med-cause/mc_analysis.R at master · jonasmago/med-cause · GitHub

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
46
47
48
49
50
51
52
'''
WHAT I WANT TO KNOW:
for each model:
- LogL (MLE)
- BIC
--> note that it is not fitted for entire data but summed across all
- Px
- Accuracy
- fitted parameters and SD???


'''
########################
# overall fitting the model
########################
columns = c(2,9,30,51,72,16,37,58,79,23,44,65,86)-8
df.mc_fitted_combined[,columns]


columns = c(9,30,51,72,16,37,58,79,23,44,65,86)-2
df.mc_fitted_combined[,columns]


########################
# best fit by participant!
########################
colnames(df.mc_fitted)
columns = c(2,9,30,51,72,16,37,58,79,23,44,65,86)

# with all
#columns = c(2,9,30,51,16,37,58,23,44,65) # without just softmax
#columns = c(2,9,16,23) # only full modles
#columns = c(2,9,30,51,23,44,65) # without just softmax

p = df.mc_fitted[,columns]
p_min = as.data.frame(apply(p,1,min))
p_dif = p - data.frame(rep(p_min,length(columns)))
d = which((p_dif == 0),arr.ind=TRUE)
data.frame (table (d[,2]))

colnames(p)

t(d)[1,]
##############################
# Chi Square test for participant distribution
##############################
p = c(0, 5, 3, 0, 0, 0, 0, 0, 4, 27, 5, 0, 0)

chisq.test(p)

m= 0; pun = 3;
-2*-193355.8