from the mean in terms of standard deviations, which is crucial for interpreting data distribution. The Z-score formula is provided, and the article distinguishes between normal distribution and skewed distributions (right and left). It also covers the calculation of mean, variance, and standard deviation, illustrating how these statistics relate to the Z-score and the distribution of data. The significance of Z-scores in machine learning is highlighted, particularly in data pre-processing to ensure consistency and comparability of data. The article concludes by emphasizing the importance of standard scaling in machine learning algorithms and invites readers to engage with the author on LinkedIn and Twitter.

Opinions

The author considers standard scaling a critical step in the data pre-processing phase of machine learning.
The article suggests that understanding Z-scores is essential for data scientists and machine learning practitioners.
It is implied that Z-scores facilitate the comparison of data from different standard distributions, leading to better results in machine learning applications.
The author provides a subjective recommendation to readers to explore further articles on related topics, indicating a belief in the value of continuous learning and exploration in the field of machine learning.

Z-Distribution or Z-Score Application in Machine Learning

Data pre-processing concept

In this article, the Z-score is a way to standardize the data to standard scale i.e. how far the data point is from the mean. The z-score can come positive or negative based on the help of mean and standard deviation values.

The data point away from the mean with some standard deviation is called a z-score.

The formula to get the z-score is shown below:

z = (data point - mean)/standard deviation

The z-score can be perfectly found in a normal distribution curve with no left skew and right skew. The below image shows these curves.

Normal Distribution: The normal distribution is a curve in which the data is spread symmetrically on both sides of the mean.

Right Skew: The data is mostly skewed on the right side because most of the data is on the right side. If we talk about outliers they are mostly on the right side too.

Left Skew: The data is mostly skewed on the left side because most of the data is on the left side. If we talk about outliers they are mostly on the left side too.

Machine Learning: Explanation of Simple Linear Regression Algorithm

A branch of artificial intelligence

medium.com

Introduction to Jupyter Notebook and its Features

Jupyter Notebook is an open-source web application

pub.towardsai.net

Mean and Standard Deviation

Mean: The mean is an average value of the data that tells about the center value of the data.

Standard deviation: It is a spread of the data around the mean with one standard deviation.

Finding the one standard deviation from the raw data

Let’s have raw data

the mean of the data is

mean = total sum of values/total number of data = 202/11 = 18.36

Let’s draw the distribution curve along with the data point for a clear visual understanding.

variance: It is the average sum of the squared difference value from the mean. As we see the difference between the number and the average of the data.

Standard deviation: When we do the square root of variance we get a standard deviation. This value of the standard deviation is the one standard deviation. The below image shows the steps to calculate the standard deviation.

In the below image the green shadow is the one standard deviation from the mean.

The 1 standard deviation consists of 68% of the data. The 2 standard deviations consist of 95% of the data. The 3 standard deviations consist of 99.7% of the data.

After finding the z-score of all the data we can draw the Standard deviation distribution.

The normal distribution curve with the spread of the data with their standard deviation value.

Application of z-score in Machine learning

To standardize the data as a part of data pre-processing.
To compare the z-score values of different standard distributions for better results.

Conclusion

This article will guide you through a brief introduction to the standard deviation distribution curve. Standard scaling is a crucial process in the data pre-processing of the machine learning algorithm.