Your machine learning model for stock valuation yields varying importance scores for features across different market conditions:

i$.$ Explain how market volatility could affect feature importance in predictive models.

$(2$ marks$)$

ii$.$ Propose methods to ensure model robustness against such variability.

$(1\mathrm{.}5$ marks$)$

iii. Discuss the challenges in applying this model for real$-$time trading decisions.

$(1\mathrm{.}5$ marks$)$