Title: Support Vector Regression desde una perspectiva multiobjetivo: Una aplicación a la esperanza de vida |
Authors: Ivanova, Daria |
Tutor: Ortiz Henarejos, Lidia Cánovas Cánovas, María Josefa |
Editor: Universidad Miguel Hernández de Elche |
Department: Departamentos de la UMH::Estadística, Matemáticas e Informática |
Issue Date: 2020-06 |
URI: http://hdl.handle.net/11000/8580 |
Abstract:
En el presente proyecto se lleva a cabo una breve recopilación de métodos bien conocidos basados en el aprendizaje automático Support Vector Machine (SVM) y Support Vector Regression (SVR) para la resolución de problemas de clasificación y regresión, respectivamente.
El primer método, a grandes ras... Ver más
In the present project a brief compilation of well known methods based on machine learning is brought off. These methods are Support Vector Machine (SVM) and Support Vector Regression (SVR), for the resolution of classification and regression problems, respectively.
The first method, broadly speaking, consists of the search of hyperplanes capable of separating the different classes. The second method also lies in searching a regressor hyperplane that describes the data as accurately as possible.
The SVR is the successor of the SVM, that is why there are less related studies or literature. Nevertheless, it has a wide application because the response variable is continuous. In this project we will focus on SVR technique.
Considering that the SVR turns out to be a Multi-objective Programming (MOP) problem, an alternative approach and resolution to the existing one is proposed.
Once the different optimization models are defined, R and MATLAB programming languages are used to program the pertintent algorithms that allow these problems to be solved. This implementation,
which solves different SVR models in a generic way, constitutes one of the main original contributions of the work.
To illustrate the different optimization models proposed, the algorithms will be applied over a real database, which will allow us to make predictions, using the above-mentioned technique, about the
life expectancy in Organisation for Economic Co-operation and Development (OECD) countries
|
Keywords/Subjects: máquinas de soporte vectorial support vector machine SVM regresión de soporte vectorial SVR programación lineal programación cuadrática support vector regression linear programming quadratic programming multi-objective programming |
Knowledge area: CDU: Ciencias sociales: Demografía. Sociología. Estadística CDU: Ciencias aplicadas: Gestión y organización. Administración y dirección de empresas. Publicidad. Relaciones públicas. Medios de comunicación de masas |
Type of document: info:eu-repo/semantics/bachelorThesis info:eu-repo/semantics/bachelorThesis info:eu-repo/semantics/bachelorThesis |
Access rights: info:eu-repo/semantics/openAccess |
Appears in Collections: TFG - Estadística Empresarial
|