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
   336   337   338   339   340   341   342   343   344   345   346