Page 99 - Kỷ yếu hội thảo khoa học lần thứ 12 - Công nghệ thông tin và Ứng dụng trong các lĩnh vực (CITA 2023)
P. 99

Thanh Trung Nguyen, Hai Bang Truong                                              83




























                     Fig. 2. Representation of the same data in aggregated form. The arithmetic mean is represented
                     by a dotted line and the geometric mean by a dashed one [10]

                     Several  studies  involved  conducting  experiments  to  compare  various  aggregation
                     methods, such as arithmetic and geometric means, and median, in different contexts
                     of  setting  with  asymmetric  information.  The  study  concluded  that  although  the
                     geometric average performed reasonably well, it was not the most effective approach
                     [14].
                       In the context of studies on the role of social influence on the wisdom of crowds,
                     which is a main direction of research, the geometric mean is often used to analyze the
                     accuracy of group decisions. In their research [19], the authors reported that in cases
                     where the solution space is wide and estimates will exhibit high variance and a wide
                     range  of  positive  values,  the geometric  mean is  a  more  accurate  measure  of  the
                     wisdom of the crowd, as it captures the central tendency of the population better than
                     the arithmetic mean.


                     4.3   The Median

                     Several studies have indeed examined the use of the median as a measure of central
                     tendency in different contexts related to crowd predictions for comparing with other
                     aggregation methods. Becker [9] used a prediction market to study the accuracy of
                     crowd  predictions  for  political  events  and  found  that  the  median  forecast
                     outperformed  other  measures  of  central  tendency  such  as  the  mean  or  mode.
                     Similarly, Hueffer [20] used a prediction market to forecast the spread of a disease
                     and  found  that  the  median  forecast  was  a  better  predictor  of  the  actual  outcome
                     compared to other measures. Palan et al. [14] examined the use of the median as a
                     measure  of  central  tendency  in  survey  responses  and  found  that  it  improved  the
                     accuracy of the aggregated responses.








                     ISBN: 978-604-80-8083-9                                                  CITA 2023
   94   95   96   97   98   99   100   101   102   103   104