Demanda energética de máquinas agrícolas na implantação da cultura do sorgo forrageiro

Forage sorghum can be grown in areas and environmental conditions dry and warm, where the productivity of other forage plants can often be uneconomical. The soil disturbance can be made only on the lines of planting (direct seeding) or entirely from the area for seeding (conventional tillage), as plowing, harrowing, subsoiling and chiseling (minimum tillage). The displacement speed ideal for planting is one in which the groove is opened and closed without removing the over-ground, allowing the distribution of seed spacing and depth constant. The experiment was conducted in a soil classified as Typic Oxisol at Lageado Experimental Farm, Faculty of Agronomic Sciences, UNESP, Botucatu campus. This study aimed to evaluate the response of sorghum in four forward speeds (3, 5, 6 and 9 km h-1) and four systems of soil management: SD (direct seeding), GP (harrow + sowing), LPG (disc harrow and two light disking + sowing) and CR (scarification and seeding). Data was subjected to analysis of variance in a factorial 4 x 4 and a randomized block design with split plots. The following parameters were determined: average speed, average strength of the drawbar, the average power drawbar, theoretical field capacity of the tractor-equipment, fuel consumption per hour. For the conditions under which the experiment was conducted, it was concluded that the hourly fuel consumption was not influenced by tillage systems and was inversely proportional to the increase of speed work, and that the change of speed in the sowing operation did not provide additional the values of average traction force on the bar of the tractor-planter.

Deposição e danos mecânicos em sementes de sorgo utilizando um mecanismo dosador de fluxo contínuo em ensaio de bancada

This study aimed to evaluate the behavior of the deposition and mechanical damage in seeds using a continuous flow metering system under different slope and speed. Part of the study was conducted at Agricultural Research Foundation Agricultural - (FAPA), where seeds that are deposited by a metering system were collected, and the quality analysis verifying the percentage of mechanical damage were conducted at the Faculty of Agricultural Sciences, UNESP, city of Botucatu– SP. The mechanism deposition was subjected to three different speed conditions (4,7, and 10 km.h-1) and three differents working slopes, ( 3%, 8%, and 16%). The results were submitted to Tukey test (p ≤ 0.05), and an analysis of variance with F test at 5% significance level was performed. The results showed an interaction between the factor slope and speed of work, increasing the metering mechanism speed, results in a reduction of the seed deposition at a 3% slope but a working speed of 10 km h-1did not reduce the rate of seed deposition until the slope reaches 16%. Both the slope factor and the working speed caused at least 3.9 and 4.2% more damage to the seeds, respectively.

Processamento da amostra das silagens de grãos úmidos de sorgo com baixo e alto tanino na determinação da digestibilidade aparente em equinos

There were used four castrated male horses, crossbred. It was used randomized block by design for four replications. The objective was to evaluate the effect of using samples, ground or in natura, in the precision of the mobile nylon bag technique (MNBT) in comparison to the total collection (CT) method in oder to estimate the apparent digestibility coefficient (DC) of nutrients of high-moisture grains silage of sorghum with low (SSLT) and high (SSHT) tannin in the feeding equine. No effect was observed for MNBT and CT methods with in natura sample for DC of DM, OM, CE, CP to SSLT (P>0.05), whose average values were 81.71, 84.23, 82.27 e 81.43%, respectively. The MNBT did not showacuracy to evaluate of DC of the SSHT, with ground and in natura samples. It was concluded that MNBT containing in natura sample should be used to determinating the DC of SSLT to DM, OM, CE and CP for equines.

Produção e qualidade de silagens de sorgo em sistema de integração lavoura-pecuária no cerrado de baixa altitude

Pós-graduação em Agronomia - FEIS

Generación automática de reglas de categorización de texto en un método híbrido basado en aprendizaje

En este artículo se evalúan diferentes técnicas para la generación automática de reglas que se emplean en un método híbrido de categorización automática de texto. Este método combina un algoritmo de aprendizaje computacional con diferentes sistemas basados en reglas en cascada empleados para el filtrado y reordenación de los resultados proporcionados por dicho modelo base. Aquí se describe una implementación realizada mediante el algoritmo kNN y un lenguaje básico de reglas basado en listas de términos que aparecen en el texto a clasificar. Para la evaluación se utiliza el corpus de noticias Reuters-21578. Los resultados demuestran que los métodos de generación de reglas propuestos producen resultados muy próximos a los obtenidos con la aplicación de reglas generadas manualmente y que el sistema híbrido propuesto obtiene una precisión y cobertura comparables a la de los mejores métodos del estado del arte.

Método híbrido para categorización de texto basado en aprendizaje y reglas

En este artículo se presenta un nuevo método híbrido de categorización automática de texto, que combina un algoritmo de aprendizaje computacional, que permite construir un modelo base de clasificación sin mucho esfuerzo a partir de un corpus etiquetado, con un sistema basado en reglas en cascada que se emplea para filtrar y reordenar los resultados de dicho modelo base. El modelo puede afinarse añadiendo reglas específicas para aquellas categorías difíciles que no se han entrenado de forma satisfactoria. Se describe una implementación realizada mediante el algoritmo kNN y un lenguaje básico de reglas basado en listas de términos que aparecen en el texto a clasificar. El sistema se ha evaluado en diferentes escenarios incluyendo el corpus de noticias Reuters-21578 para comparación con otros enfoques, y los modelos IPTC y EUROVOC. Los resultados demuestran que el sistema obtiene una precisión y cobertura comparables con las de los mejores métodos del estado del arte.