Nội dung text 04_08_Simulating Model Output.pdf
pacmann.io © 2022 – Pacmann AI 46 Simulate Model Output
pacmann.io © 2022 – Pacmann AI 47 Simulating Model Output ● After fitting a model , we want to make sure our model can learn / imitate problem we are modelling ● To do so we can generate / simulate data based on posterior or posterior predictive check ● If our model working properly then the simulated data should look similar to observed data
pacmann.io © 2022 – Pacmann AI 48 Simulating Model Output For example in globe tossing to produce number of Water obtained from 9 toss Since the tossing is binomial event, we can create Random Variables using Binomial Distribution . X ~ Binomial (N,θ) Hence we need two parameters : 1. n (Number of Trials) , in this case is 9 2. θ (Probability / Water Proportion)
pacmann.io © 2022 – Pacmann AI 49 Simulating Model Output To run simulation / posterior predictive check we can derive as : posterior