91Ó°ÊÓ

Explanatory variables, also known as predictor variables or independent variables, are the variables in a regression analysis that are used to explain variations in the response variable. In our example, we have three explanatory variables: \( x_1 \), \( x_2 \), and \( x_3 \). These variables are used to build a linear regression model that predicts the outcome variable (denoted by \( y \)). Each explanatory variable is associated with a coefficient (e.g., \( \beta_1 \), \( \beta_2 \), \( \beta_3 \)), which quantifies the effect of the explanatory variable on the response variable. The simplest model without interaction terms would look like this:

\[ y = \beta_0 + \beta_1 x_1 + \beta_2 x_2 + \beta_3 x_3 + \epsilon \]
Here:

• \( \beta_0 \) is the intercept, signifying the expected value of \( y \) when all explanatory variables are zero.
• \( \epsilon \) is the error term, which accounts for variability in \( y \) that cannot be explained by the predictors.

To master the concept of explanatory variables, remember they are the inputs we use to predict the output.

Interaction terms in a regression model determine if the effect of one explanatory variable on the response variable depends on the level of another explanatory variable. These terms are added to the model as the product of two explanatory variables. For example, if we want to model the interaction between \( x_1 \) and \( x_2 \), we add the term \( \beta_{12}(x_1 \times x_2) \). The equation then looks like this:

\[ y = \beta_0 + \beta_1 x_1 + \beta_2 x_2 + \beta_{12} (x_1 \times x_2) + \epsilon \]
Here, \( \beta_{12} \) captures the effect of the interaction between \( x_1 \) and \( x_2 \) on the response variable \( y \).

Interaction terms are crucial when the impact of one predictor on the dependent variable is modified by another variable. For example, the combined effect of study time (\( x_1 \)) and tutoring hours (\( x_2 \)) on a student's grades (\( y \)) can be different from their individual effects. To include such interactions, we can extend models with more than one interaction term as needed, like:

• \( \beta_{23}(x_2 \times x_3) \) for interaction between \( x_2 \) and \( x_3 \).

Regression coefficients are the values that multiply the explanatory variables in a regression equation. They represent the strength and direction of the relationship between each explanatory variable and the response variable. In our models, \( \beta_1 \), \( \beta_2 \), and \( \beta_3 \) are the coefficients for \( x_1 \), \( x_2 \), and \( x_3 \) respectively.

For instance, in the equation:

\[ y = \beta_0 + \beta_1 x_1 + \beta_2 x_2 + \beta_3 x_3 + \epsilon \]
If \( \beta_1 \) is 2, for every one unit increase in \( x_1 \), \( y \) increases by 2 units, assuming other variables stay the same.

Regression coefficients for interaction terms tell us how the interaction between variables affects the response variable. In the equation,

\[ y = \beta_0 + \beta_1 x_1 + \beta_2 x_2 + \beta_{12}(x_1 \times x_2) + \epsilon \]
\( \beta_{12} \) shows how the combined effect of \( x_1 \) and \( x_2 \) influences \( y \).

Understanding regression coefficients:

• Positive coefficients indicate a direct relationship between the predictor and response.
• Negative coefficients indicate an inverse relationship.
• The magnitude of the coefficient shows the strength of the relationship.
• Customized coefficients for interaction terms highlight the complexity of these combined relationships.