Registered Data

Contents

[CT074]

Session Date & Time : 3D (Aug.23, 15:30-17:10)
Type : Contributed Talk
Abstract : This work establishes the central limit theorem and moderate deviation principle for stochastic shell model of turbulence driven by multiplicative noise. The method of weak convergence introduced by Budhiraja and Dupuis has been followed in order to establish our results. The equivalence of Laplace principle and large deviation principle under Polish spaces contributes to reduce the complexity.
Classification : 60F05, 60H15
Author(s) :
- Sridevi C.S. (Bharathiar University, Coimbatore, Tamil nadu)

Session Date & Time : 3D (Aug.23, 15:30-17:10)
Type : Contributed Talk
Abstract : Real-world dynamical systems can be susceptible to events with a low probability of occurrence but severe repercussions. The aim is to asymptotically quantify the likelihood of these events in dynamical systems represented by stochastic differential equations (SDEs). First, we will go through the mathematical framework for investigating such situations. Then, we will demonstrate a numerical obstacle when using a rare events method due to diverging the tilting factor, aka Radon-Nikodym derivative. A solution will be proposed and shown using multiple scenarios.
Classification : 60F10, 65C20, 49M05, 49M29, 65C05
Author(s) :
- Mnerh Alqahtani (University of Hafr Al Batin)
- Tobias Grafke (University of Warwick)

Session Date & Time : 3D (Aug.23, 15:30-17:10)
Type : Industrial Contributed Talk
Abstract : Estimating the number of significant components(factors, resp.) from principal component analysis(explanatory factor analysis, resp.) in datasets of finance/biology is essential. However, statistical software's default estimation method behaves pathologically for big datasets. We analyze the phase transition of the default rule as to the intra-class correlation of various data-generation models, and introduce a more acceptable estimation by random matrix theory for large sample correlation matrices. We also compare our rule to retention rules proposed to date.
Classification : 60F15, 62H25
Author(s) :
- Atina Husnaqilati(Department Mathematics Tohoku University)
- Yohji Akama (Tohoku University)

Session Date & Time : 3D (Aug.23, 15:30-17:10)
Type : Contributed Talk
Abstract : Classifying samples in incomplete datasets is a common non-trivial task. Missing data is commonly observed in real-world datasets. Missing values are typically imputed, followed by classification of the now complete samples. Often, the focus is to optimise the downstream classification performance. In this talk, we highlight the serious consequences of using poorly imputed data, demonstrate how the common quality measures for measuring imputation quality are flawed and introduce an improved class of imputation quality measures.
Classification : 62D10, 68T07, 68T01, Assessing the imputation of missing data
Author(s) :
- Michael Roberts (University of Cambridge)

Session Date & Time : 3D (Aug.23, 15:30-17:10)
Type : Contributed Talk
Abstract : We studied stability properties of the expected utility function in Bayesian optimal experimental design. We proved a convergence rate of the expected utility with respect to a likelihood perturbation. This rate is uniform over the design space. As an example we have non-linear Bayesian inverse problems with Gaussian likelihood satisfying general assumptions. Theoretical convergence rates are demonstrated numerically in three different examples.
Classification : 62K05, 62F15, 35R30, Bayesian Inverse Problems
Author(s) :
- Tapio Helin (Lappeenranta University of Technology)
- Duc-Lam Duong (Lappeenranta University of Technology)
- Jose Rodrigo Rojo Garcia (Lappeenranta University of Technology)