ALGORITHMICALLY Detecting Autism From Forum Behavior

Page 5 of 5 [ 78 posts ]  Go to page Previous  1, 2, 3, 4, 5

cyberdad
Veteran
Veteran

User avatar

Joined: 21 Feb 2011
Age: 58
Gender: Male
Posts: 36,036

26 Oct 2011, 7:15 am

Verdandi wrote:
cyberdad wrote:
Strangely I am struggling to remember anyone on Wrong Planet who I would classify as obsessed about a topic or story. Perhaps Aspies are not that much different in this regard as we all have boundaries.


I am obsessed about autism.


Aren't we all...



Moog
Veteran
Veteran

User avatar

Joined: 25 Feb 2010
Age: 47
Gender: Male
Posts: 17,671
Location: Untied Kingdom

26 Oct 2011, 7:45 am

Surfman wrote:
Someone like Moog, who barely constructs a whole sentence


Surfman wrote:
i thankyou havben watch ingg a new seris off Numbers?


:P


_________________
Not currently a moderator


kfisherx
Veteran
Veteran

User avatar

Joined: 3 Nov 2010
Age: 62
Gender: Female
Posts: 1,192

26 Oct 2011, 8:15 am

swbluto wrote:
wavefreak58 wrote:
But an algorithm akin to a binary search is not going to happen.


I don't need a definite "yes" or "no", I'm fine with a "ranking order" like, say, you're socialization patterns suggest you are 60% autistic or... you're less autistic than 95% of the boards!


Uh... You mean like the AQ test or any of the other tests that are in the sticky above? :roll:

You have already been through those and you tested within the range which says you are either on the sprectrum or are BAP. That said, I also think you have other mental issues as well. What is your shrink saying about all this? Weren't you recently evauluated for something?



sunshower
Veteran
Veteran

User avatar

Joined: 17 Aug 2006
Age: 126
Gender: Female
Posts: 3,985

26 Oct 2011, 8:29 am

swbluto wrote:

Lol, if you haven't noticed, there's been no aspie like me before and most NTs do have this thing called "social insight" and there *might* be some social insight in my original post so, maybe, it could've been written by an NT. An iNTp to be more exact. :lol:

Of course, maybe not. Like I'd know. (All I know is that my cognition and social/language awareness seems to be autistic.)


Referring to the bolded, don't be so sure. You say you are different to other aspies because you have social insight. I have extremely highly developed social insight, probably more so than many of my NT friends (who come to me for advice) yet I am 100% certain I am Asperger's.

Adults with Asperger's can be more complicated than any stereotyped child presentation would ever suggest. I neatly fit into the stereotypical asperger's presentation as a child, but as an adult I am superficially indistinguishable from others for the most part. I consciously developed my own forms of social analysis and awareness, which are actually stronger than most peoples BECAUSE they were consciously and deliberately developed, plus obsessed over for years (a special interest).

I would never diagnose anyone over the internet with Asperger's, but the way you write and think swbluto definitely is suggestive of AS. I find what really distinguishes me from others is my sensory sensitivity, and my alien thought process - extreme in-depth analysis and logic driven decision making (discounting emotion).


_________________
Into the dark...


swbluto
Veteran
Veteran

User avatar

Joined: 26 Feb 2011
Age: 39
Gender: Male
Posts: 2,899
Location: In the Andes, counting the stars and wondering if one of them is home to another civilization

26 Oct 2011, 10:27 pm

Thanks to all who responded! I'm eager to hear more insights or opinions! :D



OJani
Veteran
Veteran

User avatar

Joined: 23 Feb 2011
Age: 52
Gender: Male
Posts: 2,505
Location: Hungary

27 Oct 2011, 6:30 am

Ok, you solicited. :)

The reason why it's very hard to even rank autistics is that you would have to be able to measure individual traits and then summarize them in one number, even if it would be intended only for ranking. It's not even clear what individual traits to identify. Traits that can be considered interesting, relevant, independent, indicative etc. Lots of statistical analyses, and what statistical data we have at hand? Even official diagnoses are not entirely reliable, as we all know.

In a current thread pedantic writing style is being discussed. I'd suggest starting a new thread about an algorithm that would rank posts and their respective authors by grammar and usage of "big" words, maybe for style, if you can program that. :P You'd have a pretty darn effective tool to spot Aspies. :D


_________________
Another non-English speaking - DX'd at age 38
"Aut viam inveniam aut faciam." (Hannibal) - Latin for "I'll either find a way or make one."


OJani
Veteran
Veteran

User avatar

Joined: 23 Feb 2011
Age: 52
Gender: Male
Posts: 2,505
Location: Hungary

27 Oct 2011, 7:33 am

Sticking to your original idea, first you'd have to download all threads, say, on the first three pages. I've done something similar before when I was after possible candidates on a dating portal. I downloaded, parsed data by usual string functions and stored it in a spreadsheet, automatically. You have to introduce fields for who's been quoted, four would be enough. The more delicate part would be determining the diagnostic status of occurrent members. You could go for implementing a simple diagnosed or not binomial approach, or a more sophisticated method derived from quiz scores taken from your previous thread or from the sticked one, making possible to create a weighing system, if you wish.


_________________
Another non-English speaking - DX'd at age 38
"Aut viam inveniam aut faciam." (Hannibal) - Latin for "I'll either find a way or make one."


Surfman
Veteran
Veteran

User avatar

Joined: 1 Aug 2010
Age: 63
Gender: Male
Posts: 3,938
Location: Homeward bound

27 Oct 2011, 8:25 am

OJani wrote:
Sticking to your original idea, first you'd have to download all threads, say, on the first three pages. I've done something similar before when I was after possible candidates on a dating portal. I downloaded, parsed data by usual string functions and stored it in a spreadsheet, automatically. You have to introduce fields for who's been quoted, four would be enough. The more delicate part would be determining the diagnostic status of occurrent members. You could go for implementing a simple diagnosed or not binomial approach, or a more sophisticated method derived from quiz scores taken from your previous thread or from the sticked one, making possible to create a weighing system, if you wish.


Wow that sounds like it would work.

The quiz makes diagnosis super easy, but without the benefit of a multi choice test, a heuristic could be applied incognito anywhere, including other forums, love letters and employer records.

A facial image or photo could be run through an aspie face recognition program for confirmation



Wrackspurt
Veteran
Veteran

User avatar

Joined: 17 Aug 2007
Age: 48
Gender: Female
Posts: 733

27 Oct 2011, 8:33 am

Writing (putting words down) is not genetic it's literally a trained behavior. So how could one determine if a person is on the spectrum based on their writing alone?



wavefreak58
Veteran
Veteran

User avatar

Joined: 26 Sep 2010
Age: 68
Gender: Male
Posts: 4,419
Location: Western New York

27 Oct 2011, 9:16 am

Wrackspurt wrote:
Writing (putting words down) is not genetic it's literally a trained behavior


The capacity to develop symbolic representations of reality is intrinsic to the structure of our brains and most assuredly linked to our genetics.

What is taught are the specific symbols and languages each culture ends up using.


_________________
When God made me He didn't use a mold. I'm FREEHAND baby!
The road to my hell is paved with your good intentions.


hyperlexian
Veteran
Veteran

User avatar

Joined: 21 Jul 2010
Age: 54
Gender: Female
Posts: 22,023
Location: with bucephalus

27 Oct 2011, 10:16 am

swbluto wrote:
Lol, if you haven't noticed, there's been no aspie like me before

there are lots of aspies like you on the board, who have almost identical patterns of communication. they doubt their diagnosis and seek to find ways they cannot possibly fit. they see other people on WP as evidence that they themselves could not be aspies.

if they DO come to terms with the diagnosis, they want to pick and choose which members get to be included in the aspie club. if someone is too social, too happy, too responsive to other member, has to much empathy, wears purple socks, or likes the smell of roses then they can't POSSIBLY be aspies too as those members are different from the OP.

it's kind of predictable to be honest, and not at all rare. all i can say is that you will not find definitive answers here as to whether you (or anyone else) is an aspie. only your medical professionals really know for sure.


_________________
on a break, so if you need assistance please contact another moderator from this list:
viewtopic.php?t=391105


Mdyar
Veteran
Veteran

User avatar

Joined: 28 May 2009
Age: 61
Gender: Male
Posts: 2,516

27 Oct 2011, 4:52 pm

Quote:
It could be OCD or something like schizophrenia. I was seeing some weird things just recently.


I missed this, and very sorry to hear that. If I was there in that state of mind, it certainly would evoke much anxiety, to believe that you are losing footing to this.

As far as an idea of developing an algorithm you would really need one, and one only: The content or theme of the query signifies the state of an impaired theory of mind, e.g. the detailed focus on all facets of " NT" behaviors, the constant 'driven' questions regarding the missed cues, or possible faux pas, or " what did this mean in that specific context" as in all the nuances and with all these permutations, adinfinitum.

OCD, schizophrenic folk, ADHD people, bipolar, etc. Don't engage in these dissections-- Sorry. This is what Zen was alluding to, and I believe miss. Lex, now.

This is really simple. No need for inputs on verbosity or supposed "Aspie" cognitive styles. It is true there are correlates to different degrees vs. the "NT." But by analysing grammar and aggrandized use of vocabulary on the board as indicators is really pointless. NT's write or can write with similar.

The bottom line here is nature of these theory of mind queries, or the fact that would you would even consider the "query" in and of itself.

It is really simple, but this idea may be hard to comprehend in and of itself to someone already impaired, hence you come up with " algorithms" and as in Mr. swbluto's case, this " math" is his comfort zone. There is nothing going to be" binary" or 0 1 "on" "off" determinant., as mentioned.
95 percent sure means you have 5 percent chance of a miss-- so what, and that is provided you could hone this in to this degree.

Get down to the basics here, and move ahead in harmony with the good doctor.

Hope all goes well.



Last edited by Mdyar on 30 Oct 2011, 10:36 pm, edited 1 time in total.

Surfman
Veteran
Veteran

User avatar

Joined: 1 Aug 2010
Age: 63
Gender: Male
Posts: 3,938
Location: Homeward bound

27 Oct 2011, 5:55 pm

Hes gotta be a freaken aspie



Wrackspurt
Veteran
Veteran

User avatar

Joined: 17 Aug 2007
Age: 48
Gender: Female
Posts: 733

28 Oct 2011, 9:36 am

wavefreak58 wrote:
Wrackspurt wrote:
Writing (putting words down) is not genetic it's literally a trained behavior


The capacity to develop symbolic representations of reality is intrinsic to the structure of our brains and most assuredly linked to our genetics.

What is taught are the specific symbols and languages each culture ends up using.


Eh, I still don't buy it. Only difference I can see is that someone on the spectrum can tend to ramble and talk in circles.. but then again anyone can do that.