OUTLINE OF THE PRESENTATION
                                          Generating 
       01Introduction                  Recommendations     04
             The Proposed 
       02      Solution                        Next Steps  05
         Clustering The Customers
       03
                                                                      01
                                                   Problematic
    The floating Customer loyalty 
    The popularity of smartphone                                          The change of the way of how marketing is done
    The growing role of wireless technologies in people’s life            The marketing decision-makers must focus on one-
    The problem of excessive offers and promotions                           to-one marketing and personalized services 
                                           Recommender Systems
                              Discovering the customer behavior patterns 
                             is the key enabler for the success of retailers 
                                                                                                                             02
                         Problematic
     Clustering the customers is usually the first step in the process of analyzing the 
     purchasing behavior in retail. 
     But most of studies construct models based on RFM Model:
     Recency: refers to when the customer did the most recent transaction.
     Frequency: refers to how often customers do transactions.
     Monetary: refers to how much does a customer spend in our store. 
     The Problem: The RFM model captures only partial information of a real user state. 
          Exploit other parameters to extract meaningful 
       knowledge and understand the customer purchasing 
                             habits                           03
                                            Problematic
    The core Recommender System algorithm employs a hybrid approach:
            Apriori Algorithm.
            K-means algorithm.
    To build a recommendation model, three phases are required :
    1.  Information collection phase: Collect the data about the customers, their movements and 
        their purchases. 
    2.  Learning phase: Apply the machine learning algorithm to extract meaningful knowledge from 
        the data obtained in the previous phase. 
    3.  Prediction/recommendation phase: Apply the algorithm to send recommendation or 
        prediction based on the customer’s preferences.                                                    04
                The Proposed Solution
      In order to identify the behavior of customers, an important challenge for 
     retailers is collect all the data about the customers:
      Their personal data 
      Their purchases 
      Their movements inside the mall
     The data collected about the movements and the purchases of the customers 
     will help the marketing decision-makers to achieve better knowledge of 
     purchasing behavior and to apply the adequate marketing strategies.
                                                  05