Page 288 - Contributed Paper Session (CPS) - Volume 6
P. 288

CPS1932 Carlos A. M. D.



                              Spatial analysis for forced displacement and war
                                           actions: Colombian case
                                            Carlos A Mantilla Duarte
                               Universidad Industrial de Santander, Bucaramanga, Colombia

                  Abstract
                  From  2011  to  2016,  Colombia  Government  and  Colombian  Revolutionary
                  Armed Forces held dialogues in order to end the armed conflict in Colombia.
                  Spatial  relationships  can  be  explained  using  spatial  data  analysis
                  methodologies to compare spatial configuration before and after dialogs. This
                  relationships was analyzed using spatial statistical methodologies like Moran’s
                  I  (Global and Local)  and  Spatial  Regression  comparing likelihood functios
                  from SAR, SEM and SDM models. The SDM Model fitted better in both cass.
                  It’s necessary to analyze missing data to adjust Spatial Panel Data Models or
                  complement with functional data analysis.

                  Keywords
                  Spatial  Analysis;  Spatial  Models,  Spatial  Correlation;  Changes  in  Spatial
                  Relationships

                  1. Introduction
                     Since  2011,  Colombia  government  began  negotiations  with  Colombian
                  Revolutionary Armed Forces (a.k.a. FARC) with the objective to end an armed
                  conflict with more than 50 years of duration. Mantilla (2017) describes spatial
                  relationships  between  some  common  factors  about  armed  conflict  in
                  Colombia. But, the spatial configuration of the variables shows some changes
                  that are not described in the work of Mantilla (2017).
                     This paper research a different approach to describe the changes that the
                  variables show using Global and Local Spatial Correlation Moran’s Indexes
                  and Spatial Regression likelihood function from SAR, SEM and SDM models.
                  The first part contains a summary of methods used in the analysis. After this,
                  the paper exhibit the results of analysis and, finally, exposes a short discussion
                  about the models fitted and contains conclusions,

                  2. Methods
                     The selected variables are based on the work of Mantilla (2015, 2017) and
                  Mantilla & Angulo (2017). The data set corresponds to open data by Office for
                  the Coordination of Humanitarian Affairs United Nations in Colombia (OCHA-
                  UN),  The  Observatory  of  Human  Rights  in  Colombia,  the  Directorate  of
                  Criminal Investigation and INTERPOL of the National Police (Colombia) - DIJIN


                                                                     277 | I S I   W S C   2 0 1 9
   283   284   285   286   287   288   289   290   291   292   293