ABSTRACT
The different types of sugar employed in the food industry exhibit chemical similarity and are mostly dominated by sucrose. Owing to the sugar origin of and differences in production, the presence of certain minor organic compounds differs. To differentiate between sugars based on their botanical source, geographical origin, or storage conditions, commercial brown sugars and sugar beet extracts were analyzed by 1H NMR spectroscopy applying a segmented analysis by means of multivariate curve resolution-alternating least squares (MCR-ALS). Principal component analysis and partial least squares-discriminant analysis yielded excellent differentiation between sugars from different sources after the application of this preprocessing strategy; without loss of chemical information and with direct interpretation of the results. By applying a segmented analysis via MCR-ALS to 1H NMR sugar data, similar spectroscopic profiles could be differentiated. This improved the selectivity of 1H NMR spectroscopy for sugar source differentiation which can be useful for industrial sugar authentication purposes.
Subject(s)
Carbohydrates , Sugars , Multivariate Analysis , Least-Squares Analysis , Magnetic Resonance SpectroscopyABSTRACT
Esta investigación tiene como principal propósito analizar la probabilidad de deserción temprana de los estudiantes de la Universidad de los Llanos mediante un modelo de regresión logística. Además, identificar cuáles son las variables de mayor incidencia en la deserción temprana. Para la construcción del modelo se tomó como muestra la información de 574 estudiantes que ingresaron en la cohorte 2015-2 y que para el periodo 2018-1 eran registrados como vigentes o reportaban como último periodo matriculado cualquiera de los cuatro primeros semestres. Los resultados dan cuenta de que un buen puntaje en la prueba Saber 11, el ser mujer, no haber reprobado años durante el bachillerato, el haber cursado estudios antes y si los padres conviven, disminuye la probabilidad de deserción; así mismo, el haber egresado de un colegio privado aumenta esta probabilidad. De todas las facultades en las que se puede matricular el estudiante, las que mayor riesgo tienen de deserción temprana son la de Ciencias Básicas e Ingeniería y la Facultad de Ciencias de la Salud.
The main purpose of this study is to analyze the early dropout probability in the University of Los Llanos students using a logistic regression model and to identify which are the variables with a higher incidence value. To build the model, data were taken from 574 students who had been admitted in the 2015 second semester and were registered as active or whose last semester was one of the last four. As result, the dropout factor decreases if the student had a high score in Saber 11 test, is a woman, had not to fail any school year, had taken some courses before and his parents are alive and increase if the student comes from a private school. The faculties with the high dropout index are Basic Sciences and Engineering, and Health Sciences.
