Page 198 - AI Ver 3.0 class 10_Flipbook
P. 198
Model
Clustering Model Training
Unlabelled Data Clustering on the Basis of Type
Clustering – Example 1
Rohan is the owner of daily needs departmental store. He wants to improve his marketing strategies by
grouping customers based on their purchasing behaviour. The dataset contains different details about the
customers like age, purchasing power, and their annual expenditure (Money spent to buy from the store).
Your aim is to group the customers into clusters like:
a. High income high spending patterns
b. Young and low income customers with average spending patterns
c. Middle aged customers with moderate spending patterns.
In this example clustering is used where the algorithm will analyse the data and group customers into clusters
based on similarities in their spending behaviour and income.
The outcome will be clusters under the following categories:
• Cluster 1: High-income customers with high spending.
• Cluster 2: Young customers with low spending.
• Cluster 3: Middle-aged customers with moderate spending.
This is how clustering technique works. The clustering model will be able to identify clusters based on some
similarities or patterns which are not defined in the input. For example, age, purchasing power and the annual
amount spent on the purchases (in store) are the only features known, but clusters based on age and purchasing
power have been grouped together and given as output. Once the clusters are made, the store can offer
personalised discounts or loyalty programs to high-value customers and target low-spending customers with
promotional offers to increase their spending.
Clustering – Example 2
Tiya’s hobby is listening to music. She likes to listen to K-pop and Jazz songs whereas she dislikes songs with slow
tempo and low bass. We have grouped all the songs that belong to the Jazz and K-pop category in one cluster
that she likes, while songs with slow tempo into another cluster. Now if she listens to a new song with slow tempo,
could you predict if she likes the song X or not?
This is how clustering works; the clustering model will be able to identify clusters based on some kind of similarity
or pattern which is not defined in the input. For example, the only two features tempo and bass of the song is
known, the clusters based on likes and dislikes have been grouped together to give the output.
Similar techniques are used by OTT platforms like Netflix/Hotstar for recommendations.
Difference between Classification and Clustering
Classification: It's a supervised learning technique that assigns data points to predefined categories or labels.
Labelled data is fed to the model. For example, classifying emails as "Spam" or "Not Spam."
196 Touchpad Artificial Intelligence (Ver. 3.0)-X

