Mailing-List: contact user-help@ctakes.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@ctakes.apache.org
From: Stuti Awasthi <stutiawasthi@hcl.com>
To: "user@ctakes.apache.org" <user@ctakes.apache.org>
Subject: RE: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?
Thread-Topic: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?
Thread-Index: AdDaXIVzK6v17FHjRSCaSG6OV8sZsQBefolAAA1DehAAAGpSoAANMEmQAfD3QHU=
Date: Mon, 31 Aug 2015 16:05:41 +0000
Message-ID: 
 <2270F3695BF2C04DB9FE975FB546FF3C37813CF2@NDA-HCLC-MBS03.hclc.corp.hcl.in>
References: 
 <2270F3695BF2C04DB9FE975FB546FF3C377EE6A3@NDA-HCLC-MBS03.hclc.corp.hcl.in>
 <E084D8EFE2B03A408B324458C5212E94531255FD@CHEXMBX3B.CHBOSTON.ORG>
 <2270F3695BF2C04DB9FE975FB546FF3C377F347D@NDA-HCLC-MBS03.hclc.corp.hcl.in>
 <E5A9FA5ABBF1CA4085D4F0794852A51E52B2B1EE@CHEXMBX3B.CHBOSTON.ORG>
 <2270F3695BF2C04DB9FE975FB546FF3C377F3DEE@NDA-HCLC-MBS03.hclc.corp.hcl.in>,<E5A9FA5ABBF1CA4085D4F0794852A51E52B2B335@CHEXMBX3B.CHBOSTON.ORG>
In-Reply-To: <E5A9FA5ABBF1CA4085D4F0794852A51E52B2B335@CHEXMBX3B.CHBOSTON.ORG>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Thanks Guergana,=0A=
I agree that DocTime can be utilized for rough temporality but still the ch=
allenge remains for ordering the  such events with same doctime annotation.=
 For example the same example I sent in this thread, "The MRI scan on admis=
sion revealed an impending cord compression at the level of T10." ,All the =
Doctime for the events are identified as "BEFORE" but actually in this case=
, patient had the cord compression BEFORE MRI scan revealed this. It will b=
e tricky to handle such situations.. =0A=
=0A=
Anyways, I have gone through few sections of your document and it helped me=
 to clear my doubts. Thanks. I would defintly like to read the upcoming pap=
er.=0A=
=0A=
Thanks=0A=
Stuti Awasthi=0A=
=0A=
=0A=
________________________________________=0A=
From: Savova, Guergana [Guergana.Savova@childrens.harvard.edu]=0A=
Sent: Saturday, August 22, 2015 12:25 AM=0A=
To: user@ctakes.apache.org=0A=
Subject: RE: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?=0A=
=0A=
If I understand you correctly, your text does not have mentions of temporal=
 expressions but just events where one event is a container of many other e=
vents. If this is the case, the way to position this on a time line is to g=
et the DocTimeRel of the container (BEFORE, OVERLAPS, AFTER the document cr=
eation time) and position it accordingly. The events that it contains will =
be embedded in it. I agree that this provides rough temporality but this is=
 the best that could be done if no concrete dates are provided for events.=
=0A=
=0A=
More on narrative containers, etc. in this document, section 5:=0A=
http://clear.colorado.edu/compsem/documents/THYME%20Guidelines.pdf=0A=
=0A=
Is this making sense? Could you send an example?=0A=
--Guergana=0A=
=0A=
=0A=
-----Original Message-----=0A=
From: Stuti Awasthi [mailto:stutiawasthi@hcl.com]=0A=
Sent: Friday, August 21, 2015 8:42 AM=0A=
To: user@ctakes.apache.org=0A=
Subject: RE: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?=0A=
=0A=
Thanks Guergana,=0A=
=0A=
I have downloaded and build cTakes Timelanes at my end and explored it also=
. It is a really great tool and it shows that how much effort has been put =
to develop this.=0A=
One of the things which I noticed in Timelanes is if my input doesn't conta=
in any TimeEvent and only contains relation between Events-Events then no T=
imelane is plotted. What shall be done for such cases ?=0A=
=0A=
Thanks &Regards=0A=
Stuti Awasthi=0A=
=0A=
-----Original Message-----=0A=
From: Savova, Guergana [mailto:Guergana.Savova@childrens.harvard.edu]=0A=
Sent: Friday, August 21, 2015 5:57 PM=0A=
To: user@ctakes.apache.org=0A=
Subject: RE: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?=0A=
=0A=
Hi Stuti,=0A=
There is code in cTAKES (TimeLanes) that takes all pairwise temporal relati=
ons (e.g. CONTAINS, BEFORE, etc.) and the relation of the event to the docu=
ment creation time (docTimeRel) discovered by the cTAKES temporal modules a=
nd derives a timeline. The code to TimeLanes is in the apache ctakes sandbo=
x as "timelineViz" or something like that.  It is an older copy and it will=
 be updated soon - there are common libraries in TimeLanes and some other s=
oftware, and testing all end uses takes some effort. The code is not perfec=
t but it is a start.=0A=
=0A=
A paper describing the cTAKES temporal relation module will be published he=
re within couple of weeks:=0A=
https://urldefense.proofpoint.com/v2/url?u=3Dhttp-3A__jamia.oxfordjournals.=
org_content_early_recent_&d=3DBQIFAg&c=3DqS4goWBT7poplM69zy_3xhKwEW14JZMSdi=
oCoppxeFU&r=3DSeLHlpmrGNnJ9mI2WCgf_wwQk9zL4aIrVmfBoSi-j0kfEcrO4yRGmRCJNAr-r=
CmP&m=3D-lyFyfLLkAwmcn2CZAWQyyp8UCCXkkPIzUpCPA2x21I&s=3DGlu83TBTMbtxPT_X4eD=
tF2uLznglfyj0xuNykUvMSv0&e=3D=0A=
=0A=
You are welcome to use the TimeLanes code for your use case, improve on it =
and contribute back the improvements! Extracting temporality from text and =
creating timelines of events is a very challenging task.=0A=
--Guergana=0A=
=0A=
=0A=
-----Original Message-----=0A=
From: Stuti Awasthi [mailto:stutiawasthi@hcl.com]=0A=
Sent: Friday, August 21, 2015 2:03 AM=0A=
To: user@ctakes.apache.org=0A=
Subject: RE: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?=0A=
=0A=
Thanks Tim, I understood the difference clearly would like to device an app=
roach to utilize these concepts to prepare sequential event based timeline =
for the patients.=0A=
It would be helpful if you can point some useful links or suggest your thou=
ghts.=0A=
=0A=
Thanks &Regards=0A=
Stuti Awasthi=0A=
=0A=
-----Original Message-----=0A=
From: Miller, Timothy [mailto:Timothy.Miller@childrens.harvard.edu]=0A=
Sent: Wednesday, August 19, 2015 6:30 PM=0A=
To: user@ctakes.apache.org=0A=
Subject: Re: Meaning of "CONTAINS" TLink.. Is it equivalent to Overlap ?=0A=
=0A=
Sounds like an awesome application!=0A=
=0A=
CONTAINS is similar to OVERLAP but more specific. It means that one event i=
s fully contained inside an event or time expression:=0A=
=0A=
|--------- a -----------|=0A=
      |-----b-------|=0A=
=0A=
specifically BEGIN(a) <=3D BEGIN(b)  AND END(a) >=3D END(b)=0A=
=0A=
whereas OVERLAP is a broader category that just means there is some overlap=
:=0A=
=0A=
|--------a----------|=0A=
    |-----------b-------|=0A=
=0A=
is an example of an OVERLAP relation that is not contains.=0A=
=0A=
Technically CONTAINS is a subset of OVERLAP, but it is one that is so usefu=
l and common that we keep it as its own category.=0A=
=0A=
Tim=0A=
=0A=
=0A=
On 08/19/2015 04:53 AM, Stuti Awasthi wrote:=0A=
> Hi All,=0A=
>=0A=
> I am a newbee in ctakes so please pardon my ignorance.=0A=
> I am trying to extract temporal relation between Events-Events and Events=
-Time. For this I am using cTakes Temporal module via api's.  Once I get th=
e temporal relation, I would like to put them in sequence order to understa=
nd patients medical history.=0A=
>=0A=
> Now issue is for most of the relationship(Tlink) between Evenet-Event=0A=
> and Event-Time, I receive a category  "Contains". In most of the=0A=
> ctakes related papers, I read that TLInk belongs to "Before,=0A=
> Aftre,Overlap" so what does this "contains" means and how I can=0A=
> utilize this to build user's sequential timeline=0A=
>=0A=
> Eg:=0A=
> The MRI scan on admission revealed an impending cord compression at the l=
evel of T10.=0A=
>=0A=
> ################ EventEventRelation Annotation Output ###############=0A=
> EVENT(DocTime) | Relationship Category | Event(Doctime)=0A=
> scan ( BEFORE )  | CONTAINS             | revealed ( BEFORE )=0A=
> scan ( BEFORE )  | CONTAINS             | compression ( BEFORE )=0A=
>=0A=
> Please suggest.=0A=
>=0A=
> Thanks &Regards=0A=
> Stuti Awasthi=0A=
>=0A=
>=0A=
> ::DISCLAIMER::=0A=
> ----------------------------------------------------------------------=0A=
> ----------------------------------------------------------------------=0A=
> --------=0A=
>=0A=
> The contents of this e-mail and any attachment(s) are confidential and in=
tended for the named recipient(s) only.=0A=
> E-mail transmission is not guaranteed to be secure or error-free as=0A=
> information could be intercepted, corrupted, lost, destroyed, arrive=0A=
> late or incomplete, or may contain viruses in transmission. The e mail an=
d its contents (with or without referred errors) shall therefore not attach=
 any liability on the originator or HCL or its affiliates.=0A=
> Views or opinions, if any, presented in this email are solely those of=0A=
> the author and may not necessarily reflect the views or opinions of=0A=
> HCL or its affiliates. Any form of reproduction, dissemination,=0A=
> copying, disclosure, modification, distribution and / or publication of t=
his message without the prior written consent of authorized representative =
of HCL is strictly prohibited. If you have received this email in error ple=
ase delete it and notify the sender immediately.=0A=
> Before opening any email and/or attachments, please check them for viruse=
s and other defects.=0A=
>=0A=
> ----------------------------------------------------------------------=0A=
> ----------------------------------------------------------------------=0A=
> --------=0A=
>=0A=
>=0A=