Module 8.8: Mean Absolute Percentage Error (MAPE)

Introduction

Mean Absolute Percentage Error (MAPE) is one of the most widely used evaluation metrics in Machine Learning and Artificial Intelligence for regression problems.

MAPE measures prediction accuracy by calculating percentage errors between actual values and predicted values.

It helps determine how much prediction error exists in percentage form.

This metric is widely used in Artificial Intelligence, Forecasting Systems, Finance Analytics, Sales Prediction, Demand Forecasting, and Business Intelligence.

Learning Objectives

• Understand Mean Absolute Percentage Error (MAPE).
• Learn regression evaluation.
• Understand percentage-based prediction error.
• Learn MAPE calculation.
• Explore real-world applications.
• Understand advantages and limitations.

What is Mean Absolute Percentage Error (MAPE)?

Mean Absolute Percentage Error (MAPE) is an evaluation metric used to measure prediction accuracy in regression models.

It calculates the average percentage difference between actual values and predicted values.

Unlike MSE and RMSE, MAPE expresses errors in percentage form, making interpretation easier.

In simple words:

MAPE measures how much prediction error exists as a percentage.

MAPE Formula

MAPE is calculated using actual values and predicted values.

MAPE=\frac{100}{n}\sum_{i=1}^{n}\left|\frac{y_i-\hat{y}_i}{y_i}\right|

Where:

• n = Total Number of Data Points
• y_i = Actual Value
• ŷ_i = Predicted Value

How MAPE Works

MAPE generally follows these steps:

1. Find actual values.
2. Find predicted values.
3. Calculate prediction differences.
4. Divide errors by actual values.
5. Take absolute values.
6. Calculate average percentage error.

Simple Example of MAPE

Suppose a sales prediction model generates the following results:

Step 1: Calculate Percentage Errors

• |100−90| /100 = 0.10
• |200−210| /200 = 0.05
• |300−270| /300 = 0.10

Step 2: Calculate Average

MAPE=\frac{(0.10+0.05+0.10)}{3}\times100=8.33\%

Final MAPE:

8.33%

Why MAPE is Important

MAPE is important because it expresses prediction errors as percentages.

This makes model performance easier to understand for both technical and non-technical users.

Low MAPE means:

• Smaller prediction errors.
• Higher prediction accuracy.
• Better forecasting performance.
• Improved model reliability.

MAPE vs RMSE vs MSE

MAPE in Artificial Intelligence

Artificial Intelligence systems widely use MAPE for regression model evaluation.

Applications include:

• Sales Forecasting
• Revenue Prediction
• Stock Market Forecasting
• Demand Forecasting
• Business Analytics
• Supply Chain Prediction

Real-World Applications

1. Finance

Financial institutions use MAPE for forecasting market behavior and investment analysis.

2. Retail and E-Commerce

Companies use MAPE for sales prediction and customer demand forecasting.

3. Business Analytics

Organizations measure revenue prediction performance using MAPE.

4. Supply Chain Management

Businesses forecast inventory and logistics demand using MAPE.

Basic Python Example

actual = [100,200,300]

predicted = [90,210,270]

mape = (
(abs(100-90)/100) +
(abs(200-210)/200) +
(abs(300-270)/300)
) / 3 * 100

print(mape)

Output:

8.333333333333334

This example demonstrates simple MAPE calculation for regression evaluation.

Advantages of Mean Absolute Percentage Error

• Easy percentage interpretation.
• Useful for business forecasting.
• Compares models easily.
• Widely used in regression evaluation.
• Helpful for non-technical understanding.

Limitations of Mean Absolute Percentage Error

• Cannot handle zero actual values properly.
• Sensitive to very small actual values.
• Mainly useful for regression tasks.
• May create misleading results in certain datasets.

Key Concepts

• MAPE evaluates regression model performance.
• Measures average percentage prediction error.
• Low MAPE indicates better accuracy.
• Useful for forecasting applications.
• Widely used in Artificial Intelligence.

Interview Questions

1. What is Mean Absolute Percentage Error (MAPE)?

MAPE is a regression evaluation metric that measures average prediction error in percentage form.

2. Write the MAPE formula.

MAPE=\frac{100}{n}\sum_{i=1}^{n}\left|\frac{y_i-\hat{y}_i}{y_i}\right|

3. Why is MAPE easy to interpret?

MAPE uses percentage values, making prediction errors easy to understand.

4. Give examples of MAPE applications.

Sales Forecasting, Revenue Prediction, Finance Analytics, and Demand Forecasting.

Assignment

1. Define Mean Absolute Percentage Error (MAPE).
2. Write the MAPE formula.
3. Differentiate MAPE, RMSE, and MSE.
4. Explain percentage-based prediction error.
5. List five real-world applications.

Quiz

Q1. MAPE is mainly used for?

• A. Classification Problems
• B. Regression Problems
• C. Browser Testing
• D. HTML Styling

Answer: B. Regression Problems

Q2. MAPE expresses prediction errors in?

• A. Binary Format
• B. Percentage Form
• C. Color Format
• D. Matrix Format

Answer: B. Percentage Form

Q3. Low MAPE indicates?

• A. Poor Model Performance
• B. Higher Error Rate
• C. Better Prediction Accuracy
• D. Database Failure

Answer: C. Better Prediction Accuracy

Summary

In this tutorial, you learned Mean Absolute Percentage Error (MAPE) and its importance in AI Model Evaluation.

You explored the MAPE formula, regression evaluation, percentage-based error calculation, applications, advantages, limitations, and real-world examples.

Understanding MAPE is essential because it is one of the most widely used regression evaluation metrics for forecasting and predictive analytics in Machine Learning and Artificial Intelligence.

Next Tutorial

Module 9: Deep Learning Fundamentals

“`