Big Data Forecasting Applied Nearest Neighbor Method

Abstract

Nearest neighbor is a method of approaching the case by calculating the proximity between the new and the old case in order to make the forecasting process. The value of the approach used in Nearest Neighbor itself is a classification approach that searches all training data that is relatively similar to the data test. Each approach calculates all new and old data so that in the process of searching the value of proximity needs testing of Big data in order to find how much time to produce in forecasting based on old cases. This study had been experimented by using 2,500 cases with 1 minute and 52 seconds, 18,000 of which required longer time i.e. 16 minutes and 30 seconds. The differences in the number of the old cases influence the process of time calculation.