What is Data Normalisation?

S3-SA3-0368

Grade Level:

Class 9

AI/ML, Data Science, Physics, Economics, Cryptography, Computer Science, Engineering

Definition

What is it?

Data Normalisation is a process of scaling data so that all values fall within a specific range, usually between 0 and 1, or -1 and 1. It helps make different types of data comparable by giving them a common scale without distorting differences in the value ranges.

Simple Example

Quick Example

Imagine comparing student marks from two different exams. One exam was out of 50 marks, and another was out of 100 marks. To compare fairly, we can normalise them. If a student got 40/50, it's 80% (0.8), and if another got 70/100, it's 70% (0.7). Now, 0.8 and 0.7 are comparable, showing who performed better relative to their respective exams.

Worked Example

Step-by-Step

Let's normalise a student's marks (out of 100) to a scale of 0 to 1. Suppose a student scored 75 marks. The minimum possible score is 0, and the maximum is 100.

Step 1: Identify the original value (X), minimum value (Min), and maximum value (Max).
X = 75
Min = 0
Max = 100

---Step 2: Use the Min-Max Normalisation formula: Normalised Value = (X - Min) / (Max - Min).

---Step 3: Substitute the values into the formula.
Normalised Value = (75 - 0) / (100 - 0)

---Step 4: Calculate the result.
Normalised Value = 75 / 100
Normalised Value = 0.75

---Answer: The normalised mark for 75 is 0.75.

Why It Matters

Data normalisation is crucial in fields like AI/ML, where different features (like age and income) need to be on the same scale for models to learn effectively. It's used by data scientists to build fair predictive models and by engineers designing systems that process varied sensor data. Understanding this helps you build smarter apps and analyze complex data.

Common Mistakes

MISTAKE: Applying normalisation only to a part of the data, not the whole dataset. | CORRECTION: Always calculate the minimum and maximum values from the entire dataset before applying the normalisation formula to all individual data points.

MISTAKE: Using normalisation when the data has extreme outliers, which can compress the useful range of other values. | CORRECTION: For data with significant outliers, consider other scaling methods like standardisation (Z-score normalisation) or robust scaling, which are less affected by extreme values.

MISTAKE: Confusing normalisation with standardisation. | CORRECTION: Normalisation scales data to a fixed range (like 0-1), while standardisation scales data to have a mean of 0 and a standard deviation of 1.

Practice Questions

Try It Yourself

QUESTION: Normalise the value 25 from a dataset where the minimum value is 10 and the maximum value is 50, using Min-Max Normalisation. | ANSWER: 0.375

QUESTION: A sensor records temperature readings in Celsius. If the lowest recorded temperature is -5 degrees Celsius and the highest is 35 degrees Celsius, normalise a reading of 15 degrees Celsius to a 0-1 range. | ANSWER: 0.5

QUESTION: You have a list of cricket scores: [50, 10, 120, 70, 30]. First, find the minimum and maximum scores. Then, normalise the score of 70 to a range of 0 to 1. | ANSWER: Min = 10, Max = 120. Normalised 70 = (70 - 10) / (120 - 10) = 60 / 110 = 0.545 (approx)

MCQ

Quick Quiz

Which of the following is the primary purpose of Data Normalisation?

To remove duplicate entries from a dataset

To scale data to a specific, common range

To convert data into text format

To increase the total number of data points

The Correct Answer Is:

Data Normalisation's main goal is to scale numerical data to a common range, like 0 to 1, making different features comparable. Options A, C, and D are not the primary purposes of normalisation.

Real World Connection

In the Real World

In an app like Google Maps, normalisation might be used to compare different factors for recommending a route. For example, traffic congestion (measured in minutes delay) and road quality (measured on a scale of 1-5) are very different. Normalising both to a 0-1 scale allows the app's algorithm to weigh them fairly to suggest the best route to your destination in your city.

Key Vocabulary

Key Terms

SCALING: Adjusting the range of data values | MIN-MAX NORMALISATION: A common method to scale data between 0 and 1 | DATASET: A collection of related data | FEATURE: An individual measurable property or characteristic of a phenomenon being observed | OUTLIER: A data point that differs significantly from other observations

What's Next

What to Learn Next

Great job understanding Data Normalisation! Next, explore 'Data Standardisation' (Z-score normalisation). It's another important scaling technique that handles outliers differently and is vital for many advanced data analysis tasks.