In GegznaV/RcmdrPlugin.BioStat: R Commander Plug-in for Basic Data Management and Analysis Tasks

head(trees)

lm(Girth ~ Height, data = trees)

# modelio formulės: Y ~ X1 + X2, data = duomenys
# lm(formula, data = ___)
lm_modelis <- lm(Girth ~ Height + Volume, data = trees)
# Rezultatai
print(lm_modelis)
# Modelio tinkamumo duomenims diagnostika
summary(lm_modelis)

broom::glance(lm_modelis)
broom::augment(lm_modelis)
broom::tidy(lm_modelis)
biostat::coef_standardized(lm_modelis)
sigr::wrapFTest(lm_modelis)
predict(lm_modelis)
trees$prediction <- predict(lm_modelis)

# lm(formula, data = ___)
# lm_modelis <- lm(formula, data = dframe)
# print(lm_modelis)
# broom::glance(lm_modelis)
# broom::augment(lm_modelis)
# broom::tidy(lm_modelis)
# biostat::coef_standardized(lm_modelis)
# sigr::wrapFTest(lm_modelis)
# predict(lm_modelis)
# dframe$pred <- predict(lm_modelis)
# 
# ggplot(dframe, aes(x = pred, y = outcome)) + 
       # geom_point() +  
       # geom_abline(color = "blue")
       # 
# predict(model, newdata)
# WVPlots::GainCurvePlot(frame, xvar, truthvar, title)
WVPlots::GainCurvePlot()
# In the plot: A relative gini coefficient close to one shows that the model correctly sorts high unemployment situations from lower ones. 

# RMSE  # galima interpretuoti: tipinė prognozės klaida.
# sd(Y) # Y mūsų modeliuojamo kintamojo reikšmė
#       # sd - tipinis skirtumas tar konkrečios ir vidutinės kainos.
#       Jei RMSE < sd(Y), tai mūsų modelis geriau nuspėja kainas nei 
#       modelis, kuris spėtų, kad kaina = vidutinė kaina.
#       
# R^2   - galioja tik mokymo duomenim      

# Cross-validation only evaluates modeling process, not future prediction.
# Remember, cross-validation validates the modeling process, not an actual model. 
library(vtreat)
splitPlan <- kWayCrossValidation(nRows, nSplits, NULL, NULL)

GegznaV/RcmdrPlugin.BioStat documentation built on May 8, 2023, 7:41 a.m.