Why Should You Know Linear Regression?

• easy to use
• easy to interpret
• applicable for a huge variety of problems
• fairly accurate

\(\Rightarrow\) among the most widely used algorithms by data scientists

Why Should You Know General Linear Models?

• a specific case of GLM’s is linear regression with multiple predictors
• lots of problems with discrete outcomes
• error models have to stay interpretable even when the outcome has to be positive

## 
## Call:  glm(formula = Infant.Mortality ~ Fertility + Education + Catholic, 
##     data = swiss)
## 
## Coefficients:
## (Intercept)    Fertility    Education     Catholic  
##    7.867371     0.159631     0.103403    -0.006246  
## 
## Degrees of Freedom: 46 Total (i.e. Null);  43 Residual
## Null Deviance:       390.3 
## Residual Deviance: 298.3     AIC: 230.2

How Does The App Help You Understand GLM’s?

• no need to learn a language like R
• you can concentrate on the algorithm
• you can play around with different predictors and compare their effects

How Does The App Help You Explain GLM’s?

• no need to explain any code syntax
• focus on the interpretation
• explain the effects of in-/exclusion of certain predictors on
• the model quality
• the coefficients and their significance
• the model visualizations
• the sign of other coefficients (see examples to the right)

glm (
    Infant.Mortality ~ Fertility
                       + Examination,
    data = swiss
)$coefficients

## (Intercept)   Fertility Examination 
##  8.71870014  0.13718600  0.09710969

glm (
    Infant.Mortality ~ Examination,
    data = swiss
)$coefficients

## (Intercept) Examination 
## 20.62898680 -0.04162888