LLEAD – Layered Learning for Early Anomaly Detection: A Metalearning Approach

Description:

This project aims to understand the feasibility of applying a Layered Learning approach to different anomaly detection and forecasting problems, from public datasets containing time series data. With features extracted from such data with TSFEL, a metalearning study will be developed to identify the best approaches in the different layers of the complex problem, either between existing forecasting algorithms or with the combination of different feature types.

 

Outcome:

The thesis is focused on the development of AICOS skills under the topic of time series forecasting. It will study the potential of layered learning techniques for early anomaly detection problems with time series data.

 

Author: Bruno Ribeiro

Type: MSc thesis

Partner: INESC TEC – Instituto de Engenharia de Sistemas e Computadores, Tecnologia e Ciência

Year: 2021