Predicting the Growth of Total Number of Users, Devices and Epidemics of Malware in Internet Based on Analysis of Statistics with the Detection of Near-Periodic Growth Features

摘要

This paper attempts to predict new epidemics of computer threats based on statistical analysis and the detection of likeness through near-periodic functions. As main data, the authors use data on the growing number of Internet users, the increasing number of devices connected to the global network and data on computer viruses activity statistics over the years (detected epidemics of computer viruses activity, such as computer worm attacks, Trojan viruses, and other activities of this kind). Analysis of the data shows that the increasing of users number and various devices is well described by the model of limited growth of Gompertz with logistical dependence. Data analysis on computer viruses' epidemics shows the presence of trend and oscillatory components. To obtain accurate predictions of anticipated future epidemics of computer viruses, it is necessary to find a method which helps to divide the trend component and the oscillatory component without significantly losing information about the observed process. The work attempts to build such almost periodic functions and to compare simulated data with real-life data. Based on the processed data, the authors estimate the limit of the current growth in the number of Internet users in 5.4 billion users (while maintaining the current level of technology), as well as highlight the main almost-period lasting 7-9 years and the main almost-period at 16 years. The maximum appearance of next-generation malware can occur in 2020, which shows the existence of mechanisms for their functioning already at present time.).