Surveying Different Aspects of Anomaly Detection and Its Applications


Neda Noori - Islamic Azad University, Zanjan branch, Zanjan, Iran Leila Boti - Islamic Azad University, Zanjan branch, Zanjan, Iran Ebrahim Nowzarpoor Shami - Islamic Azad University, Zanjan branch, Zanjan, Iran


Detecting anomalies is a significant issue which is being investigated in different levels of research and application. Many techniques of anomaly detection are widely and specifically developed for special applied domains while other techniques are mostly general. The purpose of this article is to provide a concise and comprehensive summary of surveys and researches related to anomaly detection. We have categorized the available techniques based on the certified methods; and to distinguish between normal and abnormal behaviors, we have defined some key hypotheses which are used by techniques. When a given technique turns out to be efficient for a specific domain, its hypotheses can provide strategies for accessing technique-efficiency in that domain. For each category, we have provided a basic technique for anomaly detection, and then we have shown how different basic techniques which are available in any category are distinguished from the defined basic technique. This procedure gives us a brief and simple understanding of the techniques which belong to any category. In addition, we will define the pros and cons of the techniques of each category. We hope that this article provides a better understanding of different aspects of the investigated issues, and that how the techniques which are developed in one area can turn out to be efficient in other areas which have not been part of the presuppositions at first.