Statistics for Decision Makers - 29.01 - Decision Trees

From Training Material
Jump to navigation Jump to search
title
29.01 - Decision Trees
author
Bernard Szlachta (NobleProg Ltd) bs@nobleprog.co.uk

Example 1 。

What should we invest money in?

ClipCapIt-140526-161958.PNG

Explanation 。

Symbol Meaning
ClipCapIt-140526-162648.PNG decision node
ClipCapIt-140526-162629.PNG uncertainty, event, chance node
ClipCapIt-140526-162543.PNG end node

Probabilities 。

What should we invest money in?

Decisiontree2.svg

Expected Value。

Decisiontree3.svg

Exercise 1。

  • A company thinks about introducing a new social networking platform called GooGl-.
  • Because GooGl- would need to compete with existing platforms like FaceHook, Witter or WeiBu, there is a good chance that it will fail (98%)
  • The company estimates that initial investment would be $10 million
  • If they succeed, the business could be worth around $1 billion
  • Should the company launch the GooGl- platform?

Dependent Uncertainties。

Imagine that NobleProg wants to invest in self-study, but the major cost is developing a testing platform

  • NobleProg is using Drupal, an open source software
  • It is possible that the community develop part of the modules which NobleProg require in the next two months (0.9), and the cost will decrease from 400 to 200
  • In this case our EV is 180 for the "wait" scenario

Decisiontree4 (1).svg

Read More。

http://vserver1.cscs.lsa.umich.edu/~spage/ONLINECOURSE/R4Decision.pdf