Simulate recognition data with Excel
Generate a “study list” of 10 items, represented as vectors (can use simple 4 feature items as in the example, or more complex). There are no specific rules for what these items must look like – they can be any combination of 4 numbers, but I recommend using single digits as in the example from lecture.
Parameters: u (probability of encoding)si (probability of source being the item itself)sr (probability of source being a random number) Note: sr = 1-si, this is simplified from the real model.c (probability of copying correctly)Here are some suggested starting values for your parameters:u = 0.3, si = 0.7, sr = (1-si) = 0.3, c = 0.8
Use the parameters above to create 10 memory “traces”, as if a subject were going through the process of encoding these 10 items. When creating these 10 traces, use the process described below:
1. Each feature of each study item presented is
a. Encoded with some probability u
b. The source of representation is chosen with probabilities si (from item) and sr (random choice)
c. Feature is copied correctly with some probability c (if not copied correctly, replace with random value)
WHAT TO TURN IN
:1) Show the 10 original items that make up the study list (this will be a list of 10 vectors).
2) For one run of the loop described above, show the memory “traces” after study as 10 vectors, which should be imperfect and incomplete copies of the study items.
3) Come up with a (simple) way to compare each trace to each actual item from the study list to give each simulation run a memory “score”. Run the simulation 100 times, plot a histogram of the memory scores, and show the average of the 100 memory scores.a. Extra credit: Complete step 3 several times, changing parameters like u and c, to see how the memory score changes. Plot a histogram for each run to see the effect of varying your parameter.
4) Attach your Excel File showing your simulation.