Page 284 - Artificial Intellegence_v2.0_Class_11
P. 284

Example 4: Consider regression equation for 55 students with x = right forearm length (cm) and y = height. The forearm
              lengths were in the range from 22 cm to 31 cm. The regression equation is:
                                                           y = 30.3 + 1.49x
              Let us see some practice questions.
              1.   One student’s right forearm length was 27 cm, and his height was 76 inches. What is the estimated height for this
                  student, based on the regression equation? What is the residual?

              Answer:     Estimated Height => 30.3 + 1.49*27 = 70.53
                          Residual = y – ŷ => 76 – 70.53 = 5.47

                  Note that Residual = Observed value – Predicted value (e = y – ŷ). Each data point has one residual.
              2.   One student’s right forearm length was 22 cm, and her height was 62 inches. What is the estimated height for this
                  student, based on the regression equation? What is the residual?
              Answer:     Estimated Height => 30.3 + 1.49*22 = 63.08

                          Residual => 62 – 63.08 = –1.08



                             Reboot

                    1.   State the two types of Regression.

                    2.  How many variables are used in linear regression?
                    3.  State the equation of the line of best fit?

                    4.  Why is it called the line of best fit?

                    5.  State two applications of regression.





                      Correlation

              The word correlation is used in daily life to denote some forms of association. We might say that we have noticed a
              correlation between smog and asthma attacks. However, in statistical terms, we use correlation to express an association
              between  two  quantitative  variables.  It  measures  the  strength  or  degree  of  relationship  between  two  variables.  The
              relationship may be causal. We also presume that the association is linear, i.e., one variable increases or decreases a set
              amount for a unit increase or decrease in the other.

                           Perfect       High        Low                      Low        High       Perfect
                           Positive    Positive     Positive      No       Negative     Negative    Negative
                         Correlation  Correlation  Correlation  Correlation  Correlation  Correlation Correlation

















                    282     Touchpad Artificial Intelligence (Ver. 2.0)-XI
   279   280   281   282   283   284   285   286   287   288   289