Is FOFC working

[cg09]

Will someone please check me? in 6.7.0 FOFC is not working, unless I am doing something quite wrong. I attach a picture of the model and the FOFC output. The sample size is 1,000 and all coefficients are positive.

[jdramsey]

That's right, actually. Once you remove the impure X6, there are no clusters of size 4.

…

On Fri, Apr 26, 2019 at 5:42 PM cg09 ***@***.***> wrote: Will someone please check me? in 6.7.0 FOFC is not working, unless I am doing something quite wrong. I attach a picture of the model and the FOFC output. The sample size is 1,000 and all coefficients are positive. [image: Screen Shot 2019-04-26 at 5 38 23 PM] <https://user-images.githubusercontent.com/16354745/56838001-892ff180-684a-11e9-902c-05b21f73c7d9.png> [image: Screen Shot 2019-04-26 at 5 38 47 PM] <https://user-images.githubusercontent.com/16354745/56838011-9056ff80-684a-11e9-88fe-267400679378.png> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1098>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACLFSRYBQYR7FZ2SRYTMXHLPSNZKNANCNFSM4HI2QN7Q> .

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213 jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

[cg09]

Not so. X1, X2, X3, X4 implies 3 tetrads, as do X1, X2, X3 and X5, as do X7, X8, X9 with X4 or X5. From the Kummerfeld Ramsey paper: "Let a *vanishing quartet* be a set of 4 indicators in which all 3 tetrads among the 4 variables are entailed to vanish by the PETS theorems....The algorithm first calls *FindPureClusters*, which tests each triple to see if it has the property that adding any other member of *O* creates a vanishing quartet; if it does have the property it is added to the list *PureList* of pure triples. " The rest is merging clusters. On Fri, Apr 26, 2019 at 6:35 PM Joseph Ramsey <notifications@github.com> wrote:

…

That's right, actually. Once you remove the impure X6, there are no clusters of size 4. On Fri, Apr 26, 2019 at 5:42 PM cg09 ***@***.***> wrote: > Will someone please check me? in 6.7.0 FOFC is not working, unless I am > doing something quite wrong. I attach a picture of the model and the FOFC > output. The sample size is 1,000 and all coefficients are positive. > > [image: Screen Shot 2019-04-26 at 5 38 23 PM] > < https://user-images.githubusercontent.com/16354745/56838001-892ff180-684a-11e9-902c-05b21f73c7d9.png > > > [image: Screen Shot 2019-04-26 at 5 38 47 PM] > < https://user-images.githubusercontent.com/16354745/56838011-9056ff80-684a-11e9-88fe-267400679378.png > > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#1098>, or mute the thread > < https://github.com/notifications/unsubscribe-auth/ACLFSRYBQYR7FZ2SRYTMXHLPSNZKNANCNFSM4HI2QN7Q > > . > -- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213 ***@***.*** Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1098 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AD4Y3OJ3UJUWQXIQ3MAIMB3PSN7RVANCNFSM4HI2QN7Q> .

[jdramsey]

@cg09 Are you still worried about this?

[cg09]

It seems to be working ok, with the correct version. Shuyan has been running it. We need to change the manual and interface description to note the conditions under which it works with categorical variables.

…

On Wed, Aug 7, 2019 at 2:11 PM Joseph Ramsey ***@***.***> wrote: @cg09 <https://github.com/cg09> Are you still worried about this? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1098?email_source=notifications&email_token=AD4Y3OJRI5ZKS5MUW36UL6TQDMF6VA5CNFSM4HI2QN72YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3ZIJ4Q#issuecomment-519210226>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AD4Y3OMSC6I3JA7IYQQYMMLQDMF6VANCNFSM4HI2QN7Q> .

[jdramsey]

OK, so this is a description problem, then. I'll leave it open for now.

[jdramsey]

@cg09 Could you or Shuyan write up a description of the conditions under which FOFC works with categorical variables? I can add it to the manual. I just don't know those conditions.