Review of Tiastis ’ slides on Dynamic Treatment Regime (iii) 11/07/2007
Discussion for today
* * ( , , ) { ( ) ( , 0) | , } l a E Y a Y a L l A a 1 1 1 j j j j j j j j
SA Murphy (JRSSB, 2003) Dynamic programming: ( , ) sup { [ | , , ]}, J L a E Y L a a 0 1 1 K K a K K K K * ( , ) arg sup { [ | , , ]}, d L a E Y L a a 1 1 K K K a K K K K The subsequently for each , j ( , ) sup { [ ( , ) | , , ]}, J L a E J L a L a a K j j j 1 a K j 1 j 1 j j j 1 j j * ( , ) arg sup { [ ( , ) | , , ]},......... d L a E J L a L a a 1 1 1 1 j j j a K j j j j j j j * * * Afterwards, the optimal rules are { , ,..., }. d d d 1 2 K
Too Naïve for a complex process
Recommend
More recommend