In this project, we will evaluate methods to imputation and utilize them on data gathered by Equal Employment Opportunity Commission. In the beginning, I discuss a number of imputation methods and analyze concept of multiple imputation (MI). Next, I evaluate difficulties with missing data and describe an artificial data simulation.
This project provides a group of new algorithms designed to deliver globally optimal solutions for Gaussian mixture models. The Expectation-Maximization algorithm is definitely a popular and straightforward way for the estimation of Gaussian mixture models along with its natural extension, model-based clustering. Having said that, although the algorithm is not hard to use and numerically it is very stable but it provides solutions which are locally optimal.
The heat transfer in the filling phase of injection moulding is studied, based on Gunnar Aronsson’s distance model for flow expansion. The choice of a thermoplastic materials model is motivated by general physical properties, admitting temperature and pressure dependence.
This project is on Stochastic Diffusion Problem. The primary goal of this project was to design iterative methods for solving the linear systems which usually appear from using the stochastic finite element approach to steady-state stochastic diffusion problems.
This project is about modeling and optimization of transmission networks. These types of networks carry out a vital task in communication, energy transmission, micro-electronics, etc. The intention of this project is to take into account a couple of different problems in connection with transmission networks.
In this project report, we will talk about Forest supply chains its planning levels and optimization. This project is about modelling and solving large-scale planning problems in the supply chain in the forest industry. We have included five research papers. The first
The objective of this section is to offer some background material and inspire, in addition to in short describe, the content of the first 2 papers of this thesis. At the very hart of Hilbert space theory and its apps lies the idea of an orthonormal sequence, i.e. a series of pairwise orthogonal elements of norm one from the underlying space.
In this dissertation, we research classification algorithms created by regularization schemes. The design of these algorithms and their error analysis are completely explained. These algorithms rely on convex risk minimization with Tikhonov regularization. They require an admissible convex loss function, a hypothesis
Efficient numerical methods are created to examine optical bistability in a multilayer structure with Kerr nonlinearity. It corresponds to computing multiple solutions of the one-dimensional nonlinear Helmholtz equation for an array of frequencies and incident intensities. The input-output characteristics is calculated as
This dissertation concerns a particular basis for the coordinate ring of the character variety of a surface. Let G be a connected reductive linear algebraic group, and let S be a surface whose fundamental group pi is a free group. Then the