Page 341 - AI_Ver_3.0_class_11
P. 341
To do so, the first step is to train the machine with all different vegetables one by one, which may be done as follows:
● If shape of a vegetable is round with a depression at top and is
red in colour, then it will be labelled as tomato.
● If the shape of a vegetable is long and conical, although
cylindrical and nearly spherical, and is orange/red in colour,
then it will be labelled as carrot.
● If the shape of a vegetable is long finger-like, has a small tip at
the tapering end, and is green in colour, then it will be labelled
as a lady finger.
Now, if you show a new vegetable in front of the machine and ask the machine to identify it, since the machine has
already been trained from previous data, it will use the learned data wisely to classify the vegetable based on its shape
and colour and would confirm the vegetable.
More examples of classification problems include:
● Given a handwritten character, classify it as one of the known characters.
● Given recent user behaviour, classify it as churn or not.
In Artificial Intelligence, classification is the
process of labelling a set of data (structured or
unstructured) into different classes or groups METAL GLASS PLASTIC PAPER
where we can assign a label to each class. For
example, cities in India have different coloured
dustbins for different types of waste: green
coloured dustbins for biodegradable waste,
blue dustbins for non-biodegradable or plastic
waste, yellow dustbins for paper waste, and red dustbins for metallic waste. Hence, we classify the waste into four
different categories while also labelling each category.
How Classification Works?
In machine learning, classification involves sorting data into specific groups or classes based on their features.
Classification Process
Classes or Features or Classification Prediction or
Categories Attributes Training Model Inference
Here’s an overview of the process:
• Classes or Categories: Data is organised into different classes or categories, each representing a distinct outcome.
For example, a binary classification scenario has two classes: positive and negative.
• Features or Attributes: Each data instance is characterised by features or attributes that provide information about it.
These features are essential for the classification model to distinguish between different classes. For example, in email
classification, features might include words in the email, sender information, and the email subject.
Machine Learning Algorithms 339

