Dear Prof. Imaj,
Yes you are right. That is my confusion. The optimal matching ratio =1
has been running for 24hrs by now. I have sample size of Control (2692)
vs Treated (1306). I tried use full match with min.controls = 1, and
max.controls = 1 to try to substitute the optimal ratio = 1 case. But
the result still match all the samples instead of 2 times of the number
of treated. Why is that?
Thank you.
Jeanie
On Wed, 2006-03-15 at 14:40 -0500, Kosuke Imai wrote:
Dear Jeanie,
So, the problem is that the code works when ratio=2, but it doesn't when
ratio=1 (the default)? Depending on the number of observations (and the
total number of possible configurations), optimal matching can be
computationally intensive. You might want to increase the tolerence level
to a larger value, the defualt is tol=0.001. You can also change the
values of max.control and min.control to make the problem more manageable
for the algorithm. See
http://www.stat.lsa.umich.edu/~bbh/optmatch/fullmatch.html
Good luck with your project,
Kosuke
-----------------------------------------------------
Kosuke Imai Office: Corwin Hall 041
Assistant Professor Phone: 609-258-6601
Department of Politics eFax: 973-556-1929
Princeton University Email: kimai(a)Princeton.Edu
Princeton, NJ 08544-1012
http://imai.princeton.edu
-----------------------------------------------------
On Wed, 15 Mar 2006, Jeanie (Jie) Na wrote:
Dear Professor,
I am using matchit for one project and run into some problems that I
could not figure out on my own. I hope you could give me some help.
My data is as follows:
26 level of CENTERs - recoded to 25 CENTER.num binary (0/1) variables;
8 level of ANAESTHC - recoded to 7 ANAESTHC.num binary vars;
4 level of PREDPATC - recoded to 3 PREDPATC... binary vars;
3 level of OP_C - recoded to 2 OP_C... binary vars;
3 level of surg2appc - recoded to 2 surg2appc.. binary vars;
3 level of ASA_R. - recoded to 2 ASA_R. binary vars;
The rest of the vars in the model are continuous vars.
The code has run for 20 hours upto now. I dont know if I need to
terminate it. Is there anything wrong?
The same model but using optimal with ratio = 2 runs for about 20
minutes to get the result.
m.out.all <- matchit(NASOTUBB ~ CENTER.3 + CENTER.10 + CENTER.13 +
CENTER.15 + CENTER.17 + CENTER.19 + CENTER.21 + CENTER.22 + CENTER.23 +
CENTER.34 + CENTER.2 + CENTER.5 + CENTER.6 + CENTER.7 + CENTER.8 +
CENTER.11 + CENTER.12 + CENTER.14 + CENTER.16 + CENTER.18 + CENTER.30 +
CENTER.31 + CENTER.32 + CENTER.33 + CENTER.113
+ ANAESTHC.0 + ANAESTHC.5 + ANAESTHC.4 + ANAESTHC.6 + ANAESTHC.3 +
ANAESTHC.1 + ANAESTHC.7
+ PREDPATC...neither.nor + PREDPATC...PONV + PREDPATC...no.PONV
+ OP_C...knee.arthroscopy + OP_C...cholecystectomy +
OP_C...general.surgery + OP_C...bone.surgery +
OP_C...upper.extremity.surgery + OP_C...thyroid.surgery +
OP_C...ENT.Eye.surgery + OP_C...hysterectomy +
OP_C...gynaecologic.surgery + OP_C...breast.surgery +
OP_C...hip.replacement
+ surg2appc...other + surg2appc...abdominal.laparoscopic
+ ASA_R...1 + ASA_R...2
+ BMI_N + YEAREXPN + PONV_GAB + PONV_GAN + KPONVRN,
data = matchuse1, method = "optimal", distance = "logit")
Thank you very much in advance for any suggestions. Looking forward to
your reply.