Page 300 - Artificial Intellegence_v2.0_Class_11
P. 300

One  more  common  example  of  classification  is  to  identify  spam  emails.  To  create  an  email  spam  filter  program,  a
              developer can train a machine learning algorithm with a set of spam-like emails labeled as “spam” and normal emails
              labeled as “not-spam”.
              The reason behind this is to train an algorithm that can identify spam emails from a set of emails and filter out spam
              emails.


























              In order to understand ‘Classification’, let us revise the concept of ‘Supervised Learning’, because classification is type of
              supervised learning.

              Supervised learning is a subcategory of Machine Learning and Artificial Intelligence. It is also known as supervised
              machine learning. In supervised learning, we train the machine using well-labelled data, which means the data is already
              marked with the correct answer. After that, the machine is given a new dataset (training data) so that the supervised
              learning algorithm can analyse this data and produce a correct outcome.

              Let us take an example of vegetables. Now, you want to create a machine that can identify the vegetables one by one.
              To do so, the first step is to train the machine with all different vegetables one by one as follows:

              •   If shape of a vegetable is rounded with a depression at top and
                  Red in colour, then it will be labeled as Tomato.
              •   If shape of a vegetable is long and conical, although cylindrical
                  and nearly-spherical cultivars and orange/red in color, then it will
                  be labeled as carrot.
              •   If shape of a vegetable is long finger like, having a small tip at the
                  tapering end and green in colour, then it will be labeled as 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 trained from previous data, it will use the learnt data wisely to
              classify the vegetable based on its shape and color 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 as churn or not.
              Supervised learning is further categorised into two:
              •  Classification: Process of categorizing a dataset into classes or groups, for example, “Orange” or “Green” or “spam”
                  and “not spam”.

              •  Regression: Process used to predict real or continuous values, for example, salary, rainfall, product sales etc.
                    298     Touchpad Artificial Intelligence (Ver. 2.0)-XI
   295   296   297   298   299   300   301   302   303   304   305