# Difference between revisions of "SC²S Colloquium - October 27, 2017"

(Created page with "{| class="wikitable" |- | '''Date:''' || October 27, 2017 |- | '''Room:''' || 02.07.023 |- | '''Time:''' || 3:00 pm, s.t. |- |} == Jonas Bürger: A Monte Carlo approach for c...") |
|||

(2 intermediate revisions by the same user not shown) | |||

Line 10: | Line 10: | ||

== Jonas Bürger: A Monte Carlo approach for clustering uncertain data using sparse grids density estimation == | == Jonas Bürger: A Monte Carlo approach for clustering uncertain data using sparse grids density estimation == | ||

− | + | In this talk an approach for the clustering of uncertain data is presented. To take the uncertainty of the input data into account, a Monte-Carlo approach is introduced. Therefore the randomized input data is clustered multiple times and representatives for these clusterings are chosen afterwards. The clustering itself is done by using density estimation. For this technique, a cluster is defined as a region of high density. For doing the density estimation, sparse grids are used in order to be able to handle high dimensionality problems. The result of the underlying thesis is a program for clustering questions from the stackexchange network by using the described technique. | |

+ | |||

+ | == Ingo Mayer: Analysis of sparse-grids-based financial time series prediction == | ||

+ | This Bachelor's Thesis applies the delay embedding theorem on financial time series to forecast future values using sparse grids. Sparse grids reduce the impact of the curse of dimensionality and are therefore ideally suited for this regression task. The prediction is applied to the S&P 500 index as a good representative of the global capital market and to a selection of cryptographic currencies to determine the role of volatility in the forecast. For this purpose an automated pipeline was developed in Python, using the general sparse grid toolbox SG++. | ||

[[Category:ShowComingUp]] | [[Category:ShowComingUp]] | ||

[[Category:news]] | [[Category:news]] |

## Latest revision as of 00:03, 27 October 2017

Date: |
October 27, 2017 |

Room: |
02.07.023 |

Time: |
3:00 pm, s.t. |

## Jonas Bürger: A Monte Carlo approach for clustering uncertain data using sparse grids density estimation

In this talk an approach for the clustering of uncertain data is presented. To take the uncertainty of the input data into account, a Monte-Carlo approach is introduced. Therefore the randomized input data is clustered multiple times and representatives for these clusterings are chosen afterwards. The clustering itself is done by using density estimation. For this technique, a cluster is defined as a region of high density. For doing the density estimation, sparse grids are used in order to be able to handle high dimensionality problems. The result of the underlying thesis is a program for clustering questions from the stackexchange network by using the described technique.

## Ingo Mayer: Analysis of sparse-grids-based financial time series prediction

This Bachelor's Thesis applies the delay embedding theorem on financial time series to forecast future values using sparse grids. Sparse grids reduce the impact of the curse of dimensionality and are therefore ideally suited for this regression task. The prediction is applied to the S&P 500 index as a good representative of the global capital market and to a selection of cryptographic currencies to determine the role of volatility in the forecast. For this purpose an automated pipeline was developed in Python, using the general sparse grid toolbox SG++.