Page 288 - Artificial Intellegence_v2.0_Class_11
P. 288

Correlation is not Causation

              The correlation is a statistical method that indicates whether a pair of variables has a linear relationship and will change
              together. It does not state the reasons of the relationship, but it tells that a relationship does exist.
              Causation shows that an event is the direct result of the occurrence of another event, i.e. a causal relationship exists
              between the two events. This is also called cause and effect. For example, a speeding car leads to an accident. The
              accident is due to causation.
              Causation takes a step ahead than correlation. It states that any change in the value of one variable will definitely cause
              a change in the value of the second variable. This that one variable makes the other happen. This is also called as cause
              and effect.
              In statistics, the phrase "correlation does not imply causation"
              means  that  the  relationship  between  two  variables  cannot
              be  reasonably  deduced  based  solely  on  their  observed
              association.
              •     "Correlation is not causation" means that if two things are   Causation           Causation
                  related, does not, necessarily mean that one thing leads to
                  the other.
              •     For example, just because Indians tend to eat more in cold
                  weather and less in hot weather does not mean that cold
                  weather leads to crazy shopping for eatables.                            Correlation
              •     Another example is due to less RAM, our mobile phone   Sale of fans                Consumption of
                  freezes. This means no playing games or text messaging                                  ice-creams
                  through the phone.



                        At a Glance


                    •  Regression is a Supervised machine learning algorithm.
                    •  Linear regression is an algorithm used to predict a relationship between two different variables.
                    •  A scatter plot is a graph which uses Cartesian coordinates to display values for mainly two variables in a
                     dataset.
                    •  To plot the points on the scatterplot, you show each one as an ordered pair.
                    •  Outliers are data points on the scatterplot that do not follow the pattern of the dataset.
                    •  Regression focuses on predicting the value of the variable that is dependent on the second variable.
                    •  The equation for linear regression is given by y = mx + b where m is the slope and b is the y-intercept.
                    •  The line that passes close to most of the data points is called the ‘Line of Best Fit’ or ‘Regression Line’.
                    •  The vertical distance between the observed responses in the dataset and the line of best fit is called the
                     residual error (e). Each data point has one residual.
                    •  Correlation is used to express an association between two quantitative variables.
                    •  The correlation coefficient is measured on a scale that varies from + 1 to –1.
                    •  "Correlation is not causation" means that if two things are related, does not, necessarily mean that one thing
                     leads to the other.
                    •  Crosstabs determine a relationship between two variables. This relationship is exhibited in tabular form.







                    286     Touchpad Artificial Intelligence (Ver. 2.0)-XI
   283   284   285   286   287   288   289   290   291   292   293