Go to JKU Homepage
What's that?

Institutes, schools, other departments, and programs create their own web content and menus.

To help you better navigate the site, see here where you are at the moment.

Uncertain Data and Query Processing

Supervisor: a.Univ.-Prof. DI Dr. Josef Küng
Co-Supervisor: Dr. Maksim Goman

Introduction, Motivation

Currently, complexity of business processes grows, streams of data from sensors and interchange with third-party software accelerate and increase, big data analysis became
a routine activity, in-memory databases became normal practice. Complex business analysis is going to be performed real-time on every minor data update and control dashboards
are constantly updated. Uncertainty is usually present in many kinds of data, mainly in forecasts and estimations, imprecise measurements and ambiguous observations where
uncertainty may not be excluded. In relation to massive data streams, risk may originate directly from uncertainty in the raw data. Beyond risk analysis, applications of uncertain
data processing can be real-time scheduling of atomic operations (e.g. in manufacturing), quality control, monitoring of user satisfaction, worldwide fraud management in financial
transactions, etc. For instance, let us imagine a logistic supply chain in which products loose their value according to known distribution. For instance, a cold supply chain.
Certain parameters can influence the actual condition of each item of products of many types and this is accounted for in the distributions. Some of the parameters of all or some
items are regularly updated, although this information may have random error too. Real condition of each unit is unknown until it is sold or disposed. Nevertheless, possible loss
due to natural deterioration and need for additional delivery to certain hubs or regions can be computed probabilistically. Value of uncertain aggregation with some probabilistic
threshold can be used in a decision rule in a decision support or inventory management system as well as in certain cost/benefit or sales trend prognoses in ERP constantly.

Bachelor or Master Theses

Depending on the study background of a student, following Sub-Topics can be chosen in the frame of a Bachelor Thesis or Master Thesis:

1. Review state-of-the-art methods for uncertain data processing, existing uncertain query models and semantics
2. Syntax of an uncertain query similar to SQL in relational databases
3. Uncertain comparison operators for uncertain data analysis (<,>,=)
4. New model of an uncertain query using chance constraint method and SQL-like syntax
5. Principles and syntax for uncertain conditioning operators like GROUP BY or WHERE a>b and their implementation
6. Principles and syntax for uncertain aggregation operators (MIN, MAX, AVG, etc.) and their implementation

A student needs to review the problems and current methods of the chosen topic of interest.
Define a sample data set, build the model, and develop the artefact of the solution.
Demonstrate application of the developed solution to the data.
Discuss and summarize the results paying attention to comparison with alternative techniques.

Instructions and literature will be given by the supervisor

Successful thesis with a topic in our area is a ticket to a team of unique professionals who develop modern data processing and querying algorithms at companies and research institutions.