Page 290 - Proceeding The 2nd International Seminar of Science and Technology : Accelerating Sustainable Innovation Towards Society 5.0
P. 290

he 2  International Seminar of Science and Technology
                                    nd
                                   “Accelerating Sustainable innovation towards Society 5.0”
                                                       ISST 2022 FST UT 2022
                                                          Universitas Terbuka

          On the other hand, objects with different characteristics are said to
          have  dissimilarities  so  that  the  difference  between  the  two  can  be
          measured  in  the  perception  distance  which  is  translated  into  a
          perception  index  such  as  a  sustainability  index.  The  distance
          determination technique is based on Euclidian Distance with Equation
          (4), where d1,2   =  Euclidian  distance,  X,  Y,  Z  =  Attribute,  1,2  =
          Observation.
                                      2
                                                  2
                            2
           d  2 , 1  =  (X 1  − X 2 ) + Y 1  2 ) + (Z 1  − Z 2 ) + ...      (4)
                               ( −Y
               The euclidian distance between the two points (d1,2) then in the
          MDS,  is  projected  into  a  two-dimensional  euclidian  distance  (Ḋ1,2)
          based on the regression formula in Equation (5), where a = intercept,
          b = slope, c = error.
           D1,2 = a + b D1,2 + c                                   (5)
          In MDS, two points or the same object are mapped in a single point
          that  is  close  to  each  other.  The  technique  used  is  the  ALSCAL
          algorithm and is easily available in almost every statistical software
          (SPSS and SAS). Rap-gametRescue in principle makes iterations of
          the regression process so that the smallest e value is obtained and
          tries to force the intercept in the equation to be equal to 0 (a=0). The
          iteration stops if the stress is < 0.25 [3]. The stress value is shown in
          Table. 3. For attributes as much as m, stress can be formulated in
          Equation (6).
                               2    2  ) 2                       (6)
                   1  m      (D ijk  − d ijk  
                           i
           stress  =      j         
                      = 
                   m  k 1        d 2    
                            j  ijk   
                              i

                              Table 3. Stress Value
                     No.  Stress value      Suitability
                     1    > 20 %            Poor
                     2    10 - 20 %         Moderate
                     3    5 - 10 %          Good
                     4    2,5 - 5 %         Very good
                   Source: (Kavanagh P. & Pitcher, [15]

          260                          ISST 2022 – FST Universitas Terbuka, Indonesia
                    International Seminar of Science and Technology “Accelerating Sustainable
                                                         Towards Society 5.0
   285   286   287   288   289   290   291   292   293   294   295