Computer Science ETDs
Publication Date
Spring 4-14-2021
Abstract
Automated approaches for parameter and algorithm selection greatly democratize fields such as machine learning, saving time and money as hiring experts can be prohibitively expensive. Unfortunately, anomaly detection is difficult to automate due to subjectivity and class imbalance. An anomaly detection system is presented that incorporates human-in-the-loop techniques and is dynamic, scalable, and able to work with non-annotated data. By focusing on meta-features of the input data, the system can intelligently choose the most promising anomaly detection methods. The system is agnostic to the medium of data; it only expects the data to be sequential in nature.
Language
English
Keywords
Anomaly Detection, Time Series, Short Text
Document Type
Dissertation
Degree Name
Computer Science
Level of Degree
Doctoral
Department Name
Department of Computer Science
First Committee Member (Chair)
Abdullah Mueen
Second Committee Member
Trilce Estrada
Third Committee Member
Jedidiah Crandall
Fourth Committee Member
George Luger
Recommended Citation
Freeman, Cynthia. "AutoML for Anomaly Detection of Time Series and Sequences of Short Text." (2021). https://digitalrepository.unm.edu/cs_etds/113
