-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathAnalyzeFitness.r
45 lines (31 loc) · 2.81 KB
/
AnalyzeFitness.r
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
# Anne S. Warlaumont
# Test to see if there are statistically significant differences between Realistic and Abstract simulations at generation 100
# If you run with the following command, a file containing the output will be saved to the directory containing this script:
# R CMD BATCH AnalyzeFitness.r
rm(list=ls())
library(lme4)
library(lmerTest)
# Change the directory as needed:
#simdata = read.csv('~/Downloads/MultipleRunsData.csv',header=T)
simdata = read.csv('/Volumes/Storage/NeuralNetVocalControlEvolutionRuns/multRunsSummary_Ready/MultipleRunsData.csv',header=T)
simdata$medProFitnessScaled = scale(simdata$medProFitness)
simdata$medPerFitnessScaled = scale(simdata$medPerFitness)
simdata$generationScaled = scale(simdata$generation)
t.test(subset(subset(simdata,useVocalTract==0),generation==100)$medProFitness,subset(subset(simdata,useVocalTract==1),generation==100)$medProFitness)
t.test(subset(subset(simdata,useVocalTract==0),generation==100)$medPerFitness,subset(subset(simdata,useVocalTract==1),generation==100)$medPerFitness)
t.test(subset(subset(simdata,useVocalTract==0),generation==500)$medProFitness,subset(subset(simdata,useVocalTract==1),generation==500)$medProFitness)
t.test(subset(subset(simdata,useVocalTract==0),generation==500)$medPerFitness,subset(subset(simdata,useVocalTract==1),generation==500)$medPerFitness)
aggdataPro = aggregate(medProFitness~runfolder+useVocalTract,subset(simdata,generation>25 & generation < 150),mean)
t.test(subset(aggdataPro,useVocalTract==0)$medProFitness,subset(aggdataPro,useVocalTract==1)$medProFitness)
aggdataPer = aggregate(medPerFitness~runfolder+useVocalTract,subset(simdata,generation>25 & generation < 150),mean)
t.test(subset(aggdataPer,useVocalTract==0)$medPerFitness,subset(aggdataPer,useVocalTract==1)$medPerFitness)
proFitModel = lmer(medProFitnessScaled~(1|runfolder)+useVocalTract+generationScaled+useVocalTract*generationScaled,data=subset(simdata,generation<=100 & generation > 50))
summary(proFitModel)
quartz(); qqnorm(resid(proFitModel))
perFitModel = lmer(medPerFitnessScaled~(1|runfolder)+useVocalTract+generationScaled+useVocalTract*generationScaled,data=subset(simdata,generation<=100 & generation > 50))
summary(perFitModel)
quartz(); qqnorm(resid(perFitModel))
var.test(subset(subset(simdata,useVocalTract==0),generation==100)$medProFitness,subset(subset(simdata,useVocalTract==1),generation==100)$medProFitness)
var.test(subset(subset(simdata,useVocalTract==0),generation==100)$medPerFitness,subset(subset(simdata,useVocalTract==1),generation==100)$medPerFitness)
var.test(subset(subset(simdata,useVocalTract==0),generation==500)$medProFitness,subset(subset(simdata,useVocalTract==1),generation==500)$medProFitness)
var.test(subset(subset(simdata,useVocalTract==0),generation==500)$medPerFitness,subset(subset(simdata,useVocalTract==1),generation==500)$medPerFitness)