Data Science - Big Data and Machine Learning

  • Statistical network monitoring
    The application of network analysis has found great success in a wide variety of disciplines; however, the popularity of these approaches has revealed the difficulty in handling networks whose complexity scales rapidly. One of the main interests in network analysis is the online detection of anomalous behaviour. To overcome the curse of dimensionality, we compose the monitoring procedures which reduce the network complexity so that the structural information is preserved. The methods are mainly based on statistical process control which are optimised with different mathematical network modelling and machine learning techniques.
    Team: Malinovskaya, Otto
    Year: 2020
  • Detection of structural breaks in stochastic processes with spatial dependencies
    With growing availability of high-resolution spatial data, like high-definition images, 3d point clouds of LIDAR scanners, or communication and sensor networks, it might become challenging to timely detect changes and simultaneously account for spatial interactions. To detect local changes in the mean of isotropic spatiotemporal processes with a locally constraint dependence structure, we propose a monitoring procedure, which can completely be run on parallel processors. This allows for a fast detection of local changes, i.e., only a few spatial locations are affected by the change. Due to parallel computation, high-frequency data could also be monitored. We, therefore, additionally focus on the processing time required to compute the control statistics.
    Leaders: Prof. Dr. Philipp Otto
    Year: 2019
  • Estimation of spatial dependency structures
    In spatial econometrics, the classical approach would be to replace the unknown spatial dependence structure with a linear combination of an unknown scalar, which has to be estimated, and a pre-defined matrix of spatial weights. This non-stochastic weighting matrix describes the dependence. One might obtain insight into its structure by examining the spatial covariogram or semivariogram. In practice, however, the true underlying matrix cannot easily be assessed, and therefore has to be estimated by maximizing certain goodness-of-fit measures, such as the log-likelihood, in-sample fits, information criteria, or cross validations over certain classical weighting schemes. In contrast to this classical approach, the project aims to develop methods to estimate the entire spatial weighting matrix. Moreover, the procedures should account for endogenous effects.
    Leaders: Prof. Dr. Philipp Otto
    Year: 2019
  • Spatial and spatio-temporal GARCH models
    The project is concerned with a subfield of spatial statistics, which deals in particular with the analysis of random processes in space. When analyzing such processes, it can often be found that observations that are located in spatial proximity to each other are similar. If, for example, land prices in a municipality are high, high prices can also be expected in the surrounding municipalities. In addition to this spatial dependency in the level of observations, a spatial dependency in the dispersion of observations as well as the conditional heteroskedasticity can also be determined. In the project, models for this will be developed and extended. The spatial models are analogous to the ARCH model of Robert F. Engle (1982) in time series analysis, who was awarded the Nobel Prize for Economics in 2003.
    Leaders: Prof. Dr. Philipp Otto
    Year: 2019
    Sponsors: Deutsche Forschungsgemeinschaft
  • TransMIT - Resource-optimised transformation of combined and separate drainage systems in existing quarters with high settlement pressure
    Increasing heavy rainfall events and growing urban districts pose great challenges for urban drainage. Using three neighbourhoods in the cities of Braunschweig, Hanover and Hildesheim as examples, it will be shown how urban development and water management aspects can be linked in the long term in neighbourhood planning.
    Leaders: Dr.-Ing. M. Beier; Prof. S. Köster, ISAH; Prof. Sester, ikg
    Team: Yu Feng, Udo Feuerhake
    Year: 2019
  • Object detection in airborne laser scanning (ALS) data using deep learning
    In partnership with the Lower Saxony State Office for Preservation of Historic Monuments, we are developing a method for automatically detecting archaeological objects in airborne laser scanning data. The type of objects to be detected are mainly those of interest by archaeologists such as heaps, shafts, charcoal piles, pits, barrows, bomb craters, hollow ways, etc. They could be point, linear, or areal objects. To this end, we are using deep learning techniques; namely, convolutional neural networks (CNNs) to classify height images from the region of interest. A combination of multiple (in most cases 5) CNN classifiers are then used to detect and localize objects of interest in a digital terrain model acquired from the region of interest.
    Team: Kazimi, Thiemann, Sester
    Year: 2018
    Sponsors: MWK Pro*Niedersachsen
  • Interdisciplinary Center for Applied Machine Learning - ICAML
    The ICAML (Interdisciplinary Center for Applied Machine Learning) aims at increasing the accessibility of machine learning across disciplines. Therefore, three fundamental components are developed and used.
    Team: Artem Leichter
    Year: 2018
    Sponsors: Federal Ministry of Education and Research
    Lifespan: 11/2017-11/2019
  • USEfUL
    Due to its location at the center of Europe and the global operating companies, logistics and mobility have always been of outstanding importance in Hanover, a city rebuilt car-friendly after the war. A growing city is associated with increasing mobility and supply needs as well as an individually and systemically caused need of logistics for supply and disposal.
    Team: Wage, Feuerhake
    Year: 2018
    Sponsors: BMBF: 03SF0547
  • Ja, wo laufen sie denn?
    Für Profi-Trainer oder auch einfache Hobby-Kicker. Vielen Fußballbegeisterten wird der Weg zum Taktikfuchs durch eine automatisierte Spielanalyse am Computer erleichtert. Ausgeklügelte Verfahren ermöglichen eine einfachere Bewertung der Leistung der Akteure.
    Team: Feuerhake, Sester
    Year: 2017
  • 3D object extraction of high-resolution 3D point clouds
    National Survey Departments acquire area-wide, controlled airborne laser scanning (ALS) datasets with different point densities, which are at least classified into ground and non-ground points. The Working Committee of the Surveying Authorities of the Laender of the Federal Republic of Germany (AdV) is discussing about an update cycle of 10 years for ALS point clouds. The national survey departments also acquire 3D point clouds from aerial images every 2-3 years with high overlapping ratios using a method called Dense Image Matching (DIM). Those DIM point clouds have a high point density, which is equal to the original aerial image resolution. In addition, those DIM point clouds also contain radiometric information from the aerial images, but only reconstruct the surface due to image correlation. This project is split into four distinctive topics.
    Team: Politz, Sester
    Year: 2017
    Sponsors: Forschungs- und Entwicklungsvorhaben zwischen den Landesvermessungsämtern Niedersachsen, Schleswig-Holstein und Mecklenburg-Vorpommern
  • Scene analysis - pattern recognition in person tracks
    The project deals with the automatic recognition of conspicuous movement patterns from given trajectories as (x, y, t) sequences of objects. For this purpose, patterns are defined that indicate conspicuous behavior for which automatic extraction methods are to be researched. This includes, among others, the qualitative description of the recognition rates of patterns.
    Team: Fischer, Sester
    Year: 2017
  • Real Time Prediction of Pluvial Floods and Induced Water Contamination in Urban Areas (EVUS)
    This project aims at developing a fast forecast model for pluvial floods in the city of Hannover. The main goal of the subproject for ikg is to integrate new sensors for the flood prediction models.
    Team: Feng, Sester
    Year: 2017
    Sponsors: BMBF Georisiken