conducted discrete choice experiment but have been analyzed for 2 months. Anybody helps me with this mixed logit problem...
My experiment was conducted with 3 alternatives (option 1, option 2, option 3) and 3 attributes (soc, man, inc)
I also have a few individual characteristics and want to use these:
avg.charge.cost, avg.charge.num, milage start.time.mon,
start.time.aft, start.time.eve, start.time.nig, start.time.rand (these are dummy variables)
It is hard to interpret with socio-demographic variables. Below is the mlogit code and results. I cannot interpret socio-demographic variables such as "start.time.mon:2" since 2 is just the name of option which is consist of 3 alternatives (soc, man, inc) . It is not like boat, charter, something like in example......
Therefore, it doesn't have meaning such as start.time.mon:2, start.time.eve:3,milage:2 ...
- How can I delete intercept:2, intercept:3? and
- How should I put individual characteristics in this situation?
Call:
mlogit(formula = choice ~ soc + man + inc | avg.charge.cost +
avg.charge.num + milage + start.time.mon + start.time.aft +
start.time.eve + start.time.nig | 0, data = dce.mlogit.alt3,
rpar = c(soc = "n", man = "n", inc = "n"), R = 100, halton = NA,
panel = TRUE)
```Result:
Call:
mlogit(formula = choice ~ soc + man + inc | avg.charge.cost +
avg.charge.num + milage + start.time.mon + start.time.aft +
start.time.eve + start.time.nig | 0, data = dce.mlogit.alt3,
rpar = c(soc = "n", man = "n", inc = "n"), R = 100, halton = NA,
panel = TRUE)
Frequencies of alternatives:choice
1 2 3
0.29233 0.36408 0.34358
bfgs method
2 iterations, 0h:0m:34s
g'(-H)^-1g = 4.72E+03
last step couldn't find higher value
Coefficients :
Estimate Std. Error z-value Pr(>|z|)
(Intercept):2 0.33469658043 0.12180762852 2.7477 0.006001 **
(Intercept):3 0.22479454593 0.12506518319 1.7974 0.072269 .
soc 0.04314020545 0.00076835684 56.1461 < 0.00000000000000022 ***
man -0.10143611100 0.00434805932 -23.3291 < 0.00000000000000022 ***
inc 0.00003393733 0.00000221364 15.3310 < 0.00000000000000022 ***
avg.charge.cost:2 0.00000111634 0.00000108569 1.0282 0.303840
avg.charge.cost:3 -0.00000034602 0.00000105725 -0.3273 0.743454
avg.charge.num:2 0.00042453222 0.00316500460 0.1341 0.893297
avg.charge.num:3 -0.00443059693 0.00327233817 -1.3540 0.175751
milage:2 -0.00051653573 0.00028698077 -1.7999 0.071877 .
milage:3 0.00019542444 0.00029649619 0.6591 0.509823
start.time.mon:2 0.14603008484 0.10239227800 1.4262 0.153816
start.time.mon:3 0.15078992009 0.10450330490 1.4429 0.149043
start.time.aft:2 -0.00503292803 0.12710403716 -0.0396 0.968414
start.time.aft:3 -0.05512992835 0.12544537138 -0.4395 0.660318
start.time.eve:2 0.01147774556 0.06731536435 0.1705 0.864611
start.time.eve:3 -0.09856076724 0.06888887389 -1.4307 0.152510
start.time.nig:2 0.06017213241 0.06093596402 0.9875 0.323415
start.time.nig:3 -0.02837111182 0.06219073407 -0.4562 0.648250
sd.soc 0.09994166256 0.00024140374 414.0021 < 0.00000000000000022 ***
sd.man 0.09956856967 0.00208219538 47.8190 < 0.00000000000000022 ***
sd.inc 0.09998613851 0.00001984986 5037.1208 < 0.00000000000000022 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Log-Likelihood: -20771
McFadden R^2: -2.8906
Likelihood ratio test : chisq = -30864 (p.value = 1)
random coefficients
Min. 1st Qu. Median Mean 3rd Qu. Max.
soc -Inf -0.02426942 0.04314020545 0.04314020545 0.11054983 Inf
man -Inf -0.16859409 -0.10143611100 -0.10143611100 -0.03427813 Inf
inc -Inf -0.06740569 0.00003393733 0.00003393733 0.06747356 Inf