China
Africa
Explore chapters and articles related to this topic
Determinants of health complaints of Bodetabek commuter workers using Bayesian multilevel logistic regression
Published in Yuli Rahmawati, Peter Charles Taylor, Empowering Science and Mathematics for Global Competitiveness, 2019
Based on Table 2, the DIC for a one-level null-model binary logistic regression is 2,530.58. A null model or empty model is a model that does not contain explanatory variables. In the first step, by comparing the DIC of a one-level null model with a two-level null model, we can conclude that DIC has decreased from 2,530.58 to 2,394.49. It indicates that the two-level regression model with random effects is more appropriate to be used. In the next step, we compare the DIC null two-level binary logistic regression model with the two-level DIC conditional model. A conditional model is a model that includes all explanatory variables. Based on Table 2, the two-level null-model DIC is greater than the two-level DIC conditional model. Thus, it can be concluded that a two-level binary logistic regression model containing all explanatory variables is more appropriate to be used to analyze health complaints experienced by Bodetabek commuter workers.
Analysis of the worst-case scenarios in an elite football team: Towards a better understanding and application
Published in Journal of Sports Sciences, 2021
Andrew R. Novak, Franco M. Impellizzeri, Arjav Trivedi, Aaron J. Coutts, Alan McCall
All independent variables were included within all models. Multicollinearity was investigated for each model prior to analysis, and where two variables were correlated above 0.8, the variable with a lower correlation with the WCS was removed from the model. Positional group was a nominal fixed effect. Therefore, interpretations of positional group are provided relative to the reference value of Central Midfielders (the group with the most observations). We calculated coefficient estimates and 95% confidence intervals (CI) for each independent variable in each WCS model. To assess model fit, the models were each compared against a null model (with only random effects), using the Akaike Information Criterion (AIC). Model residuals were inspected for deviation from normality, and Cook’s distance was calculated to investigate influential points (cut-off = 1). We calculated marginal and conditional R2 values to assess the explained variance excluding and including random effects.
External and internal maximal intensity periods of elite youth male soccer matches
Published in Journal of Sports Sciences, 2023
Songmi Kim, Stacey Emmonds, Paul Bower, Dan Weaving
To assess model fit across all models, the models were each compared against a null model (with only random effects included), using the Akaike Information Criterion (AIC) (with lower the AIC indicating better model fit). The magnitude and direction of difference (effect sizes [ES] ± 95% confidence intervals [CI]) as well as significance (p-value) were determined for comparisons. For the current study, ES was classified as trivial (<0.2), small (0.2–0.59), moderate (0.6–1.19), large (1.2–2.0) and very large (>2.0) (W. G. Hopkins, 2010). All statistical analyses were conducted using R Studio with alpha level set at p < 0.05 for all comparisons.
Effects of land use on time-of-day transit ridership patterns
Published in Transportmetrica A: Transport Science, 2022
The fitness of the model was statistically evaluated using the Comparative Fit Index (CFI) and Goodness of Fit Index (GFI). We also evaluated the performance of the model by comparing the final model with the null model that does not include any explanatory variables, using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). The LGCMs were developed using the ‘lavaan’ package in R (Rosseel et al. 2015).