This lecture discussed the prescriptive approach for decision making through the techniques of fast & frugal trees. This was demonstrated through two cases in the lecture. First, we addressed the question of whether the possibility of flagging banks at the risk of failing is reasonable. Using data, there was an improvisation on an existing tree by Aihman et al by including threshold values. Further, the logit models were discussed along with F&F models. It was also observed that the logit models were better than the F&F trees in terms of hit rate & false alarms. However, with less data, the results were reversed. F&F models are found to be better than the logit models.
Secondly, The case of NATO checkpoints in Afghanistan saw seven suicide attacks that caused a considerable number of civilian casualties. The idea here was to develop an F&F tree to minimise the casualties. Prof. Konstantinos explained that using a reasonably simple F&F tree, the number of casualties could have been possibly reduced by 40%. This was an interesting example that could be applied to a real life scenario.