Descriptive Statistics of Malware Data

Abstract

Exploring and analysing a dataset provides insights into what kind of data is present and how it can be used. This is especially useful for malware datasets. As an area that is growing bigger due to the implementation of machine learning techniques, having knowledge about a dataset may assist in any future machine learning task that can be done on the dataset. This work aims to gain statistical insights about a dataset of malware and to explore patterns of different families of malware. This will provide a gateway to enable categorizing malicious files based on their properties. One of the outcomes of this work is the discovery of patterns and insights as to how different attributes of a malware specimen can act as an indicator of its maliciousness.