Page 137 - Artificial Intellegence_v2.0_Class_12
P. 137

Introduction to Commonly used Algorithms

             he algorithms that transform a data set into a model are kno n as machine learning algorithms, and they are the heart
            of machine learning.  he appropriate algorithm to use  supervised, unsupervised, classification, regression, etc.)  ill rely
            on the type of problem you're trying to solve, the computational po er you have at your disposal, and the type of data
            you have.  et us no  study some commonly used AI algorithms.

                                                 Supervised Learning Algorithms

                         Regression                         Classi cation                    Decision Trees
                                                      .

                                                      .


                                                      .
                                                      .

                                                      .
                                                          .    .   .    .    .     .   .


              egression  predicts  a  real  number   he   classification   algorithm  Decision  trees,  one  of  the  most
              either rounded or  ith a decimal point)  categorises   the   dataset   into  popular   supervised   learning
             as  the  last  component  of  the  result.  groups  based  on  several  criteria.  algorithms,  are  named  for  their
              he technique  ill use both dependent   he  AI  model  learns  from  the  tree like structure  even if the tree is
             variable and an independent variable to  training  dataset   hen  employing  a  inverted).  he  roots  of the tree are
             predict a potential result.          classification  method,  and  it  then  the training datasets,  hich connect
                                                  divides the data into several groups  to particular nodes that represent test
                                                  based on  hat it has discovered.  attribute nodes. A node that doesn't
                                                                                   continue on is referred to as a  leaf
                                                                                   since nodes frequently lead to other
                                                                                   nodes.   y  branching  out  into  sub
                                                                                   nodes until it reaches the conclusion,
                                                                                   the decision tree can classify the data
                                                                                   that is presented to it.

             K Nearest Neighbour Algorithm
             It is one of the simplest  upervised  earning based  achine  earning algorithms.  he   Y
                  algorithm assumes similarity bet een the ne  case and the e isting cases and
             assigns the ne  instance to the category that matches the e isting cases the most
             closely.  he      algorithm can be applied to both classification and regression.

             Other examples include  - Support-vector machines,  Naive  Bayes,  and Linear             Category B
             discriminant analysis, and K Nearest.                                                   New Data Point

                                                                                             Category A
                                                                                                                X







                                                                                       C apstone  P roj e ct  135
   132   133   134   135   136   137   138   139   140   141   142