Predicted Missing Imputation on Dengue Fever Spread Data with K-Nearest Neighbor (K-NN)

Abstract

Dengue Hemorrhagic Fever (DHF) is a disease caused by dengue virus with Aedes Aegypti intermediate.  Based on  a  survey  from  the  Health  Office  of  Jember  recorded  during January 2015 out of 300 cases of DHF patients, 7 of them are died, that why the prediction of DHF distribution is needed for prevention of spreading. Parameters that used to determine the potential  for  the  spread  of  DHF  diseases  are  rainfall  ,  rainy  day,  larva  free  and  house  index. However, the survey data is often incomplete, the missing imputation data resulted the process to  predict  the  potential  for  the  spread  of  DHF  is  still  constrained.  By used  of  K-Nearest Neighbor  (K-NN)  methods  that  can  be  used  to  predict  the  missing  imputation  data  and complete  it.  Using  the  correlations  between  attributes  attained  on Euclidean  Distance  that shows  better  performance  in  terms  of  imputation  accuracy. The method  show  MSE  below  1 and MAPE around 10 – 16%.