Please help me solve this probability problem i understand it conceptually but do not know how to solve the math behind the three cases

Suppose you are preparing for a camping trip and want to determine how cold it will be so that you know what gear to bring along. To help you make your prediction, you gather a list of historical temperatures for the dayr of your trip. You Wish to aggregate them into a single number using the approach of loss function minimization this will be your prediction. You decide that it would be better if your prediction is too lowr as opposed to too high. If your prediction is too low, you'll end up packing a lot of extra clothing, but you'll be warm. On the other hand, if your prediction is too high, you'll pack no extra clothing and you'll freeze to death. You decide that overestimating the temperature by 6 degra- is twice as bad as underestimating the temperature by the same amount. To account for this, you adopt the following loss function: u |h. yd if 3;, > h, (i.e., h underestimates yi) L(h) = Z 2- lh yil if y;