Consider the following expression:

\frac{\lambda _{2}}{\lambda _{3}y^{\lambda _{3}-1}+ \lambda _{4}\left ( 1-y \right )^{\lambda _{4}-1}}\geq 0

Obviously, it depends on

g\left ( y,\lambda_{3},\lambda_{4} \right )\equiv \lambda _{3}y^{\lambda _{3}-1}+ \lambda _{4}\left ( 1-y \right )^{\lambda _{4}-1}

but I am having difficulty solving this to determine the \left ( \lambda_{3},\lambda_{4} \right ) that will lead to a valid distribution.

How do I go about finding the parameter space i.e. the regions where this pdf would be valid?