UW CSE Dataset
All Datasets: boost-starai/BoostSRL-Datasets
by: Nandini Ramanan, Alexander L. Hayes
<< "Cora" | BoostSRL Wiki | "CiteSeer" >>
From the UW-CSE Alchemy Page.
"This data set consists of information about the University of Washington Department of Computer Science and Engineering. The data has been anonymized to comply with the University of Washington's privacy guidelines."
As usual, the version here is a .zip with the necessary background and train/test folders.
Target: advisedby
The facts contain information on fourteen labels: courselevel
, hasposition
, inphase
, professor
, projectmember
, publication
, samecourse
, sameperson
, sameproject
, student
, ta
, taughtby
, tempadvisedby
, yearsinprogram
.
Table of Contents - BoostSRL Wiki
Download: UW-CSE.zip (257 KB)
-
md5sum
: 5e8217ebdb835ff8b6ff94eb3880d96b -
sha256sum
: f16be492805bdac95cded02a3a3e590c29a68145f5ea59eb4180c300fb23b7e2
Table of Contents - BoostSRL Wiki
Linux/Mac:
- After downloading, unzip UW-CSE.zip
unzip UW-CSE.zip
- If you're using a jar file, move it into the UW-CSE directory:
mv (jar file) UW-CSE/
- Learning:
java -jar BoostSRL.jar -l -train train/ -target advisedby -trees 10
- Inference:
java -jar BoostSRL.jar -i -test test/ -model train/models/ -target advisedby -trees 10
Windows:
(Coming soon)
Table of Contents - BoostSRL Wiki
setParam: loadAllLibraries = false.
setParam: treeDepth=3.
setParam: nodeSize=1.
setParam: numOfClauses=8.
setParam: numOfCycles=8.
importLibrary: listsInLogic.
queryPred: advisedby/2.
mode: professor(+Person).
mode: student(+Person).
mode: publication(+Title, -Person).
mode: publication(-Title, +Person).
mode: taughtby(+Course, +Person, -Quarter).
mode: taughtby(+Course, -Person, +Quarter).
mode: taughtby(-Course, +Person, -Quarter).
mode: courselevel(+Course, +Level).
mode: courselevel(+Course, #Level).
mode: hasposition(+Person, +Position!1).
mode: hasposition(+Person, #Position).
mode: multiclass_hasposition(+Person).
okIfUnknown: multiclass_hasposition/1.
mode: projectmember(+Project, -Person).
mode: projectmember(-Project, +Person).
range: Position={faculty_affiliate,faculty,faculty_adjunct,faculty_emeritus}.
range: Phase={pre_quals,post_generals,post_quals}.
mode: position(+Position).
mode: phase(+Phase).
position(faculty_affiliate).
position(faculty).
position(faculty_adjunct).
position(faculty_emeritus).
phase(pre_quals).
phase(post_generals).
phase(post_quals).
mode: advisedby(+Person, +Person).
mode: inphase(+Person, +Phase!1).
mode: inphase(+Person, #Phase).
mode: multiclass_inphase(+Person).
okIfUnknown: multiclass_inphase/1.
mode: tempadvisedby(-Person, +Person).
mode: tempadvisedby(+Person, -Person).
mode: yearsinprogram(+Person, #Integer).
mode: ta(+Course, -Person, +Quarter).
mode: ta(+Course, +Person, -Quarter).
mode: ta(-Course, +Person, -Quarter).
mode: sameperson(+Person, +Person).
mode: samecourse(+Course, +Course).
mode: sameproject(+Project, +Project).
mode: have_more_than_n_pubs(+Person, #PThresh).
mode: have_more_than_n_common_pubs(+Person, -Person, #PThresh).
mode: have_more_than_n_common_pubs(-Person, +Person, #PThresh).
mode: count_taughtby(+Person, -PThresh).
mode: count_publications(+Person, -PThresh).
mode: count_common_pubs(+Person, -Person, -PThresh).
mode: count_common_pubs(-Person, +Person, -PThresh).
usePrologVariables: true.
precompute:
commonpub(Title, P1,P2) :- publication(Title, P1), publication(Title, P2),P1\==P2.
precompute:
commonta(C,Q,P1,P2) :- ta(C,P2,Q), taughtby(C,P1,Q).
precompute1:
count_taughtby(Person,N) :- taughtby(SomeC, Person, SomeQ), all([Course, Quarter], taughtby(Course, Person, Quarter), AllCourses), N is length(AllCourses).
precompute1:
count_publications(Person,N) :- publication(Somet, Person), all(Title, publication(Title, Person), AllTitles), N is length(AllTitles).
precompute1:
count_common_pubs(P1,P2,N) :- commonpub(Somet, P1,P2), all(Title, commonpub(Title, P1,P2), AllTitles), N is length(AllTitles).
precompute2:
have_more_than_n_pubs(A,N) :-
count_publications(A,N2),
member(N,[1, 3, 5, 7, 9,11,13,15]),
N2 > N.
precompute2:
have_more_than_n_common_pubs(A1,A2,N) :-
count_common_pubs(A1,A2,N2),
member(N,[1, 3, 5, 7, 9,11,13,15]),
N2 > N.
Table of Contents - BoostSRL Wiki
BoostSRL Wiki
Home
BoostSRL Basics
- Getting Started
- File Structure
- Basic Usage Parameters
- Advanced Usage Parameters
- Basic Modes Guide
- Advanced Modes Guide
Deep dive into BoostSRL
- Default (RDN-Boost)
- MLN-Boost
- Regression
- Cost-sensitive SRL
- Learning with Advice
- Approximate Counting
- One-class Classification (coming soon)
- Discretization of Continuous Valued Attributes
- Lifted Relational Random Walks
- Grounded Relational Random Walks
Datasets
Applications of BoostSRL