ABSTRACT Soil mapping is a permanent demand, but the traditional method does not allow fast execution and low cost. Digital soil mapping (DSM) aims to improve the process by working with models that treat soil spatial variability quantitatively. In this perspective, the objective of the study is to perform DSM of the Parnaíba River Delta, Northeastern Brazil, through the decision tree (DT) integration technique using a set of attributes derived from the digital elevation model (DEM) and satellite images as input parameters. Data matrices were created considering different soil groups. The performance of the J48 machine learning algorithm (DT) was assessed for a set of two data matrices: one elaborated for the mapping units of the pre-existing conventional pedological map and the other for a set of associations determined based on the characteristics of the landscape of the study area with close correlation with the existing soils, mainly due to the source material. From the data processing, digital soil maps were created and validated by means of error matrices, whose reference points were classified in the field and validated using a pre-existing traditional soil map of the area. The results revealed that the attributes derived from satellite images stood out from those derived from DEM in predicting soil groups. Based on the validation coefficients applied (overall accuracy, Kappa index, user’s accuracy and producer’s accuracy), the classification quality was satisfactory, despite the complexity of the environment. demand cost (DSM quantitatively perspective Delta Brazil DT (DT (DEM parameters groups J J4 preexisting pre soils material processing overall index users user s producers producer , accuracy) satisfactory environment

Abstract: One of the negative consequences of the industrialization of Mexico favored by the North American Free Trade Agreement (NAFTA), is the emergence of huge industrial corridors associated with the demand for mobility by commuters who move to their workplace. The demand produces mobility patterns that have a serious impact on air pollution in five cities in the state of Guanajuato that, despite being medium in size, outnumber Mexico City in pollution. The objective of this work is to model a data-driven agent based on the beliefs-desires-intentions model, to predict the selection of transport modes using a J48 decision tree algorithm that was designed from data from the 2015 national census (INEGI). The method is mode based l agent programmed in Net logo. The results show that: it is possible to predict the demand of transport considering the: gender, level of education, transfer times and age in the five cities of Guanajuato, in a horizon of three years. With changes in public policies related to mobility and changes in transportation patterns, air pollution would be reduced. The proposed model could be used to support public policies that improve mobility and positively impact air quality in five cities in the state of Guanajuato.

Resumo Neste artigo, técnicas preditivas de mineração de dados são aplicadas para descobrir padrões de desempenho acadêmico na competência linguística dos testes Saber 5 apresentados por alunos de escolas primárias colombianas em 2017. Para isso, foi usada a metodologia CRISP-DM e o socioeconômico, foram consideradas as informações acadêmicas e institucionais das bases de dados do ICFES. Um conjunto de dados mineráveis foi obtido usando técnicas de limpeza e transformação de dados e uma árvore de decisão foi construída com o algoritmo J48 da ferramenta Weka. Entre os preditores dos padrões descobertos estão a natureza e a localização da escola, se os alunos foram ou não reprovados, a faixa etária, a escolaridade da mãe e as taxas de TIC e eletrodomésticos nas residências. O conhecimento produzido nesta pesquisa é informação de qualidade para a tomada de decisões nos MEN e nas secretarias de educação e para os diretores de instituições de educação primária definirem planos de melhoria que resultem na qualidade da educação na Colômbia.

Abstract In this paper, predictive data mining techniques are applied to determine the academic performance from fifth grade students in the Saber 5° tests Language skill at Colombian elementary schools in 2017. We employed the CRISP-DM methodology. Socioeconomic, academic, and institutional information was available at the ICFES databases. A minable dataset was obtained using data cleaning and transformation techniques. A decision tree was built with the Weka tool J48 algorithm. Some of the predictors of the discovered patterns are the nature and location of the school, whether or not students failed a school year, the age group, the mother's educational attainment, and the rates of ICTs and household appliances. The findings of this research serve as quality information for the decision-making at the Ministry of National Education (MEN) and the secretaries of education, and for the directors of elementary educational institutions to define improvement plans that result in the quality of elementary school education in Colombia.

Resumen En este artículo se aplican técnicas predictivas de minería de datos para descubrir patrones de desempeño académico en la competencia de Lenguaje de las pruebas Saber 5° que presentaron los estudiantes de las instituciones educativas colombianas de básica primaria en el año 2017. Para tal fin, se utilizó la metodología CRISP-DM y se tuvo en cuenta la información socioeconómica, académica e institucional de las bases de datos del ICFES. Se obtuvo un conjunto de datos minable utilizando técnicas de limpieza y transformación de datos y se construyó un árbol de decisión con el algoritmo J48 de la herramienta Weka. Entre los factores predictores de los patrones descubiertos están la naturaleza y la ubicación del colegio, si los estudiantes reprobaron o no algún grado, el grupo etario, la educación de la madre y los índices de TICs y electrodomésticos en los hogares. El conocimiento producido en esta investigación es información de calidad para la toma de decisiones en el MEN y las secretarías de educación y para que las directivas de las instituciones educativas de básica primaria definan planes de mejoramiento que redunden en la calidad de la educación en Colombia.

Resumen El objetivo de este artículo es medir a través de técnicas inteligentes el nivel de vida universitaria de una población universitaria. En esta investi gación, una variable dependiente llamada calidad de vida universitaria se toma en cuenta junto con 10 variables independientes: carga académica, recursos económicos, relación con compañeros de clase, relación con docentes, currículum, actividades extracurriculares, vivienda actual, relaciones familiares, estado emocional y entorno universitario. Para el muestreo de estas variables se llevaron a cabo 127 encuestas a estu diantes de una universidad pública ubicada en la región central del país. Subsecuentemente, las variables más relevantes fueron seleccionadas a través de técnicas estadísticas con el fin de establecer un archivo de análisis desde el algoritmo de clasificación de árbol de decisiones J48 de la plataforma Weka. EL resultado demuestra, con una efectividad de más del 80 % que las variables más influyentes en la calidad de vida universitaria son: entorno universitario, vivienda actual, estado emocional y relación con los docentes, encontrando, varias veces, que la calidad de vida también puede depender de variables externas a la universidad, tales como: vivienda actual y estado emocional. Estos resultados son de gran importancia en el diseño de las políticas universitarias por venir.

ABSTRACT The objective of this article is to measure through intelligent techniques, the quality of university life in a university population. In this investigation, a dependent variable called quality of university life is taken, as well as 10 independent variables: Academic load, economic resources, rela tionship with classmates, relationship with professors, curriculum, extracurricular activities, current housing, family relationships, emotional state and university environment. In the samplings of these variables, 127 surveys were carried out on university students of a public university located in the central region of Colombia. Subsequently, the most relevant variables were selected throughout statistical techniques, in order to establish a file to be analyzed through the decision tree classification algorithm J48from the Weka platform. The results show, with over an 80 % effectiveness, that the most influential variables in the quality of life of a university student are: University environment, current housing, emotional state, and relationships with professors. Finding a lot of times that the quality of university life can also depend of external variables to the university such as: Current housing and emotional state. These results are of great importance in the design of new university policies.

This study evaluates the ability of various machine-learning techniques to predict low academic performance among Nigerian tertiary students. Using data collected from undergraduate student records at Niger Delta University in Bayelsa State, the research applies the cross-industry standard process for data mining (CRISP-DM) research methodology and the Waikato Environment for Knowledge Analysis (WEKA) tool for modelling. Five machine-learning classifier algorithms are tested-J48 decision tree, logistic regression (LR), multilayer perceptron (MLP), naïve Bayes (NB), and sequential minimal optimisation (SMO)-and it is found that MLP is the best classifier for the dataset. The study then develops a predictive software application, using PHP and Python, for implementation of the MLP model, and the software achieves 98% accuracy.

Abstract: This research study proposes a classification system to identify and prevent occupational accidents in fiber storage warehouses at a pulp and paper facility. The present analysis is based on variables including pedestrian circulation, bobcat, trailer trucks, access, pedestrian circulation zones, and handrails. The proposed methodology defines and trains the system by using occupational accident event data collected at the facility. Three different predicting algorithms are used: J48 decision-making trees, Naive Bayes, and Bayesian nets. The results show that the J48 decision-making tree algorithm accurately identifies possible occupational accidents 90% of the time. It is concluded that identifying variables involved in occupational accidents allows generating a C4.5 (J48) decision-making tree that can be used as a support tool to prevent occupational accidents.

Resumen: En este artículo se propone un sistema de clasificación para la identificación y la prevención de accidentes laborales en las bodegas de almacenamiento de fibra en una empresa de pulpa de papel. Eso se hace con base en el análisis de variables que incluyen la circulación de peatones, bobcat, tractocamiones, acceso, zonas de circulación peatonal y barandas. La metodología propuesta define y entrena el sistema propuesto con datos recopilados en planta respecto a accidentalidad e incidentes laborales. Se compararon los resultados de tres algoritmos: redes bayesianas, naive bayes y árboles de decisión. Los resultados muestran que, en el 90% de los casos, el clasificador de arboles de decisión J48 identifica correctamente los casos posibles de accidentalidad laboral. Se concluye que la identificación de los atributos que generan la ocurrencia de incidentes y accidentes laborales, permite generar un árbol de decisión C4.5 (J48) que sirve de herramienta y soporte para labores de prevención de accidentes laborales.

ABSTRACT Accountability requires the existence of reliable and valid information and auditing is one of fundamental bases for the accountability process. Thus educating the optimum method is of great importance. With the presence of a great volume of information, using the prediction methods can contribute the auditors in this respect. This research aims to compare a variety of methods of teaching of that. In current research, J48, random forest, vector machine and neural network have been used. Research population involves the corporates accepted by Tehran Stock Exchange in 2008-2017. Here, 19 financial and non-financial independent variables have been applied in two groups of test and training. Also, the independent auditor report has been classified into two groups of acceptable and conditional. Comparing the above-mentioned methods has indicated that random forest algorithm with the prediction accuracy average as 78.83% was the most optimum model to predict the report type in the both groups and other models involving J48, support vector machine, CART decision tree and finally, artificial neural network were of the most accuracy.

RESUMEN La rendición de cuentas requiere la existencia de información confiable y válida y la auditoría es una de las bases fundamentales para el proceso de rendición de cuentas. Por lo tanto, educar el método óptimo es de gran importancia. Con la presencia de un gran volumen de información, el uso de los métodos de predicción puede contribuir a los auditores a este respecto. Esta investigación tiene como objetivo comparar una variedad de métodos de enseñanza de eso. En la investigación actual, se han utilizado J48, bosque aleatorio, máquina de vectores y red neuronal. La población de investigación involucra a las empresas aceptadas por la Bolsa de Teherán en 2008-2017. Aquí, se han aplicado 19 variables independientes financieras y no financieras en dos grupos de prueba y capacitación. Además, el informe del auditor independiente se ha clasificado en dos grupos de aceptable y condicional. La comparación de los métodos mencionados anteriormente ha indicado que el algoritmo de bosque aleatorio con un promedio de precisión de predicción de 78.83% fue el modelo más óptimo para predecir el tipo de informe en ambos grupos y otros modelos que involucran J48, máquina de vectores de soporte, árbol de decisión CART y finalmente, La red neuronal artificial fue de la mayor precisión.

ABSTRACT Neurodegenerative diseases are being treated by modulating adenosine receptors with more effective, safe and selective antagonists. The objective of the study was to develop a methodology to obtain classification models based on decision tree algorithms and descriptors from 0D to 2D of non-congenital families of organic compounds to qualitatively predict ligand-RAA2A affinity. For this purpose, a non-congeneric database of 315 antagonists was constructed and cured with its inhibition constant in nano molar, labeled as potent and weak. The Dragon and ISIDA / QSPR programs were used to calculate molecular descriptors and five groups of descriptors were obtained. In each group 50 descriptors were selected using the mRMR criterion. The database was divided into Training, Test and External series through a random selection and a generalized k-means cluster analysis. Classifiers were developed and validated using the WEKA program. The results were analyzed using the statistical tests of Friedman and Wilcoxon. The significant influence of parameter m of algorithm J48 on the predictivity was verified for the models that used the descriptors of the aug.a-b and hyb.aug.a groups of ISIDA / QSPR. The best performance model was obtained from the selected descriptors of the ISIDA-all group with a value of m = 6 and reached 90.6% prediction on the External series. The methodology developed to obtain classification models based on decision tree algorithms and descriptors from 0D to 2D of non-congenital families of organic compounds is effective in qualitatively predicting ligand-RAA2A affinity with accuracy, specificity and selectivity greater than 90 %.

RESUMEN Las enfermedades neurodegenerativas están siendo tratadas mediante la modulación de los receptores de adenosina con antagonistas más eficaces, seguros y selectivos. El objetivo del estudio consistió en desarrollar una metodología para obtener modelos de clasificación sobre la base de algoritmos de árboles de decisión y descriptores de 0D a 2D de familias no congenéricas de compuestos orgánicos para predecir cualitativamente la afinidad ligando-RAA2A. Para ello se construyó y curó una base de datos no congenérica de 315 antagonistas con su constante de inhibición en nano molar, etiquetados como potentes y débiles. Se utilizaron los programas Dragon e ISIDA/QSPR para calcular descriptores moleculares y se obtuvieron cinco grupos de descriptores. En cada grupo se seleccionaron 50 descriptores usando el criterio mRMR. La base de datos se dividió en series de Entrenamiento, Prueba y Externa mediante una selección aleatoria y un análisis de clúster k-means generalizado. Se desarrollaron y validaron clasificadores utilizando el programa WEKA. Los resultados fueron analizados mediante las pruebas estadísticas de Friedman y Wilcoxon. Se comprobó la influencia significativa del parámetro m del algoritmo J48 en la predictividad, para los modelos que usaron los descriptores de los grupos aug.a-b e hyb.aug.a del ISIDA/QSPR. El modelo de mejor desempeño se obtuvo de los descriptores seleccionados del grupo ISIDA-todos con un valor de m=6 y alcanzó 90.6% de predicción sobre la serie Externa. La metodología desarrollada para obtener modelos de clasificación sobre la base de algoritmos de árboles de decisión y descriptores de 0D a 2D de familias no congenéricas de compuestos orgánicos es efectiva para predecir cualitativamente la afinidad ligando-RAA2Acon una exactitud, especificidad y selectividad superiores al 90%.

ABSTRACT Background and aim: the aim of the study was to use a validated questionnaire to identify factors associated with the development of gastric cancer (GC) in the Mexican population. Methods: the study included cases and controls that were paired by sex and ± 10 years of age at diagnosis. In relation to cases, 46 patients with a confirmed histopathological diagnosis of adenocarcinoma-type GC, as reported in the hospital records, were selected, and 46 blood bank donors from the same hospital were included as controls. The previously validated Questionnaire to Find Factors Associated with Gastric Cancer (QUFA-GC(c)) was used to collect data. Odds ratio (OR) and 95% confidence interval (IC) were estimated via univariate analysis (paired OR). Multivariate analysis was performed by logistic regression. A decision tree was constructed using the J48 algorithm. Results: an association was found by univariate analysis between GC risk and a lack of formal education, having smoked for ≥ 10 years, eating rapidly, consuming very hot food and drinks, a non-suitable breakfast within two hours of waking, pickled food and capsaicin. In contrast, a protective association against GC was found with taking recreational exercise and consuming fresh fruit and vegetables. No association was found between the development of GC and having an income that reflected poverty, using a refrigerator, perception of the omission of breakfast and time period of alcoholism. In the final multivariate analysis model, having no formal education (OR = 17.47, 95% CI = 5.17-76.69), consuming a non-suitable breakfast within two hours of waking (OR = 8.99, 95% CI = 2.85-35.50) and the consumption of capsaicin ˃ 29.9 mg capsaicin per day (OR = 3.77, 95% CI = 1.21-13.11) were factors associated with GC. Conclusions: an association was found by multivariate analysis between the presence of GC and education, type of breakfast and the consumption of capsaicin. These variables are susceptible to intervention and can be identified via the QUFA-GC(c).

Este trabalho teve o objetivo de identificar diferenças no padrão de vocalização em função do sexo dos animais e diferentes situações de estresse. Foram utilizados 150 animais machos castrados e 150 fêmeas (linhagem Dalland®), com 100 dias de idade. Os suínos foram submetidos a diferentes situações de estresse: sede (animais sem acesso à água), fome (suínos sem acesso ao alimento), estresse térmico (ITU superior a 74) e BEA (animais com alimento e água, com ITU abaixo de 70). Foram registrados os sinais acústicos a cada 30 minutos, totalizando seis coletas para cada situação de estresse. Posteriormente, os áudios foram analisados pelo software Praat® 5.1.19, gerando um espectro sonoro. Para a determinação das condições de estresse, os dados foram processados no programa computacional WEKA® 3.5, utilizando o algoritmo de árvore de decisão C4.5, conhecido como J48 no ambiente do programa computacional WEKA®, considerando validação cruzada com amostras de 10% (10-fold cross-validation). De acordo com a Árvore de Decisão, o atributo acústico mais importante para a classificação das condições de estresse foi a Intensidade do som (nó raiz). Não foi possível identificar o sexo dos animais pelo registro vocal, utilizando os atributos testados. Foi gerada uma árvore de decisão para reconhecimento de situação de fome, sede e estresse térmico em suínos, a partir de registros da intensidade do som, da frequência de Pitch e da Formante 1.

This study aimed to identify differences in swine vocalization pattern according to animal gender and different stress conditions. A total of 150 barrow males and 150 females (Dalland® genetic strain), aged 100 days, were used in the experiment. Pigs were exposed to different stressful situations: thirst (no access to water), hunger (no access to food), and thermal stress (THI exceeding 74). For the control treatment, animals were kept under a comfort situation (animals with full access to food and water, with environmental THI lower than 70). Acoustic signals were recorded every 30 minutes, totaling six samples for each stress situation. Afterwards, the audios were analyzed by Praat® 5.1.19 software, generating a sound spectrum. For determination of stress conditions, data were processed by WEKA® 3.5 software, using the decision tree algorithm C4.5, known as J48 in the software environment, considering cross-validation with samples of 10% (10-fold cross-validation). According to the Decision Tree, the acoustic most important attribute for the classification of stress conditions was sound Intensity (root node). It was not possible to identify, using the tested attributes, the animal gender by vocal register. A decision tree was generated for recognition of situations of swine hunger, thirst, and heat stress from records of sound intensity, Pitch frequency, and Formant 1.

The aim was to find craniofacial morphology patterns in a multivariate cephalometric database using a clustering technique. Cephalometric analysis was performed in a sample of 100 teleradiographs collected from Chilean orthodontic patients. Thirty cephalometric measurements were taken from commonly used analysis. The computed variables were used to perform a clustering analysis with the k-means algorithm to identify patterns of craniofacial morphology. The J48 decision tree was used to analyze each cluster, and the ANOVA test to determine the statistical differences between the clusters. Four clusters were found that had significant differences (P<0.001) in 24 of the 30 variables studied, suggesting that they represent different patterns of craniofacial form. Using the decision tree, 8 of the 30 variables appeared to be relevant for describing the clusters. The clustering analysis is effective in identifying different craniofacial patterns based on a multivariate database. The distinct clusters appear to be caused by differences in the compensation process of the facial structure responding to a genetically determined cranial and mandible form. The proposed method can be applied to several databases, creating specific classifications for each one of them.

El objetivo fue encontrar patrones morfológicos craneofaciales, a partir de una base de datos cefalométricos multivariada, utilizando una técnica de clustering. Se realizó un análisis cefalométrico a una muestra de 100 telerradiografías pertenecientes a pacientes chilenos de ortodoncia. Treinta medidas cefalométricas obtenidas de los análisis más utilizados fueron registradas. Las variables computadas se utilizaron para realizar un análisis de clustering con el algoritmo k-medias, para identificar patrones de morfología craneofacial. El árbol de decisión J48 se utilizó para analizar cada cluster, y test de ANOVA para determinar diferencias estadísticamente significativas entre los clusters. Se encontraron cuatro clusters con diferencia estadísticamente significativas (p<0,001) en 24 de las 30 variables estudiadas, lo que sugiere que efectivamente corresponden a diferentes patrones craneofaciales. Utilizando el árbol de decisión, se pudo determinar que 8 de las 30 variables resultaron ser relevantes en la definición de los clusters. El análisis de clustering es efectivo en identificar patrones morfológicos craneofaciales usando una base de datos multivariada. Los distintos cluster encontrados, aparentemente se formarían a partir de diferencias en el proceso de compensación de la estructura facial, en respuesta a la forma mandibular genéticamente determinada. El método propuesto puede ser aplicado a múltiples bases de datos, creando clasificaciones específicas para cada una de ellas.

O estudo foi realizado com o banco de dados do Programa de Melhoramento Genético de Poedeiras de Ovos Marrons do Laboratório de Avicultura (LAVIC) do Departamento de Zootecnia da Universidade Federal de Santa Maria (UFSM), onde são utilizadas aves das raças Plymouth Rock White (PRW), Plymouth Rock Barred (PRB) e Rhode Island Red (RIR). Os dados utilizados foram referentes a duas gerações das três raças durante o primeiro ciclo de produção, dos anos de 2009 e 2010. Os parâmetros avaliados foram: taxa de postura semanal, peso dos ovos e peso corporal. O estudo foi dividido em duas abordagens de análise, a primeira análise foi realizada com os dados produtivos com o objetivo de identificar pontos críticos de seleção fenotípica, relacionados à caracterização zootécnica, e a segunda abordagem consistiu na associação dos mesmos parâmetros produtivos, com os dados climáticos de Santa Maria - RS. Os dados foram analisados pela técnica de Mineração de Dados, sendo utilizada a tarefa de classificação com a construção de árvore de decisão pelo algoritmo J48. A mineração dos dados indicou a taxa de postura média na 25ª semana de idade e posteriormente o peso médio dos ovos na 33ª semana de idade como as principais variáveis relacionadas com a diferenciação das raças estudadas, (precisão de 0,833). Na segunda abordagem, as árvores de classificação obtidas apresentaram precisão entre 0,88 e 0,91, sugerindo uma forte associação entre as variáveis ambientais e o desempenho produtivo das aves. Poedeiras PRW são mais sensíveis ao calor que as RIR e PRB e apresentam, na fase de pré-pico, uma maior taxa de postura. A Mineração de Dados permite classificar a produtividade das raças de poedeiras de ovos marrons, podendo indicar as influências ambientais sobre o fenótipo das diferentes raças estudadas.

The study was conducted with the database obtained from the Breeding Program of Brown Egg Laying Hens of the Poultry Laboratory (LAVIC) of the Department of Animal Science of the Federal University of Santa Maria (UFSM), where it is used poultry breeds such as White Plymouth Rock (PRW), Barred Plymouth Rock (PRB) e Red Rhode Island (RIR). The data used referred to two generations of these three breeds during the first production cycle of 2009 and 2010. The parameters evaluated were: weekly rate of egg laying, egg weight and body weight. The research was divided into two approaches of analysis, the first being performed with the production data, aiming to identify critical points of phenotypic selections related to zootechnical characterization, and the second approach consisted in the association of the same production parameters together with the weather data of Santa Maria - RS. The data was analyzed using the Data Mining technique, using the classification task by building the decision tree with the J48 algorithm. The data mining indicated the average rate of egg laying at 25 weeks of age and the average weight of eggs at 33 weeks of age, and pointed these as the main variables related to the differentiation of breeds studied, (accuracy of 83.3%). On the second approach, the decision trees obtained showed accuracy between 88% and 91%, suggesting a strong association between the environmental variables and the production performance of the birds. The PRW breed is more sensitive to heat than the RIR and the PRB breed and they showed a higher rate of production during the pre-peak phase. The data mining allowed to classify the productivity of the brown egg laying hens, which may indicate environmental influences on the phenotype of the different breeds studied.

Problemas locomotores impedem a ave de se movimentar livremente, prejudicando o bem-estar e a produtividade, além de gerarem lesões nas pernas dos frangos. O objetivo deste trabalho foi avaliar a influência da idade, do uso de vitamina D, da assimetria de membros e do gait score, no grau de lesões de perna em frangos de corte, utilizando mineração de dados. A análise foi realizada em um conjunto de dados obtidos de um experimento de campo, em que foram utilizados dois grupos de aves com 30 aves cada, sendo um grupo-controle e outro tratado com vitamina D. Foram avaliados o gait score, a assimetria entre os dedos dos pés direito e esquerdo, e o grau de lesões de perna. O software Weka® foi utilizado na mineração de dados. Em particular, o algoritmo C4.5 (também conhecido como J48 no ambiente Weka) foi utilizado para a geração de uma árvore de decisão. Os resultados mostraram que a idade é o fator que mais influencia o grau de lesões de perna e que os dados provenientes das avaliações de gait score não se mostraram confiáveis para estimar problemas locomotores em frangos de corte.

Locomotor problems prevent the bird to move freely, jeopardizing the welfare and productivity, besides generating injuries on the legs of chickens. The objective of this study was to evaluate the influence of age, use of vitamin D, the asymmetry of limbs and gait score, the degree of leg injuries in broilers, using data mining. The analysis was performed on a data set obtained from a field experiment in which it was used two groups of birds with 30 birds each, a control group and one treated with vitamin D. It was evaluated the gait score, the asymmetry between the right and left toes, and the degree of leg injuries. The Weka ® software was used in data mining. In particular, C4.5 algorithm (also known as J48 in Weka environment) was used for the generation of a decision tree. The results showed that age is the factor that most influences the degree of leg injuries and that the data from assessments of gait score were not reliable to estimate leg weakness in broilers.

Quando levantamentos de solos não estão disponíveis para atividades de planejamento de uso das terras, técnicas de mapeamento digital de solos podem ser úteis. Mapeadores de solos podem processar as informações espaciais rapidamente, auxiliando na execução de levantamentos de solos tradicionais ou prevendo a ocorrência de classes de solos na paisagem. Avaliaram-se técnicas de análise de decisão na predição da ocorrência de classes de solos em áreas de encostas basálticas no Sul do Brasil. Várias combinações de tipos de algoritmos de árvore de decisão e quantidade de elementos nos nós terminais das árvores de decisão foram testadas usando mapas de solos com a legenda original e com legenda simplificada. Em geral, o uso de árvores de decisão foi eficaz na predição de ocorrência de unidades de mapeamento de solos. Menor número de elementos no nó terminal das árvores de decisão produziu acurácias mais altas e a simplificação da legenda (agregação) reduziu a precisão das predições. O algoritmo J48 teve melhor desempenho que BF Tree, RepTree, Random Tree, e Simple Chart.

When soil surveys are not available for land use planning activities, digital soil mapping techniques can be of assistance. Soil surveyors can process spatial information faster, to assist in the execution of traditional soil survey or predict the occurrence of soil classes across landscapes. Decision tree techniques were evaluated as tools for predicting the ocurrence of soil classes in basaltic steeplands in South Brazil. Several combinations of types of decicion tree algorithms and number of elements on terminal nodes of trees were compared using soil maps with both original and simplified legends. In general, decision tree analysis was useful for predicting occurrence of soil mapping units. Decision trees with fewer elements on terminal nodes yield higher accuracies, and legend simplification (aggregation) reduced the precision of predictions. Algorithm J48 had better performance than BF Tree, RepTree, Random Tree, and Simple Chart.