Implementaci?n de una red neuronal artificial para predecir la calidad del sabor de la cerverza
- Authors
- Ayala Cruz, Edy Leonardo
- Format
- DoctoralThesis
- Status
- publishedVersion
- Description
Brewing beer involves a number of steps to produce this popular alcoholic drink. Nowadays, the brewing industry has improved the techniques by automating most of the processes where usually humans use to take decisions. However, in small-scale breweries, the process is still been monitored and controlled by humans. One of the most important parts is the Sensory Panel test. This involves a human measurement of the quality of the beer in each batch produced. This information is useful for the brewers who can make decisions based in this analysis. This testing is very subjective and difficult to reproduce since it depends only on the panellist impression and experience. Artificial Neural Networks (ANN) are computational models, which can accurately predict, classify and organize database on mathematical algorithms. They have been developed and used since 1943. As a result, today many industries have adopted this method to improve the process in different areas. For instance, human decisions making, which involves many subjective aspects, is one of the areas where ANN have the best potential of application. ANNs work similarly as the brain, simulating the neurons synapsis and the way they compute information. ANNs and their ability to learn is what makes this new tendency very popular worldwide. The aim of this project is to use ANNs in order to predict beer taste quality in a particular brewing company. This report will present how the data was collected, a description of an ANN and its relevant parts, the method used for training, the experiments performed, the analysis of the architectures generated, the description of the algorithm used for the training, the programming and all the components related with the functional aspects of the prototype.
El presente estudio es un dise?o e implementaci?n de una red neuronal para la predicci?n del sabor de la cerveza antes durante el proceso de elaboraci?n. Este programa permite a los dise?adores de cerveza estimar el resultado de la producci?n antes de iniciar el proceso. As?, los productores pueden garantizar un mejor resultado mejorando la eficiencia de producci?n.
- Publication Year
- 2015
- Language
- eng
- Topic
- REDES NEURONALES
PREDICCI?N SABOR CERVEZA
PROCESO DE ELABORACI?N CERVEZA
INGENIER?A DE ALIMENTOS
- Repository
- Repositorio SENESCYT
- Rights
- openAccess
- License