In order to decide upon a point estimate (a single-value summary) for the posterior distribution, we need to pick a loss function. The absolute loss function is defined as the sum of the absolute difference between the predicted parameter and its true value, weighted by the posterior distribution. The quadratic loss function is defined as the sum of the square of the difference between the predicted parameter and its true value, weighted by the posterior distribution. a. What point estimate does the absolute loss function lead to? b. What point estimate does the quadratic loss function lead to? 5. What are the three arithmetic processes that can lead to a normal distribution? Give concrete examples if necessary. 6. What are the five steps when designing a linear model under the Bayesian framework.