Explore chapters and articles related to this topic
Working with categorical outcome variables
Published in Ewen Harrison, Pius Riinu, R for Health Data Science, 2020
Pearson’s chi-squared () test of independence is used to determine whether two categorical variables are independent in a given population. Independence here means that the relative frequencies of one variable are the same over all levels of another variable.
Model Fit
Published in Peter Cummings, Analysis of Incidence Rates, 2019
Andrews (Andrews 1988a, 1988b) described a modification of Pearson’s chi-squared test which accounts for the uncertainty of the model estimates. This test has been described in textbooks (Cameron and Trivedi 2005 pp266–267, 2013a pp193–196) and implemented in a Stata command for count models, chi2gof (Manjón and Martínez 2014). The command allows the user to categorize the outcome counts into strata and then test the null hypothesis that the frequency distribution of the observed counts matches the frequency distribution of the model predicted counts. It is best to use strata in which the counts are not too small. This command can be used after Stata’s poisson and nbreg (negative binomial regression) commands. When the command is used after the Poisson regression model for the fall count data, and if cell counts of 0, 1, 2, … , 7, and 8 or more are chosen as the strata, the output looks like this: . chi2gof, cells(9) table Chi-square Goodness-of-Fit Test for Poisson Model: Chi-square chi2(8) = 124.40 Prob>chi2 = 0.00 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Fitted Cells Abs. Freq. Rel. Freq. Rel. Freq. Abs. Dif. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 0 199 .4393 .2166 .2227 1 115 .2539 .3262 .0724 2 66 .1457 .2507 .105 3 24 .053 .131 .078 4 15 .0331 .0524 .0192 5 10 .0221 .017 .005 6 6 .0132 .0047 .0085 7 3 .0066 .0011 .0055 8 or more 13 .0331 2.9e-04 .0328 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
LASSO-based machine learning algorithm to predict the incidence of diabetes in different stages
Published in The Aging Male, 2023
Qianying Ou, Wei Jin, Leweihua Lin, Danhong Lin, Kaining Chen, Huibiao Quan
Statistical analyses were performed using R software version 4.0.2 and Statistical Product Service Solutions (SPSS) version 25.0 (SPSS, Chicago, IL). Continuous variables were described as mean ± standard deviation (SD). Categorical data were presented as number (percentage). Pearson’s Chi-squared test was used for comparison of categorical variables. Mann–Whitney U test was used to evaluate the significant differences of continuous variables between groups. The odds ratio (OR) values and 95% confidence intervals (CIs) were used to express the independent predictive ability of the predictors. A two-sided p value .05 was considered statistically significant. The “glmnet’” (4.1-2) and “ggplot2’” packages in R were used for LASSO logistic regression. The “mlbench” and caret packages were used for logistic regression, SVM, RF, LDA, NB, and Treebag analysis. The “rms” package was used for nomogram and calibration plots. ROC curve analyses were conducted using the “pROC” packages. In addition, the nomogram was subjected to 1000 bootstrap resamples for internal validation to assess their predictive accuracies. Correlation analysis among different predictive factors was performed by “corrplot” package.
Dietary Intake is Associated with miR-31 and miR-375 Expression in Patients with Head and Neck Squamous Cell Carcinoma
Published in Nutrition and Cancer, 2022
Tathiany Jéssica Ferreira, Caroline Castro de Araújo, Ana Carolina da Silva Lima, Larissa Morinaga Matida, Ana Flávia Mendes Griebeler, Alexandre Siqueira Guedes Coelho, Antônio Paulo Machado Gontijo, Cristiane Cominetti, Eneida Franco Vêncio, Maria Aderuza Horst
Pearson’s chi-squared test was used to compare the sample characterization data. Analysis of variance (one-way ANOVA) was used to compare the mean food consumption and the expression of miR-31 and miR-375 in patients according to the different anatomical sites (oral cavity, oropharynx, hypopharynx/larynx groups). For the association between food consumption and the expression of miR-31 and miR-375, a stepwise multiple logistic regression model was adjusted for age, tumor site, cancer stage, smoking and drinking status, and energy intake. The critical level of significance adopted was 5%. All analyses were performed in R software, version R-3.4.3. The power of the test was estimated a posteriori and was calculated based on an average effect size of 0.15, for a sample of 67 individuals. The calculations showed that with a significance level of 5%, the statistical power is equivalent to 88%. The software used was G-Power® (version 3.1.9.2).
The effect of the revised WHO classification on the incidence of grade II meningioma
Published in British Journal of Neurosurgery, 2020
Lindsey S. Bulleid, Zoe James, Alistair Lammie, Caroline Hayhurst, Paul A. Leach
Using retrospective data from the University Hospital of Wales (UHW) neuro-oncology multi-disciplinary team meetings over a forty-month period all adult and paediatric patients with a potential diagnosis of meningioma were identified. We also used the Institutional online reporting system Welsh Clinical Portal to identify patients who had undergone surgery within a three-year period (January 1st 2015 until December 31st 2017) and reviewed histological reports. Only patients with cranial meningioma were included. One patient was later reported to have had a haemangiopericytoma and was excluded. Neuropathologists at UHW had followed the 2007 WHO Classification of Tumours of the Central Nervous System but classified brain invasion as a stand-alone criterion for grade II meningioma, anticipating the 2016 revision of the grading criteria. We compared incidence rates in our cohort with other reported incidence rates. Statistical analysis was performed using Pearson Chi Squared test.