Student Projects
Topic: Online Centroid Decomposition for Dynamic Time Series
The aim of this thesis to implement an incremental version of the Centroid Decomposition method
Status
- Open
Type:
- BSc/MSc
Student:
- TBD
Supervisors:
Topic: Centroid Decomposition Based Recovery for Segmented Time Series
The aim of this thesis is to investigate and implement the Centroid Decomposition method to recover blocks of missing values in time series. The special focus will be to compare the accuracy of the recovery technique using complete hydrological time series and segmented hydrological time series. The segmentation of the time series will be an issue to investigate.
Status
- completed
Type:
- BSc
Student
- Jonathan Nagel
Supervisors:
Topic: Missing Block Recovery using Centroid Decomposition Method
The aim of this Master project is to investigate and implement the Centroid Decomposition method to recover blocks of missing values in shifted time series. As an example for the block recovery in shifted time series, consider the humidity, precipitation and temperature time series shown in
Status
- completed
Type:
- MSc projektarbeit
Student:
- Eszter Börzsönyi
Supervisors:
Topic: SQL Implementation of Centroid Decomposition
The aim of this Vertiefung is to investigate and implement the Centroid Decomposition method using SQL queries. The task is to implement using PL/SQL the Centroid Decomposition and evaluate the solution empirically on hydrological time series. The goal of the SQL implementation is to optimize the efficiency of the decomposition computation by reducing the memory allocation and preserving the accuracy for large data sets. The solution should scale up to large number of time series with hundreds of millions of observations.
Status
- completed
Type:
- Informatik-Vertiefung
Student:
- Alessandro De Carli
Supervisors:
Topic: SQL Implementation of QR Formalization and Householder Transformation
The aim of this Vertiefung is to investigate and implement the Singular Value Decomposition (SVD) method using SQL queries. SVD is a matrix decomposition method that decomposes a matrix V into three matrices L, Σ and RT. The product of the three matrices is equal to V. The SVD method is performed by using the Householder transformation and the QR factorization.
Status
- completed
Type:
- Informatik-Vertiefung
Student:
- Robert Stucki
Supervisors:
Topic: Implementation of Approximation Functions for Time Series
Topic: The aim of this summer project is to investigate and implement different types of functions able to approximate time series. The two main types of functions are the continuous approximati- on functions and the piecewise approximation functions functions. An empirical comparison between the two types of functions should be provided as a final output of the summer project.
Status
- completed
Type:
- Summer project
Student:
- Rojomon Nagaroor
Supervisors:
Topic: Engine Tool for Recovering Missing Values
The goal of this project is to develop a new online search tool that efficiently supports the recovery and the visualization of missing values in databases that store hydro- logical measurements. The recovery of missing values is based on using some replacement techniques for the approximated missing values. In order to assist the development of single missing values imputation strategies, basic techniques with their corresponding visualizations have to be implemented.
Status
- completed
Type:
- MSc projektarbeit
Student:
- Michal Koltonik