Paquete estadístico para análisis de regresión
Archivos
Fecha
2012-07-03
Autores
Herrera Pallares, Sandro
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
Esta tesis resume los conceptos fundamentales del análisis de regresión. Se muestra una descripción detallada de los algoritmos "paso a paso", "introducir todas las variables" y "mejores subconjuntos". Se explican cuidadosamente los supuestos de los modelos matemáticos. Se presentan varias pruebas de hipótesis para demostrar la normalidad de los residuos y la homogeneidad de sus varianzas.
Se presentan los aspectos principales del análisis, diseño e implementación del software “Linear Regression Package 1.0”. Además se elaboró un manual de usuario.
Se presentan dos aplicaciones con datos reales. En la primera se encuentran varios modelos para resolver un problema de quimioinformática utilizando los métodos “paso a paso” y “mejores subconjuntos”. En la segunda aplicación se resuelve un problema de Ingeniería Civil. Se obtienen modelos para predecir la fortaleza del hormigón. El primer modelo usa las variables originales, mientras que el segundo usa sus interacciones. En todos los casos se probó el cumplimiento de los supuestos. Todos los cálculos se realizaron con ayuda del software elaborado.
This thesis shows the fundamental concepts of regression analysis. The full description of the algorithms “stepwise”, “enter” and “best subsets” are provided. The assumptions of the mathematical models are carefully explained. Several hypothesis tests to prove the normality of the residuals and the homogeneity of their variances are explained. The main aspects of the analysis, design and implementation of software "Linear Regression Package1.0" are provided. Also a user manual was developed. Two different applications with real data are presented. In the first there are calculated several models to solve a cheminformatics problem using stepwise method and best subsets method. The second application solves a problem of Civil Engineering. Models to predict the strength of concrete are presented. The first model uses the original variables, while the second one uses their interactions. The required assumptions of the equations are proved in all cases. All calculations were performed using the software developed.
This thesis shows the fundamental concepts of regression analysis. The full description of the algorithms “stepwise”, “enter” and “best subsets” are provided. The assumptions of the mathematical models are carefully explained. Several hypothesis tests to prove the normality of the residuals and the homogeneity of their variances are explained. The main aspects of the analysis, design and implementation of software "Linear Regression Package1.0" are provided. Also a user manual was developed. Two different applications with real data are presented. In the first there are calculated several models to solve a cheminformatics problem using stepwise method and best subsets method. The second application solves a problem of Civil Engineering. Models to predict the strength of concrete are presented. The first model uses the original variables, while the second one uses their interactions. The required assumptions of the equations are proved in all cases. All calculations were performed using the software developed.
Descripción
Palabras clave
Paquete Estadístico, Análisis de Regresión, Métodos Paso a Paso, Problema de Quimioinformática, Problema de Ingeniería Civil