La presente investigación tuvo como objetivo determinar que los modelos de aprendizaje de máquina evalúan eficazmente el riesgo crediticio de personas naturales de una institución financiera de Chiclayo que el modelo clásico de credit scoring estimado mediante la Regresión Logística. La investigación es de tipo descriptivo explicativo y predictivo para lo cual se trabajó con la metodología CRISP-DM. Para el desarrollo de la investigación se utilizaron los modelos de aprendizaje de maquina tales como, Árboles de Calsificación, Redes Neuronales, Maquinas de Soporte Vectorial y el modelo clásico de la Regresión Logística; la base de datos estuvo contituída por 2464 clientes, de los cuales se utilizó el 70% de la base para el entrenamiento el 30% restante para la validación.

This research work had as main objective optimizing the general acceptability though affective tests and response surface methodology of a dry cacao powder mixture based beverage. We obtained formulations of mixtures of cacao powder with different concentrations of 15%, 17.5% and 20%, as well as lecithin concentrations of 0.1%; 0.3%; and 0.5% maintaining a constant content of sugar (25 %), Vanillin (1%) that included cacao powder with different pH values: natural (pH 5) and alkalinized (pH 6.5 and pH 8) and water by difference to 100%, generating a total of fifteen treatments to be evaluated, according to the Box-Behnen design for three factors. The treatments underwent satisfaction level tests to establish the general acceptability. The treatment that included cacao powder with a concentration of 17.5 %, pH 6.5 and lecithin concentration of 0.3 % obtained the best levels of acceptabilit...

This research work had as main objective optimizing the general acceptability though affective tests and response surface methodology of a dry cacao powder mixture based beverage. We obtained formulations of mixtures of cacao powder with different concentrations of 15%, 17.5% and 20%, as well as lecithin concentrations of 0.1%; 0.3%; and 0.5% maintaining a constant content of sugar (25 %), Vanillin (1%) that included cacao powder with different pH values: natural (pH 5) and alkalinized (pH 6.5 and pH 8) and water by difference to 100%, generating a total of fifteen treatments to be evaluated, according to the Box-Behnen design for three factors. The treatments underwent satisfaction level tests to establish the general acceptability. The treatment that included cacao powder with a concentration of 17.5 %, pH 6.5 and lecithin concentration of 0.3 % obtained the best levels of acceptabilit...