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
   193   194   195   196   197   198   199   200   201   202   203