The QMV iterative process of analysis
“when will my service tech arrive?”
Dispatch sys estimated arrival time
tech’s reported arrival time
time of tech’s prior finished job
=>
Classification e.g. before noon/afternoon
regression estimator
Classification route
Bayes Net, Perceptron, Logistic Regression
Candy bowl and consumer choice modeling
Consumer choice modeling: Understand how consumer make decisions
– Is it possible to find preference ordering for product brends
– Can we infer that there even exists a preference between brands?
Description of the Candy Bowl Data
consumer choice
-name
-gender
-candy
-candy color/flavor
-age
-ethnicity
Time between selections
“Interselection time” for candy C = # turns between selections of c
In []: plot_interselection_time(event_list, "orange", "airhead") In []: plot_interselection_time(event_list, "red", "starburst") plot_interselection_time(event_list, "orange", "airhead")
Point estimation, Confidence sets, Classification, Hypothesis testing
r: interselection time for candy, c, at a given turn
x = (“airhead”, 1), (“role”, 5), (starburst, 7),…
c, choice #, interselection time of other candies in bowl, r(c, choice #)
