Classification and regression tree analysis for predicting morphological hand types based on radiography data / Análisis de árbol de clasificación y regresión para predecir tipos morfológicos de la mano basados en datos radiográficos

SUMMARY: Analysis and systematization of the longitudinal dimensions of the phalanges of the index and ring fingers for the classification of morphological types of the hand using classification and regression trees (CART). X-rays of the hands of 50 men and 50 women (mean age 47.16 (10.1) years, range 23-65 years) were studied. Each hand, depending on the ratio of the length of the index and ring fingers, was classified into three types: radial (R, 2d>4d), indefinite (N, 2d=4d), ulnar (U, 2d>4d). Morphometry of radiographs included measurements of the lengths of the proximal (PP), middle (MP), and distal (DP) phalanges. The sex differences of the analyzed indicators are statistically significant. There were no significant bilateral differences between the phalanges of the II and IV fingers in length, regardless of sex (p>0.05). A set of rules for classifying the morphological types of the hand depending on the lengths of the phalanges of the index and ring fingers was created by constructing a binary decision tree. The CART method demonstrates the usefulness of this statistical procedure for developing a scientifically based prediction of the morphological type of the hand. The results of this study can be the basis of an algorithm for determining the morphological type of the hand depending on the length of the phalanges of the fingers.

RESUMEN: En este estudio se realizó un análisis y sistematización de las dimensiones longitudinales de las falanges de los dedos índice y anular para la clasificación de tipos morfológicos de la mano, mediante árboles de clasificación y regresión (CART). Se estudiaron radiografías de las manos de 50 hombres y 50 mujeres (edad media 47,16 (10,1) años, rango 23-65 años). Cada mano, según la proporción de la longitud de los dedos índice y anular, se clasificó en tres tipos: radial (R, 2d> 4d), indefinida (N, 2d = 4d), ulnar (U, 2d> 4d). La morfometría de las radiografías incluyó mediciones de las longitudes de las falanges proximal (FP), media (FM) y distal (FD). Las diferencias de sexo de los indicadores analizados fueron estadísticamente significativas. No hubo diferencias bilaterales significativas entre las falanges de los dedos II y IV en longitud, independientemente del sexo (p> 0,05). Se creó un conjunto de reglas para clasificar los tipos morfológicos de la mano en función de las longitudes de las falanges de los dedos índice y anular mediante la construcción de un árbol de decisión binario. El método CART demuestra la utilidad de este procedimiento estadístico para desarrollar una predicción con base científica del tipo morfológico de la mano. Los resultados de este estudio pueden ser la base de un algoritmo para determinar el tipo morfológico de la mano en función de la longitud de las falanges de los dedos.

Percepción del riesgo de COVID-19 y medidas preventivas en México / Perception of the risk of COVID-19 and preventive measures in Mexico

Introducción: las características de transmisión del virus SARS-CoV-2 incrementaron la necesidad de implementar medidas de prevención a fin de evitar su propagación; por lo tanto, hay un interés en la percepción de riesgo y una intención de la población para adoptar las medidas de protección frente a la COVID-19. Objetivo: evidenciar los factores causales que contribuyen a una percepción alta/baja del riesgo de la pandemia de COVID-19. Material y métodos: se hizo una investigación exploratoria cuantitativa, analítica y explicativa con diseño transversal. Se diseñó una encuesta de autoacceso con el procedimiento de Evaluación Dimensional del Riesgo Percibido, basado en el paradigma psicométrico. Para el análisis de los datos se utilizó la técnica árboles de clasificación. Resultados: el análisis descriptivo indicó que los encuestados tienen una preocupación por la pandemia de 3.8 en una escala de 1 a 5, que evidencia una percepción media-alta del riesgo de la pandemia de COVID-19; sin embargo, subestiman su riesgo personal (vulnerabilidad) en relación con el de los demás. Del análisis con árboles de clasificación, se obtuvo que las variables independientes que contribuyen directamente a la percepción global de riesgo son la gravedad de la pandemia, la vulnerabilidad y el poder catastrófico. Conclusiones: se encontró un efecto positivo: cuanto mayor es la percepción de severidad del virus y la susceptibilidad de contraerlo, hay una mayor toma de medidas preventivas.

Background: The characteristics of the transmission of SARS-CoV-2 virus increased the need to implement prevention measures, in order to avoid its spread; therefore, there is an interest in the risk perception and an intention in people to adopt protective measures against COVID-19. Objective: To show the causal factors that contribute to a high/low perception of risk during the COVID-19 pandemic. Material and methods: A quantitative, analytical and explanatory exploratory study was carried out with a cross-sectional design. To collect data, a self-access survey was designed, using the Dimensional Assessment of Perceived Risk procedure, based on the psychometric paradigm. For data analysis, the classification tree method was utilized. Results: The descriptive analysis indicated that the surveyed population had a concern about the pandemic of 3.8 on a scale ranging from 1 to 5, which shows a medium-high perception of COVID-19 risk; however, respondents underestimated their personal risk (vulnerability) in relation to that of others. The classification tree analysis showed that the independent variables that contribute directly to the global perception of risk are the severity of the pandemic, vulnerability and catastrophic power. Conclusions: It was found a positive effect: The greater perception of severity, and susceptibility to acquiring the virus, the more taking of preventive measures.

Clasificación del estado nutricional basada en perfiles antropométricos del personal silvoagropecuario femenino de un sector del centro-sur de Chile / Classification of nutritional status based on anthropometric profiles of agricultural and forestry female staff silvoagropecuario a sector of south-central Chile

Con el objetivo de identificar las variables antropométricas que mejor explican el estado nutricional del personal femenino del sector silvoa-gropecuario de las regiones del Maule y Bío-Bío en Chile, se determinó su estado nutricional a través de tres indicadores: el índice de masa corporal (IMC), porcentaje de masa grasa obtenido a través de impedancia bioeléctrica y mediciones subcutáneas. La presente investigación propone como hipótesis que una evaluación antropométrica permite determinar las variables que explican la composición corporal, a través de técnicas multivariantes. La identificación de perfiles antropométricos, que permitieron caracterizar la probabilidad de presentar sobrepeso en las trabajadoras, se realizó a través de árboles de clasificación, técnica estadística que se basa en el uso de variables predictoras para particionar la base de datos, en regiones similares, obteniendo grupos altamente homogéneos para una característica en particular. Como resultado se determinó que la población bajo estudio presenta altos porcentajes de obesidad. La metodología de clasificación determinó, para el caso del IMC, que las variables antropométricas que explicaron significativamente la probabilidad de presentar sobrepeso fueron el perímetro de cadera y edad; en cambio, para el caso del porcentaje de masa grasa determinado por medio de impedancia bioeléctrica, fueron perímetro de antebrazo y pliegue tricipital; y a través de pliegues subcutáneos, pliegue submentoniano, la talla y el peso.

In order to identify the anthropometric variables that best explain the nutritional status of female staff of the agriculture and forestry sector of the Maule and Bio Bio in Chile, nutritional status was determined using three indicators: body mass index (BMI ), percentage of fat mass obtained through bioelectrical impedance and subcutaneous measurements. This research proposes the hypothesis that an anthropometric assessment to determine the variables that explain the body composition through multivariate techniques. The identification of anthropometric profiles, which allowed to characterize the probability of overweight women workers, was carried through classification trees, statistical technique based on the use of predictive variables to partition the database, in similar regions, obtaining highly homogeneous groups for a particular feature. As a result it was determined that the study population has high percentages of obesity. The classification methodology determined for the case of BMI, anthropometric variables that significantly explained the probability of overweight were hip circumference and age; however, in the case of fat mass percentage determined by bioelectrical impedance they were forearm circumference and triceps skinfold; and through skinfolds, submental crease, height and weight.

Comparação de métodos para mapeamento digital de solos com utilização de sistema de informação geográfica / Comparison of methods for digital soil mapping using a geographical information system

Mapas pedológicos são fontes de informações primordiais para planejamento e manejo do uso do solo, porém apresentam altos custos de produção. A fim de produzir mapas de solos a partir de mapas existentes, neste trabalho, foram comparados métodos de classificação em estágio único (Regressões Logísticas Múltiplas Multinomiais e Bayes) e em estágios múltiplos (Classification and Regression Trees (CART), J48 e Logistic Model Trees (LMT)) com a utilização de sistemas de informações geográficas e de variáveis geomorfométricas para produção de mapas pedológicos com legenda original e simplificada. A base de dados foi gerenciada em aplicativo computacional ArcGis, em que as variáveis e o mapa original foram relacionados por meio de amostras de treinamento para os algoritmos. Os resultados dos algoritmos obtidos no software Weka foram implementados no ArcGis, para a confecção dos mapas. Foram geradas matrizes de erros para análise de acurácias dos mapas. As variáveis geomorfométricas de declividade, perfil e plano de curvatura, elevação e índice de umidade topográfica são aquelas que melhor explicam a distribuição espacial das classes de solo. Os métodos de classificação em estágio múltiplo apresentaram sensíveis melhoras nas acurácias globais, porém significativas melhoras nos índices Kappa. A utilização de legenda simplificada aumentou significativamente as acurácias do produtor e do usuário.

Soil maps are sources of important information for land planning and management, but are expensive to produce. This paper proposes testing and comparing single stage classification methods (Multiple Multinomial Logistic Regression and Bayes) and multiple stage classification methods (Classification and Regression Trees (CART), J48 and Logistic Model Trees (LMT)) using geographic information system and terrain parameters for producing soil maps with both original and simplified legend. The database was managed in ArcGis computer application in which the variables and the original map were related through training of the algorithms. The results from statistical software Weka were implemented in ArcGis environment to generate digital soil maps. The terrain parameters that best explained soil distribution were slope, profile and planar curvature, elevation, and topographic wetness index. The multiple stage classification methods showed small improvements in overall accuracies and large improvements in the Kappa index. Simplification of the original legend significantly increased the producer and user accuracies, however produced small improvements in overall accuracies and Kappa index.

Effectiveness of Repeated Examination to Diagnose Enterobiasis in Nursery School Groups

The aim of this study was to estimate the benefit from repeated examinations in the diagnosis of enterobiasis in nursery school groups, and to test the effectiveness of individual-based risk predictions using different methods. A total of 604 children were examined using double, and 96 using triple, anal swab examinations. The questionnaires for parents, structured observations, and interviews with supervisors were used to identify factors of possible infection risk. In order to model the risk of enterobiasis at individual level, a similarity-based machine learning and prediction software Constud was compared with data mining methods in the Statistica 8 Data Miner software package. Prevalence according to a single examination was 22.5%; the increase as a result of double examinations was 8.2%. Single swabs resulted in an estimated prevalence of 20.1% among children examined 3 times; double swabs increased this by 10.1%, and triple swabs by 7.3%. Random forest classification, boosting classification trees, and Constud correctly predicted about 2/3 of the results of the second examination. Constud estimated a mean prevalence of 31.5% in groups. Constud was able to yield the highest overall fit of individual-based predictions while boosting classification tree and random forest models were more effective in recognizing Enterobius positive persons. As a rule, the actual prevalence of enterobiasis is higher than indicated by a single examination. We suggest using either the values of the mean increase in prevalence after double examinations compared to single examinations or group estimations deduced from individual-level modelled risk predictions.

Árboles de regresión y otras opciones metodológicas aplicadas a la predicción del rendimiento académico / Regression trees and other methodological options applied to the prediction of academic performance

El presente trabajo se realizó con el fin de construir un algoritmo para detectar estudiantes con alto riesgo de fracaso académico e identificar los mejores predictores del rendimiento. Se caracterizaron los estudiantes que ingresaron en el primer año en el ICBP "Victoria de Girón" durante el curso 2001-2002 de acuerdo con su índice académico del preuniversitario, índice escalafonario, exámenes de ingreso, prueba de inteligencia y un indicador de su motivación profesional. Se emplearon árboles de clasificación para identificar los predictores relevantes y sus puntos de corte óptimos. Se utilizó un modelo de regresión ordinal para evaluar la importancia relativa de los predictores y proponer el algoritmo de predicción. A partir del índice escalafonario, exclusivamente, se obtuvo un procedimiento de clasificación, que permitió identificar a los estudiantes de mayor riesgo de fracaso académico. Los puntos de corte fueron 87 y 91 puntos, que definen una tricotomía para el pronóstico del rendimiento

This paper is aimed at constructing an algorithm to detect students at high risk for academic failure and at identifying the best preformance predictors. The students that were admitted in the first year at Victoria de Girón Institute of Preclinical Basic Sciences during the course 2001-2002 were characterized according to their preuniversity academic index, roster index, admission test, intelligence test and an indicator of their professional motivation. Classification trees were used to identify the relevant predictors and their optimal cut-off points. A model of ordinal regression was used to evaluate the relative importance of the predictors and to propose the prediction algorithm.Starting only from the roster index, it was obtained a classification procedure that allowed to identify students at the highest risk for academic failure. The cut-offs were 87 and 91 points, which define a trichotomy for the performance prognosis.