Page 135 - Data Science class 11
P. 135

Forecasters are accountable for the actual analytical outcomes based on a set of variables, preparing reports that
            are then used by the marketers for marketing strategy development in the near future. Other than the existing data,
            forecasting leans on three pillars:
               • management's experience
               • experts’ knowledge
               • judgement
            In predictive analysis, machines can analyse historical data, find patterns, and foresee the probability of some future
            events. For example, if one owns a chain of restaurants globally, one can easily predict which restaurants are likely to
            get more customers than expected.
            Careful planning is necessary for any kind of business. It is imperative to see the tendencies and react to any changes
            in time. Forecasting is one of the most effective planning approaches. It helps survive market changes and generates
            a suitable strategy.

            Forecasting depends on the data that is collected, usually from both the past and the present, followed by the market
            trends analysis and the development of the description of the actions to follow.

            3.1.1 types of Forecasting

            Two main forecasting approaches are:
               • Qualitative: This method is based on expert or  skillful  opinions and  the  comprehensive analytical research of
              consumers’ behaviour.
               • Quantitative: This method is based on historical statistics research.
            According to marketers, complex forecasting techniques are the most powerful, which means that both qualitative
            and quantitative forecasting techniques should bring actual results so as to build a master plan. Forecasters take the
            following aspects into consideration:

               • expert evidence
               • consumer surveys
               • regression and input-output analysis
               • moving averages, econometric models, etc.
            Now that we have learnt, what are qualitative and quantitative methods of forecasting, let us know more about the
            sub-categories that these two forecasting methods are further divided into, as shown in the diagram below:



                          QUALITATIVE                                           QUANTITATIVE






                  Expert Opinion       Consumer Survey
                     Method                Method
                                                                         Time Series  Moving    Averages  Exponential   Smoothing  Index   Numbers  Regression   Analysis  Econometric   Models  Input Output   Analysis


                          Complete         Sample         End-Use
                      Enumeration Survey   Survey         Method








                                                                                         Forecasting on Data   133
   130   131   132   133   134   135   136   137   138   139   140