This work provides an analysis of using the evolutionary algorithm EPNet to create ensembles artificial neural networks solve a range forecasting tasks. Several previous studies have tested in classification field, taking best individuals problem and creating improve performance. But no analyzed behavior detail for time series forecasting, nor used try predictions. Thus, aim this is compare ensemble approach, two linear combination methods calculate output, against individual found. Since...

Modularity is known to have benefits for neural systems and their evolution, this paper aims improve the evolutionary network algorithm EPNet take advantage of those benefits. Neural networks exist with varying degrees modularity ranging from pure modular characterized by disjoint partitions hidden nodes no communication between modules, homogeneous significant connections throughout. In are apparently that can be seen some degree if reorganized appropriately. paper, a measure presented...

Chaotic time series have been successfully predicted with the EPNet algorithm through evolution of artificial neural networks. However, input feature selection problem has either not fully explored before or compared against other algorithms in literature. This paper presents four derived from classical to evolve three different chaotic series: Logistic, Lorenz and Mackey-Glass. Additionally, some flaws prediction field that may be considered future works are discussed. A comparison previous...

Similarity search in time series databases has been the focus of a lot research. It is non-trivial problem due their peculiar structure, high dimensionality (C. Faloutsos et. al, 1994), noise, and feature correlation; it also heart many data mining applications, like clustering, classification, rule discovery query by content. In general, similarity measured with Euclidean metric, but curse remains (K. Chan A. Wai-chee, 1999). Basically, there are four major techniques that have proposed...

Resumen.En el presente trabajo se utilizó la segmentación de letras fotografías tomadas mediante cámara del Parrot AR Drone, con objetivo establecer una interacción estímulo-respuesta, donde imagen original en formato combinación colores Rojo, Verde y Azul (RGB) segmentó por color (escogiendo canal rojo).Una vez que reconoce carácter, Drone ejecuta acción correspondiente.Se utilizaron inicialmente patrones numéricos libres ruido posteriormente agregaron algunos pixeles a hacer más robusto...

Resumen.Actualmente el cultivo de especies acuícolas para consumo humano es una práctica que se realiza en todo mundo.De ahí, oxígeno disuelto uno los parámetros más importantes la calidad del agua sistemas acuícolas, por lo mantener buenos niveles concentración este gas esencial producción exitosa.Este trabajo presenta un modelo predictivo basado Redes Neuronales Artificiales (RNAs), las cuales son diseñadas con algoritmo evolutivo FS-EPNet estimar base a cantidad camarón blanco.Los...

In computer sciences, matrices are widely used for representing different kinds of information. Measuring similarity among square is an interesting open problem in sciences. Furthermore, eigenvalues and eigenvectors a powerful way characterizing matrices. this paper we introduce new measure two data the same class; idea based on evaluating effect conjugate one matrix with other matrix, vice versa. Some experimental results showed order to analyze exemplify Eigenconjugation as approach

The effects of crime are diverse and complex, ranging from psychological physical traumas faced by victims, to negative impacts on the economy a whole nation. In this paper, an agent-based simulation framework analyze its causes is proposed implemented. models simulates both 1) events as consequence set interrelated social individual-level factors, 2) opportunities, i.e., combinations circumstances that enable person commit crime. selection factors design agent supported by, based on,...

Resumen.En este artículo, se propone el uso de un ensamble clasificadores para determinar perfil académico del estudiante, basado en su promedio general y datos relacionados a los factores educativos: actividades estudio, formas aprendizaje hábitos estudio.Los usados obtuvieron cuestionario socioeconómico aplicado

Resumen.El problema de localización placas automóvil ha sido ampliamente estudiado con anterioridad, dando resultados favorables.Sin embargo,

The theft of auto parts in Mexico has become a serious problem recent years. implementation artificial intelligence methods can help identify criminal behaviors that are commonly associated with this crime. However, there is no publicly available dataset focused on solving security problem; furthermore, specific references the literature regarding identification theft. This paper proposes creation related to type theft, and application predictive models for automatic exhibited by individuals...

In computer sciences, matrices are widely used for representing different kinds of information. Furthermore, measuring similarity among is an interesting open problem. this paper we introduce a new measure two data the same class; idea based on evaluating effect conjugating singular values and vectors one matrix with other matrix, vice versa. A brief analysis value conjugation (SV-conjugation) also included in order to extend understanding concept. Finally, some experimental results...