以下是ppt內容: ---------------------------- Bayes’ Rule We are also given estimates of the probabilities of each state of the world. These are called prior probabilities. In different states of the world, different decisions may be optimal. It may be desirable to purchase information that gives the decision maker more foreknowledge about the state of the world. This may enable the decision maker to make better decisions. Example of Bayes’ rule: A lot of chips content 10 chips 80% good (with one defect out of 10) 20% bad (with 5 defect out of ten) Process information Process a good lot $1000 Process a bad lot $4000 Rework a lot then process $2000 (rework $1000 + process $1000) Inspection: $100 to test one chip EVSI? EVPI? Example cont’ States (prior prob.) Good lot: P(G)=0.8 Bad lot: P(B)=0.2 Outcomes of inspection D: defective is observed ND: nondefective is observed Thus we have: P(D|G)=0.1 P(ND|G)=0.9 P(D|B)=0.5 P(ND|B)=0.5 Joint probabilities are: P(D∩G)=P(G)P(D|G)=0.8*0.1 P(D∩B)=0.1 P(ND∩G)=0.72 P(ND∩B)=0.1 Probability of each experiment outcome: P(D)=P(D∩G)+P(D∩B)=.18 P(ND)=.82 Then we have: (posterior prob.) P(G|D)=4/9 P(G|ND)=72/82 P(B|D)=5/9 P(B|ND)=10/82 ------------------------------------- 一半以上都看不懂...... 有人能說明一下每個的機率是怎麼來的? 還有P(D|G)跟P(G|D)到底差在哪? 能麻煩具體說明一下嗎， 有請70神人了@@ -- 願為佛陀 卻成心魔 執著是苦 走火入魔 魔道亦道 分別徒勞 人心難測 善惡為何 --

※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 118.169.143.221