Dengue Predictive Model 2020 for Rawalpindi District Pakistan

Abstract

Objectives: To  draw  dengue  predictive  model  2020  for Rawalpindi district in response to dengue epidemic 2019 for timely arrest and mitigation of dengue cases. Methods: Predictive  model  was  drawn  by  using  machine learning  technique.  Numerous  tools  like Pandas,  Numpy, Matplotlib,  Sklearn,  Pylab  were  use  for  data  wrangling. Residential  data  of  12,192  dengue  cases  admitted  in  3 teaching  hospitals  (Holy  Family  Hospital,  Benazir  Bhutto Hospital and District Head Quarters Hospital) affiliated with Rawalpindi   Medical   University   was   employed   for   this purpose.  Weather  data  for  the  Rawalpindi  was  extracted from  Texas  A&M  university  weather  data  archive.  Ordinary Linear Sequential (OLS) Regression was used to estimate the relationship  of  dengue  cases  reported  during  2019  with weekly average temperature and cases to be reported during 2020 in Rawalpindi district. R2 was computed as an indicator of  model  fit.  Significance  of  association  between  reported cases  and  air  temperature  was  statistically  confirmed  by application of t-test. Results:Fitness  of  predicted  model  was  determined  in terms   of   R2and   P-value   by   application   of   t-test   of correlation.  R2of  our  predicted  model  is  0.79,  presenting positive   relationship   between   active   dengue   cases   and average   weekly   air   temperature.   Moreover,   statistically significant    association    of    reported    cases    with    air temperature   (P   <   0.086)   was   determined   with   97.5% Confidence Interval (CI). Conclusion: Dengue   predictive   model   2020   drawn   for Rawalpindi district would really be helpful to reduce dengue cases  by  application  of  appropriate  preventive  measures  in high risk zones by concerned authorities.