Malware attack forecasting by using exponential smoothing

Abstract

One of the threats on digital environment is malicious software (malware). Malware can bring harm to computer system that have connection to the internet. Malware may disclose sensitive information such as password and brings economic losses. Predicting the malware attack is vital in supporting decision-making process to avoid further damage on computer systems. The main objective of this study is to develop computational model to predict quantity of malware attack and assess the performance of exponential smoothing in forecasting malware attack. There are two types of exponential smoothing forecasting model involve in this study which is single exponential smoothing and double exponential smoothing. The forecasting performance will be evaluated by using Mean Squared Error (MSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). Monthly malware detection data in one financial institution’s computer servers from November 2014 to August 2019 will be utilized in this research. The result from this study shows that single exponential smoothing produces 0.0015 of MSE, 0.4655 of MAE and 6.0158 of MAPE. Single Exponential Smoothing produces lower value of MSE, MAE and MAPE, compared to double exponential smoothing. Thus, single exponential smoothing gives a promising result in forecasting the malware attack.