• Title, Summary, Keyword: Test of heterogeneity

Search Result 155, Processing Time 0.05 seconds

Regression Analysis of Longitudinal Data Based on M-estimates

  • Jung, Sin-Ho;Terry M. Therneau
    • Journal of the Korean Statistical Society
    • /
    • v.29 no.2
    • /
    • pp.201-217
    • /
    • 2000
  • The method of generalized estimating equations (GEE) has become very popular for the analysis of longitudinal data. We extend this work to the use of M-estimators; the resultant regression estimates are robust to heavy tailed errors and to outliers. The proposed method does not require correct specification of the dependence structure between observation, and allows for heterogeneity of the error. However, an estimate of the dependence structure may be incorporated, and if it is correct this guarantees a higher efficiency for the regression estimators. A goodness-of-fit test for checking the adequacy of the assumed M-estimation regression model is also provided. Simulation studies are conducted to show the finite-sample performance of the new methods. The proposed methods are applied to a real-life data set.

  • PDF

Nonparametric analysis of income distributions among different regions based on energy distance with applications to China Health and Nutrition Survey data

  • Ma, Zhihua;Xue, Yishu;Hu, Guanyu
    • Communications for Statistical Applications and Methods
    • /
    • v.26 no.1
    • /
    • pp.57-67
    • /
    • 2019
  • Income distribution is a major concern in economic theory. In regional economics, it is often of interest to compare income distributions in different regions. Traditional methods often compare the income inequality of different regions by assuming parametric forms of the income distributions, or using summary statistics like the Gini coefficient. In this paper, we propose a nonparametric procedure to test for heterogeneity in income distributions among different regions, and a K-means clustering procedure for clustering income distributions based on energy distance. In simulation studies, it is shown that the energy distance based method has competitive results with other common methods in hypothesis testing, and the energy distance based clustering method performs well in the clustering problem. The proposed approaches are applied in analyzing data from China Health and Nutrition Survey 2011. The results indicate that there are significant differences among income distributions of the 12 provinces in the dataset. After applying a 4-means clustering algorithm, we obtained the clustering results of the income distributions in the 12 provinces.

Material structure generation of concrete and its further usage in numerical simulations

  • Husek, Martin;Kala, Jiri
    • Structural Engineering and Mechanics
    • /
    • v.68 no.3
    • /
    • pp.335-344
    • /
    • 2018
  • The execution of an experiment is a complex affair. It includes the preparation of test specimens, the measurement process itself and also the evaluation of the experiment as such. Financial requirements can differ significantly. In contrast, the cost of numerical simulations can be negligible, but what is the credibility of a simulated experiment? Discussions frequently arise concerning the methodology used in simulations, and particularly over the geometric model used. Simplification, rounding or the complete omission of details are frequent reasons for differences that occur between simulation results and the results of executed experiments. However, the creation of a very complex geometry, perhaps all the way down to the resolution of the very structure of the material, can be complicated. The subject of the article is therefore a means of creating the material structure of concrete contained in a test specimen. Because a complex approach is taken right from the very start of the numerical simulation, maximum agreement with experimental results can be achieved. With regard to the automation of the process described, countless material structures can be generated and randomly produced samples simulated in this way. Subsequently, a certain degree of randomness can be observed in the results obtained, e.g., the shape of the failure - just as is the case with experiments. The first part of the article presents a description of a complex approach to the creation of a geometry representing real concrete test specimens. The second part presents a practical application in which the numerical simulation of the compressive testing of concrete is executed using the generated geometry.

An Experimental Study of the Effect of the Test-well Arrangement on the Partitioning Interwell Tracer Test for the Estimation of the NAPL Saturation (지하수 유동 방향에 대한 관정배열이 분배추적자 시험에 미치는 영향 분석)

  • Kim, Bo-A;Kim, Yongcheol;Yeo, In Wook;Ko, Kyung-Seok
    • Journal of Soil and Groundwater Environment
    • /
    • v.19 no.3
    • /
    • pp.111-122
    • /
    • 2014
  • Partitioning interwell tracer test (PITT) is a method to quantify and qualify a site contaminated with NAPLs (Non-Aqueous Phase Liquids). Analytical description of PITT assumes that the injection-pumping well pair is on the line of the ambient groundwater flow direction, but the test-well pair could frequently be off the line in a real field site, which could be an erroneous factor in analyzing PITT data. The purpose of this work is to study the influence of the angle of the test-well pair on the ambient groundwater flow direction based on the result from PITT. From the experiments, it was found that the obliqueness of the test-well pair to the ambient groundwater flow direction could affect the tracer test resulting in a decreased NAPL estimation efficiency. In case of an oblique arrangement of the test-well pair to the ambient flow direction, it was found that the injection of a chase fluid could enhance the estimation efficiency. An increase of the pumping rate could enhance the recovery rate but it cannot be said that a high pumping rate can increase the test efficiency because a high pumping rate cannot give partitioning tracers enough time to partition into NAPLs. The results have a implication that because the arrangement of the test-well pair is a controlling factor in performing and interpreting PITT in the field in addition to the known factors such as heterogeneity and the source zone architecture, flow direction should be seriously considered in arranging test-well pair.

Simulation of fracture in plain concrete modeled as a composite material

  • Bui, Thanh T.;Attard, Mario M.
    • Computers and Concrete
    • /
    • v.2 no.6
    • /
    • pp.499-516
    • /
    • 2005
  • A composite model is used to represent the heterogeneity of plain concrete consisting of coarse aggregates, mortar matrix and the mortar-aggregate interface. The composite elements of plain concrete are modeled using triangular finite element units which have six interface nodes along the sides. Fracture is captured through a constitutive single branch softening-fracture law at the interface nodes, which bounds the elastic domain inside each triangular unit. The inelastic displacement at an interface node represents the crack opening or sliding displacement and is conjugate to the internodal force. The path-dependent softening behaviour is developed within a quasi-prescribed displacement control formulation. The crack profile is restricted to the interface boundaries of the defined mesh. No re-meshing is carried out. Solutions to the rate formulation are obtained using a mathematical programming procedure in the form of a linear complementary problem. An event by event solution strategy is adopted to eliminate solutions with simultaneous formation of softening zones in symmetric problems. The composite plain concrete model is compared to experimental results for the tensile crack growth in a Brazilian test and three-point bending tests on different sized specimens. The model is also used to simulate wedge-type shear-compression failure directly under the loading platen of a Brazilian test.

Obesity and Risk of Bladder Cancer: A Meta-analysis of Cohort Studies

  • Qin, Qi;Xu, Xin;Wang, Xiao;Zheng, Xiang-Yi
    • Asian Pacific Journal of Cancer Prevention
    • /
    • v.14 no.5
    • /
    • pp.3117-3121
    • /
    • 2013
  • Objective: Previous epidemiologic studies demonstrated that obesity might associated with the risk of bladder cancer. However, many of the actual association findings remained conflicting. To better clarify and provide a comprehensive summary of the correlation between obesity and bladder cancer risk, we conducted a meta-analysis to summarize results of studies on the issue. Stratified analyses were also performed on potential variables and characteristics. Methods: Studies were identified by searching in PubMed and Wanfang databases, covering all the papers published from their inception to March 10, 2013. Summary relative risks (SRRs) with their corresponding 95% confidence intervals (CIs) were calculated by either random-effect or fixed-effect models. Results: A total of 11 cohort studies were included in our meta-analysis, which showed that obesity was associated with an increased risk for bladder cancer in all subjects (RR=1.10, 95% CI=1.06-1.16; p=0.215 for heterogeneity; $I^2$=24.0%). Among the 9 studies that controlled for cigarette smoking, the pooled RR was 1.09 (95% CI 1.01-1.17; p=0.131 for heterogeneity; $I^2$=35.9%). No significant publication bias was detected (p = 0.244 for Egger's regression asymmetry test). Conclusions: Our results support the conclusion that obesity is associated with the increased risk of bladder cancer. Further research is needed to generate a better understanding of the correlation and to provide more convincing evidence for clinical intervention in the prevention of bladder cancer.

Meta Analysis of Symptom Improvement through Eradication of Helicobacter pylori in Patients with Non-ulcer Dyspepsia (비궤양성 소화불량에서 Helicobacter pylori 박멸치료후 증상개선에 대한 메타분석)

  • Ohm, Sang-Hwa;Jeong, Ki-Won;Shin, Won-Chang;Cho, Jong-Rae;Shon, Hye-Suk;Pae, Ki-Taek;Kim, Sung-Jun
    • Journal of Preventive Medicine and Public Health
    • /
    • v.32 no.4
    • /
    • pp.427-434
    • /
    • 1999
  • Objectives: This study was conducted to determine, by reviewing the literature, whether treatment of Helicobacter pylori infection in patients with non-ulcer dyspepsia affects symptoms. Methods: We retrieved the literature using MEDLINE search, with nonulcer dyspepsia and Hericobacter pylori and treatment as key words, which were reported from 1984 to 1998, and manual literature search. The criteria for inclusion was as follows; 1) The paper should have confirmed nonulcer dyspepsia as case definition. 2) The paper should have peformed a randomized, blind trial. 3) Confirmation of Helicobacter pylori eradication should be done 4 weeks after treatment. 4) studies with no information on measurement of symptoms after treatment were not accepted. The percentage of patients with symptom improvement after eradication therapy for Helicobacter pylori infection was calculated. Cumulative odds ratio was compared by fixed effect model and random effect model as sensitivity and funnel plot was used to evaluate publication bias. Results: The overall effect size of symptom improvement was calculated by cumulative odds ratio. Cumulative odds ratio of random effect model was 4.16(95% CI: 1.55-11.19). Before integrating each effect sizes into common effect size, the homogeneity test was conducted and random effect model was selected(Cochran's Q=41.08 (d.f=10, p<0.001)). The heterogeneity across studies was evaluated and the different methodological aspects of studies led to differences between study results Conclusions: The results suggest that the eradication of Helicobacter pylori in patients with non-ulcer dyspepsia results more symptom improvement. In studios that shows the opposite results there are methodological aspects explaining the heterogeneity.

  • PDF

Glutathione-S-Transferase Polymorphisms (GSTM1, GSTT1 and GSTP1) and Acute Leukemia Risk in Asians: a Meta-analysis

  • Tang, Zhen-Hai;Zhang, Chi;Cheng, Pan;Sun, Hong-Min;Jin, Yu;Chen, Yuan-Jing;Huang, Fen
    • Asian Pacific Journal of Cancer Prevention
    • /
    • v.15 no.5
    • /
    • pp.2075-2081
    • /
    • 2014
  • The association between glutathione-S-transferase polymorphisms (GSTM1, GSTT1 and GSTP1) and risk of acute leukemia in Asians remains controversial. This study was therefore designed to evaluate the precise association in 23 studies identified by a search of PubMed and several other databases, up to December 2013. Using random or fixed effects models odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were calculated. Heterogeneity across studies was assessed, and funnel plots were constructed to test for publication bias. The meta-analysis showed positive associations between GST polymorphisms (GSTM1 and GSTT1 but not GSTP1) and acute leukemia risk [(OR=1.47, 95% CI 1.18-1.83); (OR=1.32, 95% CI 1.07-1.62); (OR=1.01, 95% CI 0.84-1.23), respectively] and heterogeneity between the studies. The results suggested that the GSTM1 null genotype and GSTT1null genotype, but not the GSTP1 polymorphism, might be a potential risk factors for acute leukemia. Further well-designed studies are needed to confirm our findings.

Estimate of the Basic Reproduction Number for COVID-19: A Systematic Review and Meta-analysis

  • Alimohamadi, Yousef;Taghdir, Maryam;Sepandi, Mojtaba
    • Journal of Preventive Medicine and Public Health
    • /
    • v.53 no.3
    • /
    • pp.151-157
    • /
    • 2020
  • Objectives: The outbreak of coronavirus disease 2019 (COVID-19) is one of the main public health challenges currently facing the world. Because of its high transmissibility, COVID-19 has already caused extensive morbidity and mortality in many countries throughout the world. An accurate estimation of the basic reproduction number (R0) of COVID-19 would be beneficial for prevention programs. In light of discrepancies in original research on this issue, this systematic review and meta-analysis aimed to estimate the pooled R0 for COVID-19 in the current outbreak. Methods: International databases (including Google Scholar, Science Direct, PubMed, and Scopus) were searched to identify studies conducted regarding the R0 of COVID-19. Articles were searched using the following keywords: "COVID-19" and "basic reproduction number" or "R0." The heterogeneity among studies was assessed using the I2 index, the Cochran Q test, and T2. A random-effects model was used to estimate R0 in this study. Results: The mean reported R0 in the identified articles was 3.38±1.40, with a range of 1.90 to 6.49. According to the results of the random-effects model, the pooled R0 for COVID-19 was estimated as 3.32 (95% confidence interval, 2.81 to 3.82). According to the results of the meta-regression analysis, the type of model used to estimate R0 did not have a significant effect on heterogeneity among studies (p=0.81). Conclusions: Considering the estimated R0 for COVID-19, reducing the number of contacts within the population is a necessary step to control the epidemic. The estimated overall R0 was higher than the World Health Organization estimate.

Vitamin D and fibromyalgia: a meta-analysis

  • Makrani, Atekeh Hadinezhad;Afshari, Mahdi;Ghajar, Marayam;Forooghi, Zahra;Moosazadeh, Mahmood
    • The Korean Journal of Pain
    • /
    • v.30 no.4
    • /
    • pp.250-257
    • /
    • 2017
  • Vitamin D is a cofactor responsible for autoimmune disorders. There is no agreement in the studies investigating the association between vitamin D and fibromyalgia. This study aims to combine the conflicting results of the primary studies which compared these patients with control groups regarding the serum concentration of vitamin D. This meta-analysis has been designed based on PRISMA guidelines. Relevant keywords were searched in PubMed, Science direct, Scopus, Cochrane, and Google scholar and primary studies were selected. After screening the eligible studies according to inclusion/exclusion criteria, we investigated the risk of bias in the selected studies and also the heterogeneity between the primary results using Cochrane (Q) and I-squared ($I^2$) indices. The primary results were combined using inverse variance method and Cohen statistics as well as a random effects model. Publication bias was assessed using Egger test. Sensitivity analysis was applied to investigate the influence of each primary study on the final result of the meta-analysis. Suspected factors in the heterogeneity were assessed using meta-regression models. We entered 12 eligible studies in the meta-analysis including 851 cases compared with 862 controls. The standardized mean difference of Vitamin D between the two groups was -0.56 (95% confidence interval: -1.05, -0.08). Our meta-analysis showed that vitamin D serum levels of patients with fibromyalgia was significantly lower than that of control group.