S01E05: The Placement Conundrum
Questions:
If there are 1000 houses in an area, how many houses should I see to pick a house that is among the top 5% percentile of houses?
When to stop searching for the best option (house, job, partner, etc.), given we know the number of options?
Additional Small Problem:
If there are 100 students in a class with a distribution of 90 boys and 10 girls, the teacher has selected only female students as the class representative from the past 10 years and says that she uses a random picking algorithm to get the candidate. Is she lying or telling the truth?
Additional Questions:
Can we create a strategy to select the best job offer if we know how many offers we will have?
If we do not know the total number of offers we are going to have, can we still pick a relatively good offer?
How to make sure we make a good choice (around the best) and not to lose good options in greed of better ones?
A case where we know what percentile or group we are targeting compared to a case where we just want to get round about the best, how these are different? and how will the strategies differ?
What is the optimal number of options I should see to make a good estimation of the population?
Coding Assignments:
Create a program to solve the problem of getting the best package emperically.
Plot different graphs, to see how the relationships between options seen, estimation of the population, and the best option we picked.
Create program for different variations of this problem, i.e. when we know the number of of options, when we don’t know the number of options, .
Create a program how many times we actually get a house from top X percentile given we see as many houses as many groups are there, i.e. in case of top 5% percentile it is 20 groups and hence we should see 20 houses.
← Back to Season 01