Page 119 - Artificial Intellegence_v2.0_Class_12
P. 119

A successful problem defining process requires a basic analysis and evaluation of the pro ect related problems, their
            reasons and methods.  inding the right problem definition is usually an iterative process. It can reveal more questions
            and points to consider that  ould have been ignored  ithout the problem definition process.  he questions belo
            serve as a reference point for a thorough analysis of the problem and the problems that surround it.  ust spending time
            ans ering the follo ing questions can save you  eeks and months  orking on problems that proved impossible for
            previously unkno n reasons
            •    hat is the problem that needs to be resolved

            •    hy do you need a solution to your problem
            •    o  should  e  ork on the solution to the problem
            •    hich aspects of the problem do the AI model solve
            •    o  do I need to interact  ith the solution to the problem

            •    hich category of data  ill be involved    lassification)
            •    o  much or ho  many    egression)
            •    an the data be grouped    lustering)
            •   Is there any unusual pattern in the data   Anomaly Detection)
            •    hich option should be given to the customer    ecommendation)

            It is essential to determine  hich of these questions you re asking, and in  hat  ay ans ering them helps solve your
            problem.



                          Brainy Fact


                  Organisations need analytics to give their businesses a competitive advantage.  here are five types of
                  business analytics.  et us learn about them brie y.

                                                                                    Predictive (Insight)
                                                     Diagnostic (Foresight)      sing the previously obtained
                       Descriptive (Hindsight)      Analytics that focuses on     data, predictive analytics
                      It helps in summari ing or    previous performance to     provides accurate predictions
                                                                                  of  hat is going to occur.
                   describing past events.   amples   determine the cause of an event.    achine learning techniques
                       generating financial or sales     ample    tudy the impact of
                                                                                  and statistical models are
                             reports.            medicine on patients before and   used to forecast the potential
                                                        after treatment
                                                                                outcome.  ample   Ama on's
                                                                                  recommendation engine.

                                                                        Cognitive (Inference)
                                   Prescriptive (Context)            It combines a number of smart
                              his analytics is a subset of predictive   technologies, including artificial
                              analytics that uses data analysis to   intelligence, machine learning algorithms,
                            suggest one or more courses of action.    and deep learning models, to simulate the
                              ample   aymo,  oogle s self driving   human brain and provide outcomes that
                                car uses prescreptive analytics   are compatible  ith ho  people think.
                                                                    amples   chatbots, virtual assistants










                                                                                       C apstone  P roj e ct
   114   115   116   117   118   119   120   121   122   123   124