1. What is a Project Cycle?
Definition:
A Project Cycle is a systematic series of steps or phases
followed to develop and
complete a project from start to finish. It helps ensure that
a project is well-planned,
executed efficiently, and achieves its intended goal.
In AI Context:
The AI Project Cycle is a structured framework that helps
students and professionals
build AI-based solutions to real-world problems by
following six organized phases.
Importance:
● Helps understand and define the problem clearly.
● Provides direction for data collection and model building.
● Encourages ethical thinking and innovation.
2. The Six Phases of the AI Project Cycle
Phase 1: Problem Scoping
Definition:
Problem scoping is the phase where we understand,
define, and narrow down the
actual problem that needs to be solved using AI.
Purpose:
To identify the need, the people affected, and the context
in which the problem exists.
Tool: 4W's Canvas (CBSE Prescribed Format)
Problem Statement Template:
The Problem Statement Template is a structured way to
summarise the real-world
problems that you want to solve using AI. It helps you
define the goal, object, and
action needed to create a meaningful solution.
“I want to [action] [object] in order to [goal].”
Phase 2: Data Acquisition
Definition:
Data acquisition is the process of collecting relevant data
from various sources that will
be used to train and test the AI model.
Sources of Data:
● Public datasets
● Government portals
● Sensors & IoT devices
● Online databases and APIs
Types of Data:
● Structured Data – Tables, rows (like Excel files)
● Unstructured Data – Images, audio, videos, text
Goal:
To gather the right amount of clean, relevant data for
analysis and modeling
Phase 3: Data Exploration
Definition:
Data exploration involves understanding, organizing,
cleaning, and visualizing the
collected data to identify useful patterns and remove
errors.
Key Activities:
● Remove duplicates or missing values
● Identify trends using charts and graphs
● Classify data types
Data Features:
● Features are important variables or columns that are
used by AI to make
predictions or [Link] stands for the type of data and
its characteristics.
● Example: In a weather prediction system, features can
be temperature, humidity,
and wind speed.
Phase 4: Modelling
Definition:
Modeling is the phase where a machine learning model is
selected and trained using
the dataset to learn patterns and make predictions.
Steps:
1. Select a suitable ML algorithm (e.g., decision tree,
regression, etc.)
2. Train the model using training data
3. Test it with unseen data
Types of Learning Models:
● Supervised Learning – Labelled data is used.
● Unsupervised Learning – Data is unlabelled; AI finds
patterns itself
Phase 5: Evaluation
Definition:
Evaluation means testing the model’s accuracy, efficiency,
and effectiveness in solving
the problem.
Common Evaluation Metrics:
● Accuracy – % of correct predictions
● Precision & Recall – Used for classification problems
● Confusion Matrix – Compares actual vs predicted results
Goal:
To ensure the model performs well and is reliable.
Phase 6: Deployment
Definition:
Deployment is the final phase where the trained AI model
is integrated into a real-world
application for users to interact with.
Examples of Deployment:
● Chatbots in websites
● Face detection in phones
● AI-powered apps (like Google Maps traffic alerts)
Goal:
To make the AI solution usable and beneficial to its users.
