Preguntas etiquetadas con chi-squared

Una prueba (generalmente de distribución, independencia o bondad de ajuste) o una familia de distribuciones relacionadas con dicha prueba.

78
Un ejemplo: regresión LASSO usando glmnet para el resultado binario

Estoy empezando a incursionar con el uso de glmnetla LASSO regresión donde mi resultado de interés es dicotómica. He creado un pequeño marco de datos simulados a continuación: age <- c(4, 8, 7, 12, 6, 9, 10, 14, 7) gender <- c(1, 0, 1, 1, 1, 0, 1, 0, 0) bmi_p <- c(0.86, 0.45, 0.99, 0.84,...

r self-study lasso regression interpretation anova statistical-significance survey conditional-probability independence naive-bayes graphical-model r time-series forecasting arima r forecasting exponential-smoothing bootstrap outliers r regression poisson-distribution zero-inflation genetic-algorithms machine-learning feature-selection cart categorical-data interpretation descriptive-statistics variance multivariate-analysis covariance-matrix r data-visualization generalized-linear-model binomial proportion pca matlab svd time-series correlation spss arima chi-squared curve-fitting text-mining zipf probability categorical-data distance group-differences bhattacharyya regression variance mean data-visualization variance clustering r standard-error association-measure somers-d normal-distribution integral numerical-integration bayesian clustering python pymc nonparametric-bayes machine-learning svm kernel-trick hyperparameter poisson-distribution mean continuous-data univariate missing-data dag python likelihood dirichlet-distribution r anova hypothesis-testing statistical-significance p-value rating data-imputation censoring threshold
46
Interpretación del logaritmo transformador predictor y / o respuesta

Me pregunto si hace una diferencia en la interpretación si solo el dependiente, tanto el dependiente como el independiente, o solo las variables independientes se transforman logarítmicamente. Considere el caso de log(DV) = Intercept + B1*IV + Error Puedo interpretar el IV como el porcentaje...

regression data-transformation interpretation regression-coefficients logarithm r dataset stata hypothesis-testing contingency-tables hypothesis-testing statistical-significance standard-deviation unbiased-estimator t-distribution r functional-data-analysis maximum-likelihood bootstrap regression change-point regression sas hypothesis-testing bayesian randomness predictive-models nonparametric terminology parametric correlation effect-size loess mean pdf quantile-function bioinformatics regression terminology r-squared pdf maximum multivariate-analysis references data-visualization r pca r mixed-model lme4-nlme distributions probability bayesian prior anova chi-squared binomial generalized-linear-model anova repeated-measures t-test post-hoc clustering variance probability hypothesis-testing references binomial profile-likelihood self-study excel data-transformation skewness distributions statistical-significance econometrics spatial r regression anova spss linear-model
28
Calcular la repetibilidad de los efectos de un modelo más antiguo

Acabo de encontrar este artículo , que describe cómo calcular la repetibilidad (también conocida como confiabilidad, también conocida como correlación intraclase) de una medición a través del modelado de efectos mixtos. El código R sería: #fit the model fit =

mixed-model reliability intraclass-correlation repeatability spss factor-analysis survey modeling cross-validation error curve-fitting mediation correlation clustering sampling machine-learning probability classification metric r project-management optimization svm python dataset quality-control checking clustering distributions anova factor-analysis exponential poisson-distribution generalized-linear-model deviance machine-learning k-nearest-neighbour r hypothesis-testing t-test r variance levenes-test bayesian software bayesian-network regression repeated-measures least-squares change-scores variance chi-squared variance nonlinear-regression regression-coefficients multiple-comparisons p-value r statistical-significance excel sampling sample r distributions interpretation goodness-of-fit normality-assumption probability self-study distributions references theory time-series clustering econometrics binomial hypothesis-testing variance t-test paired-comparisons statistical-significance ab-test r references hypothesis-testing t-test normality-assumption wilcoxon-mann-whitney central-limit-theorem t-test data-visualization interactive-visualization goodness-of-fit
27
¿Pueden los grados de libertad ser un número no entero?

Cuando uso GAM, me da un DF residual de 26.626.626.6 (última línea en el código). Qué significa eso? Yendo más allá del ejemplo de GAM, en general, ¿puede el número de grados de libertad ser un número no entero? > library(gam) > summary(gam(mpg~lo(wt),data=mtcars)) Call: gam(formula = mpg ~...

r degrees-of-freedom gam machine-learning pca lasso probability self-study bootstrap expected-value regression machine-learning linear-model probability simulation random-generation machine-learning distributions svm libsvm classification pca multivariate-analysis feature-selection archaeology r regression dataset simulation r regression time-series forecasting predictive-models r mean sem lavaan machine-learning regularization regression conv-neural-network convolution classification deep-learning conv-neural-network regression categorical-data econometrics r confirmatory-factor scale-invariance self-study unbiased-estimator mse regression residuals sampling random-variable sample probability random-variable convergence r survival weibull references autocorrelation hypothesis-testing distributions correlation regression statistical-significance regression-coefficients univariate categorical-data chi-squared regression machine-learning multiple-regression categorical-data linear-model pca factor-analysis factor-rotation classification scikit-learn logistic p-value regression panel-data multilevel-analysis variance bootstrap bias probability r distributions interquartile time-series hypothesis-testing normal-distribution normality-assumption kurtosis arima panel-data stata clustered-standard-errors machine-learning optimization lasso multivariate-analysis ancova machine-learning cross-validation
20
¿Cuáles son los valores correctos para precisión y recuperación en casos extremos?

La precisión se define como: p = true positives / (true positives + false positives) ¿Es cierto que, como true positivesy false positivesenfoque 0, la precisión se aproxima a 1? La misma pregunta para recordar: r = true positives / (true positives + false negatives) Actualmente estoy...

precision-recall data-visualization logarithm references r networks data-visualization standard-deviation probability binomial negative-binomial r categorical-data aggregation plyr survival python regression r t-test bayesian logistic data-transformation confidence-interval t-test interpretation distributions data-visualization pca genetics r finance maximum probability standard-deviation probability r information-theory references computational-statistics computing references engineering-statistics t-test hypothesis-testing independence definition r censoring negative-binomial poisson-distribution variance mixed-model correlation intraclass-correlation aggregation interpretation effect-size hypothesis-testing goodness-of-fit normality-assumption small-sample distributions regression normality-assumption t-test anova confidence-interval z-statistic finance hypothesis-testing mean model-selection information-geometry bayesian frequentist terminology type-i-and-ii-errors cross-validation smoothing splines data-transformation normality-assumption variance-stabilizing r spss stata python correlation logistic logit link-function regression predictor pca factor-analysis r bayesian maximum-likelihood mcmc conditional-probability statistical-significance chi-squared proportion estimation error shrinkage application steins-phenomenon