Evaluating performance and determining optimum sample size for regression tree and automatic linear modeling / [Avaliando o desempenho e determinando o tamanho ideal da amostra para árvore de regressão e modelagem linear automática]

This study was carried out for two purposes: comparing performances of Regression Tree and Automatic Linear Modeling and determining optimum sample size for these methods under different experimental conditions. A comprehensive Monte Carlo Simulation Study was designed for these purposes. Results of simulation study showed that percentage of explained variation estimates of both Regression Tree and Automatic Linear Modeling was influenced by sample size, number of variables, and structure of variance-covariance matrix. Automatic Linear Modeling had higher performance than Regression Tree under all experimental conditions. It was concluded that the Regression Tree required much larger samples to make stable estimates when comparing to Automatic Linear Modeling.(AU)

Este estudo foi realizado com dois objetivos: comparar os desempenhos da Árvore de Regressão e da Modelagem Linear Automática e determinar o tamanho ideal da amostra para estes métodos sob diferentes condições experimentais. Um abrangente Estudo de Simulação de Monte Carlo foi projetado para estes propósitos. Os resultados do estudo de simulação mostraram que a porcentagem de estimativas de variação explicada tanto da Árvore de Regressão como da Modelagem Linear Automática foi influenciada pelo tamanho da amostra, número de variáveis e estrutura da matriz de variância-covariância. A Modelagem Linear Automática teve um desempenho superior ao da Árvore de Regressão em todas as condições experimentais. Concluiu-se que a Árvore de Regressão exigia amostras muito maiores para fazer estimativas estáveis quando comparada à Modelagem Linear Automática.(AU)

Predictors of Students' Mathematics Achievement in Secondary Education / Preditores do Desempenho em Matemática de Estudantes do Ensino Médio

Abstract Acknowledging the relevance of mathematics education, as well the evidence about predictors related to achievement in this domain, the present study performed a predictive analysis of students' mathematics achievement in the National Exam for Secondary Education, employing the Regression Tree Method and a model with 53 predictors. Results indicated that the model explained 29.97% of the mathematics achievement variance. Certain variables are related to worse achievement in mathematics: Students' family monthly income equal or smaller than 2 minimum wages, be female, have not attended Primary and Secondary Education in private schools, live in North, North East and Center West regions of Brazil, be highly motivated to perform the exam to obtain Secondary Education certificate or scholarship. The results obtained highlight the role of variables related to the individual, school and family as predictors of mathematics achievement.

Resumo Considerando a relevância da formação em matemática, assim como a evidência de preditores relacionados ao desempenho nesse domínio, realizou-se no presente estudo uma análise preditiva do desempenho matemático de inscritos no Exame Nacional do Ensino Médio de 2011, empregando a abordagem de Regressão em Árvore e um modelo com 53 preditores. Os resultados indicam que o modelo explicou 29,97% da variância do desempenho em matemática na amostra teste. Determinadas variáveis relacionam-se a um pior rendimento em matemática: renda familiar de até dois salários mínimos, sexo feminino, não ter cursado escolas particulares no ensino fundamental e no ensino médio, residir nas regiões Norte, Nordeste e Centro-Oeste, e estar altamente motivado para fazer o Exame para obter certificação ou bolsa de estudos. Os resultados obtidos salientam o papel de variáveis relacionadas ao indivíduo, à escola e à família como preditoras do desempenho em matemática.

Spatial distribution combined cold and hotspots of China's healthy adult male serum CA19-9 reference value based on classification and regression tree / 西安交通大学学报(医学版)

Objective: To provide clinical criteria with regional basis of the reference values of carbohydrateantigen 19-9 (CA19-9) in healthy Chinese male adults. Methods: First, the reference values of CA19-9 in 9,382 healthy male adults distributed in 64 cities and counties in China were collected to construct the database. Then geographical factors received dimensionality reduction by comprehensive application of the principal component analysis, variation coefficient method, and correlation analysis; then the classification regression tree model was constructed. Third, the hotspot and spatial distribution of the reference values of CA19-9 in healthy male adults was drawn based on 2,322 observation data points. Results: The reference value of CA19-9 showed spatial variation, which was higher in the southern region than in the northern region, and higher in Jilin and Liaoning provinces than in the eastern central region. Hotspots were mostly distributed in the southern region, and cold spots in the northern region. Conclusion: There exist spatial variation of the reference values of CA19-9 at spatial level. Therefore, when it comes to clinical diagnosis, it is necessary to consider regional variation.

Study on Grouping Method of Score in Case-based Score Payment / 中国卫生经济

Objective:To explore the grouping methods of the case-based score from thinking of medical expenses impacted by the secondary diagnosis.Methods:Classification and Regression Trees(CART) was used to group the non-surgical cerebral infarction diseases in the discharge summary page and the results were evaluated with CV and non-parametric test (Nemenyi test).Results:The non-surgical cerebral infarction diseases could be divided into 3 groups,the P values were below 0.05 for all groups analyzed with the non-parametric test.Conclusion:It was feasible to use CART to group the non-surgical cerebral infarction diseases.

Hepatotoxicity Study of Voriconazole Therapy in Patients with Invasive Fungal Infections / 中国药学杂志

OBJECTIVE: To explore the factors that association with voriconazole-related hepatotoxicity in patients with invasive fungal infections. METHODS: Voriconazole trough plasma concentrations(ρmin) were measured by high-performance liquid chromatography(HPLC). Voriconazole-related hepatotoxicity was defined according to Common Terminology Criteria for Adverse Events(CTCAE), and the multivariate linear regression and classification and regression tree(CART) were used to explore the factors that association with hepatotoxicity. RESULTS: A total of 328 samples from 144 patients were measured, and 18 patients with signs of hepatotoxicity. There was a significant difference between the voriconazole ρmin in patients with hepatotoxicity and those without hepatotoxicity [(3.49 ± 2.31) μg•mL-1 vs (1.96 ± 1.48 ) μg•mL-1, P 2.89 μg•mL-1; ③non-poor metabolizer with voriconazole ρmin > 2.89 μg•mL-1 who received CYP2C19 inhibitor. CONCLUSION: Monitoring voriconazole ρmin and liver function could benefit to the safety of voriconazole therapy in patients with invasive fungal infections.

Measurement and Modeling of Job Stress of Electric Overhead Traveling Crane Operators

BACKGROUND: In this study, the measurement of job stress of electric overhead traveling crane operators and quantification of the effects of operator and workplace characteristics on job stress were assessed. METHODS: Job stress was measured on five subscales: employee empowerment, role overload, role ambiguity, rule violation, and job hazard. The characteristics of the operators that were studied were age, experience, body weight, and body height. The workplace characteristics considered were hours of exposure, cabin type, cabin feature, and crane height. The proposed methodology included administration of a questionnaire survey to 76 electric overhead traveling crane operators followed by analysis using analysis of variance and a classification and regression tree. RESULTS: The key findings were: (1) the five subscales can be used to measure job stress; (2) employee empowerment was the most significant factor followed by the role overload; (3) workplace characteristics contributed more towards job stress than operator's characteristics; and (4) of the workplace characteristics, crane height was the major contributor CONCLUSION: The issues related to crane height and cabin feature can be fixed by providing engineering or foolproof solutions than relying on interventions related to the demographic factors.

Initial Rotterdam CT score may predict prognosis of the patients after unilateral decompression craniectomy for treating severe traumatic brain injury / 中华急诊医学杂志

Objective To identify the factors enhancing the contusive brain hemorrhage following unilateral decompression craniectomy in patients with severe traumatic brain injury (TBI),and to explore the relationship between the initial Rotterdam CT score and clinical outcomes.Methods A prospective study of 291 consecutive patients with TBI admitted from Jan 2008 through Dec 2012 was carried out.Patients treated with unilateral decompression craniectomy were enrolled for study.Patients without preoperative or postoperative cranial CT imaging were excluded.Of them,235 patients were followed up.Gender,age,the causes of injury,preoperative general condition including Glasgow Coma Scale (GCS) score,pupillary response,laboratory data and the initial CT scans before operation,contusion hematoma size in CT scans following operation and Glasgow Outcome Scale (GOS) score were recorded.With t test,x2 test and nonparametric rank sum test,differences in the above listed variables were compared between patients with enlarged hematoma size group and those without change in hematoma size.A Classification And Regression Tree (CART) was used to predict the size of hematoma.Correlation analysis was used to find the relationship between the Rotterdam CT scores and GOS scores.Results The differences in age (t =2.034,P =0.043),first Rotterdam CT score (Z =4.838,P ＜ 0.01),GCS score (Z =4.440,P ＜ 0.01),pupillary response (Z =3.235,P =0.001),the length of time elapsed between the trauma occurred and the decompressive craniectomy (Z =3.874,P ＜ 0.01),glucose level (Z =3.880,P ＜ 0.01) and cerebrum hernia magnitude (Z =2.529,P =0.012) were significant between the patients with hematoma expanded (n =120) and those without change in hematoma size (n =115).The results of the CART indicated that Rotterdam score got from the initial head CT,glucose level and the length of time elapsed between trauma occurred and decompressive craniectomy were strong predictors of the risk for expanded hemorrhagic contusions following decompressive craniectomy.Both age and size of the removed bone-flap also could predict the risk of postoperative expansion of hemorrhagic contusions.The overall predictive accuracy of the CART model was 83.3％.Correlation analysis results indicated that Rotterdam CT score was negatively correlated with GOS (r =-0.333,P ＜ 0.01).Conclusions Initial Rotterdam CT scores,glucose level and the length of time between trauma and decompressive craniectomy may predict the risk of contusions expansion following decompressive craniectomy.Rotterdam CT score was negatively correlated with GOS.

The predictability of dentoskeletal factors for soft-tissue chin strain during lip closure / 대한치과교정학회지

OBJECTIVE: To investigate the dentoskeletal factors which may predict soft-tissue chin strain during lip closure. METHODS: The pretreatment frontal and lateral facial photographs and lateral cephalograms of 209 women (aged 18-30 years) with Angle's Class I or II malocclusion were examined. The subjects were categorized by three examiners into the no-strain and strain groups according to the soft-tissue chin tension or deformation during lip closure. Relationships of the cephalometric measurements with the group classification were analyzed by logistic regression analysis, and a classification and regression tree (CART) model was used to define the predictive variables for the group classification. RESULTS: The lower the value of the overbite depth indicator (ODI) and the higher the values of upper incisor to Nasion-Pogonion (U1-NPog, mm), overjet, and upper incisor to upper lip (U1-upper lip, mm), the more likely was the subject to be classified into the strain group. The CART showed that U1-NPog was the most prominent predictor of soft-tissue chin strain (cut-off value of 14.2 mm), followed by overjet. CONCLUSIONS: To minimize strain of the soft-tissue chin, orthodontic treatment should be oriented toward increasing the ODI value while decreasing the U1-NPog, overjet, and U1 upper lip values.

Prediction of large esophageal varices in cirrhotic patients using classification and regression tree analysis

OBJECTIVES: Recent guidelines recommend that all cirrhotic patients should undergo endoscopic screening for esophageal varices. That identifying cirrhotic patients with esophageal varices by noninvasive predictors would allow for the restriction of the performance of endoscopy to patients with a high risk of having varices. This study aimed to develop a decision model based on classification and regression tree analysis for the prediction of large esophageal varices in cirrhotic patients. METHODS: 309 cirrhotic patients (training sample, 187 patients; test sample 122 patients) were included. Within the training sample, the classification and regression tree analysis was used to identify predictors and prediction model of large esophageal varices. The prediction model was then further evaluated in the test sample and different Child-Pugh classes. RESULTS: The prevalence of large esophageal varices in cirrhotic patients was 50.8 percent. A tree model that was consisted of spleen width, portal vein diameter and prothrombin time was developed by classification and regression tree analysis achieved a diagnostic accuracy of 84 percent for prediction of large esophageal varices. When reconstructed into two groups, the rate of varices was 83.2 percent for high-risk group and 15.2 percent for low-risk group. Accuracy of the tree model was maintained in the test sample and different Child-Pugh classes. CONCLUSIONS: A decision tree model that consists of spleen width, portal vein diameter and prothrombin time may be useful for prediction of large esophageal varices in cirrhotic patients.