Big data-based predictive model for attendance rate of reserve forces training

While reserve forces are strategically important to deter war in the Republic of Korea (ROK), the training of the reservists is a challenge since they are civilians and difficult to control. In order to tackle the difficulty of predicting the attendance rate of reserve training, this paper proposes a predictive model using Big Data. The current prediction method in the military uses the last year's attendance rate, and one previous study suggests utilising daily weather information without a systematic analysis. This paper aims to test the significance of the predictor variables in the daily attendance rate. Next, to improve the prediction accuracy of the current method, a predictive model with the volume of web search data is proposed. In the case study, statistically significant predictor variables are identified, and the proposed Big Data-based predictive model improves the prediction performance in comparison to the current method with real reserve training data.