The task is to predict five attributes for each video based on its content. Formally, the goal is to learn a function \( f \) such that:

$$
f(y) \rightarrow \left(X_1, X_2, X_3, X_4, X_5\right)
$$

where \( X_1, X_2, X_3 \) are categorical variables, and \( X_4, X_5 \) are continuous variables.