|Title||Development and evaluation of an artificial immune system for anomaly detection|
The detection of anomalies is a complex task, which is investigated in science in many different research areas (besides computer science e.g. also in statistics or engineering). While machine learning models are state of the art today, there are other models for anomaly detection, such as artificial immune systems.
Artificial immune systems are algorithms inspired by the biological immune system that use the properties to discriminate between self (endogenous) and non-self (exogenous) in the immune system to detect anomalies in data. The majority of known algorithms are based on the principle of negative selection. A well-known algorithm for this is, for example, the Real Valued Negative Selection Algorithm. There are also other models, such as the Multilevel Immune Learning Algorithm.
|Home institution||Department of Computing Science|
|Type of work||practical / application-focused|
|Type of thesis||Bachelor's or Master's degree|
|Author||Torge Wolff, M. Sc.|
For the application of machine learning models, such as artificial neural networks, there are a number of well-known and widely used frameworks that simplify the training, evaluation and deployment of the models. For Artificial Immune Systems, however, such frameworks do not exist as of today.
The task of this thesis is to implement an algorithm of an artificial immune system and to evaluate it on a well-known benchmark dataset in comparison with state-of-the-art machine learning models. The exact orientation as well as the focus of the work can be defined flexibly in advance.
Basics Machine Learning