188/4 E-Commerce Group
Institute of Software Technology and Interactive Systems
Vienna University of Technology
Favoritenstrasse 9-11/188, A-1040 Vienna, Austria

Data cleaning, consolidation, and feature selection for failure prediction of Edge computing systems

Type: 
Master Thesis
Description: 
Bayesian networks can effectively model the dependencies between different types of failures and can accurately estimate availability levels in Edge computing systems. However, there are huge number of attributes that can be included to the model, such as HW and SW characteristics, previous failures rates, geographical and logical location, etc. The task is to identify most useful attributes in order to increase accuracy and decrease complexity of the model. Preprocessing data by filling missing values, combining multiple attributes into a generated feature, or selecting features with highest information gain can be considered in that sense.
State: 
published
Supervisor: 
Brandic, I.
Language: 
English
Issued: 
Apr 2017