• OK, it's on.
  • Please note that many, many Email Addresses used for spam, are not accepted at registration. Select a respectable Free email.

Statistical Analysis of intpf

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#51
Nice, mediocrity, just what I was going for

I can guess my word cloud: well, though, maybe, perhaps, guess, some, however, also, might, but, people

All my previous 500 posts are basically the same
that's thing about that TF-IDF algorithm – even if those are the words that comprise most of your posts, those words are not emphasized because they are common words among many of the posters.


Minuend and mine:
 

Attachments

Local time
Today, 13:00
Joined
Jan 1, 2009
Messages
3,773
#52
That's cheating:mad:

Though, I do have very strong opinions about digestion behavior that tend to be alienated
 

Polaris

Radioactive vision
Local time
Today, 01:00
Joined
Oct 13, 2009
Messages
2,026
#54
I’ve had the same thought as Minu. I sort of know what my cloud words would be in terms of topics I’ve been preoccupied in the last year. I don’t think it would be a nice cloud, more like looming doom and gloom in the form of negative words.

Funny how you can look at a word cloud and instantly feel like you’d rather not want to know that person. I think mine will be equally repellant. Kind of how I feel about people in general though, so no surprises there.
 

PmjPmj

Full of stars.
Local time
Today, 12:00
Joined
Sep 18, 2012
Messages
1,394
Location
UK
#57
My cloud gave me a bloody good chuckle. And reminded me that I need to try harder :<

Thank you, kind sir.
 
Local time
Today, 05:00
Joined
Oct 10, 2016
Messages
810
Location
Omnipresent
#58
I'm curious about mine now. The words for the year pic in the op was about what I expected.
 

baccheion

Active Member
Local time
Today, 07:00
Joined
May 2, 2016
Messages
224
#60
Everyone in the thread gets a cloud!
This is baccehion, gopher, gps, hado, haim, respectively
How many words have I written this year, how many unique words, and where would I rank if included in the original post? Can you run the same program on other forums?

Have you considered aggregating then LZMA-compressing posts, then sorting by compression ratio? Each post could be compressed to get an average for each member (can apply standard error), but a giant post seems more telling? I saw this used to rank songwriters. The less the compression ratio, the more "complex/rich" a person's posts. Including the overall number of posts would help filter the final list.
 

Niclmaki

Disturber of the Peace
Local time
Today, 07:00
Joined
Oct 21, 2012
Messages
390
Location
Canada
#61
Pretty cool. Do those emojis count as a word? Or are they ignored? Or get the weirdo formatting in their “count”:elephant:
 

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#64
How many words have I written this year, how many unique words, and where would I rank if included in the original post? Can you run the same program on other forums?

Have you considered aggregating then LZMA-compressing posts, then sorting by compression ratio? Each post could be compressed to get an average for each member (can apply standard error), but a giant post seems more telling? I saw this used to rank songwriters. The less the compression ratio, the more "complex/rich" a person's posts. Including the overall number of posts would help filter the final list.
Words written this year: 3482
Unique words: 1018
Unique dictionary words: 876
In terms of number of words written your rank is 68

The compression idea is interesting. Might look at that when I have some time.

The code can probably be run on other forums but it will probably need some modification. It depends on the html structure of the threads. Dunno how much similarity there is between forums in that regard.
 

Niclmaki

Disturber of the Peace
Local time
Today, 07:00
Joined
Oct 21, 2012
Messages
390
Location
Canada
#66
Nice, mediocrity, just what I was going for

I can guess my word cloud: well, though, maybe, perhaps, guess, some, however, also, might, but, people

All my previous 500 posts are basically the same





Shit and fuck were my first guesses. Not sure what to make of the milk and bowl, though. Hmmmm

What is your word cloud, serac?
500 posts! Heck all you guys post a lot. I never really looked at the post count before.
 

baccheion

Active Member
Local time
Today, 07:00
Joined
May 2, 2016
Messages
224
#67
Words written this year: 3482
Unique words: 1018
Unique dictionary words: 876
In terms of number of words written your rank is 68

The compression idea is interesting. Might look at that when I have some time.

The code can probably be run on other forums but it will probably need some modification. It depends on the html structure of the threads. Dunno how much similarity there is between forums in that regard.
What about the second graph (unique words)? I'm more interested in the density of unique words (something like unique_dictionary_words / total_dictionary_words) and the resulting graph/rank. If you end up trying the compression approach I mention, that would address everything.
 

Nebulous

fungus lord. dishwasher opossum
Local time
Today, 07:00
Joined
Mar 11, 2016
Messages
891
Location
Just North of Normal
#68
Random one of the day: Nebulous
NEAT
“Daydreaming” not surprised

This stuff’s so coolllllllll I love this kind of thing so muccchhhhh
 

crippli

disturbed
Local time
Today, 13:00
Joined
Jan 15, 2008
Messages
1,650
#69
Can you match clouds, as to find who on the site would be best to sleep with?
 
Local time
Today, 05:00
Joined
Apr 4, 2010
Messages
5,846
Location
subjective
#70
Can you match clouds, as to find who on the site would be best to sleep with?
terrific idea.

group people with similarities in a high dimension graph of complex word similitude.

Maybe discover a persons MBTI type as well.
 
Local time
Today, 23:00
Joined
Apr 26, 2013
Messages
1,194
Location
Shallow grave
#71
Dang I gotta know. Hit me plz.
 

QuickTwist

Alive - Born Anew
Local time
Today, 06:00
Joined
Jan 24, 2013
Messages
6,922
Location
...
#72
According to my psychological profile, I am a 40-year-old female INTJ.

https://applymagicsauce.com/demo.html

Also includes the big five.
Interesting shit.

It says I am a 25-29 yo Male who is very unlikely to be gay who is an INFP. I think my Big 5 was:

Liberal and Artistic (O): 57%
Organized and Hardworking (C): 53% lol
Contemplative (E): 34%
Team Working and Trusting (A): 51%
Laid Back and Relaxed (N): 42%

Thanks for doing this Serec. That is quite a project you had there. How long did it take you to complete this?

My Title seems to fit based on this so I think I will keep it for the time being.

I was going to go to sleep, look what you made me do...

Can you do my word salad Serec? I don't really use big words tho.
 

Rixus

I introverted think. Therefore, I am.
Local time
Today, 12:00
Joined
Nov 21, 2016
Messages
1,276
Location
United Kingdon
#73
According to my psychological profile, I am a 40-year-old female INTJ.

https://applymagicsauce.com/demo.html

Also includes the big five.
Apparently, my digital profile suggests I'm a 29 year old INTP, Single, Morman, Conservative Female with a strong interest in art and is highly unlikely to be gay.

And apparently, liking rock music and a couple of fitness pages makes me less intellectual. I'm also apparently quite unsatisfied with life.
 

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#75
That is quite a project you had there. How long did it take you to complete this?
No too long tbh. A sunday afternoon for the code that retrieves the threads and did the cloud thing while waiting for some computations to finish at work
 
Local time
Today, 23:00
Joined
Apr 26, 2013
Messages
1,194
Location
Shallow grave
#76
Haha that was fun. Thanks
 

PmjPmj

Full of stars.
Local time
Today, 12:00
Joined
Sep 18, 2012
Messages
1,394
Location
UK
#78
The test AK linked has me as a 32 year old male INTJ.

Accurate test is accurate. I mean sure, I'm probably an E rather than an I, but it had me close to borderline anyway.

Spooky shit bruh.
 

PmjPmj

Full of stars.
Local time
Today, 12:00
Joined
Sep 18, 2012
Messages
1,394
Location
UK
#79
Oh, wait. I analysed some emails and it now thinks I'm 26/f/INTP.

Hawt.
 

Rixus

I introverted think. Therefore, I am.
Local time
Today, 12:00
Joined
Nov 21, 2016
Messages
1,276
Location
United Kingdon
#80
Oh, wait. I analysed some emails and it now thinks I'm 26/f/INTP.

Hawt.
Ah, fellow INTP female in their late 20's. Just like me.
Are you a Conservative Mormon, as well?
 

QuickTwist

Alive - Born Anew
Local time
Today, 06:00
Joined
Jan 24, 2013
Messages
6,922
Location
...
#83
Serec, I have a favor to ask. I want to take the results from the cloud you did for me and plug it into IBM watson's personality utility and see what I get and see if it's any different.
 

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#84
Serec, I have a favor to ask. I want to take the results from the cloud you did for me and plug it into IBM watson's personality utility and see what I get and see if it's any different.
What sort of format does it take as input? The word clouds are based on giving each word a numerical weight
 

Rixus

I introverted think. Therefore, I am.
Local time
Today, 12:00
Joined
Nov 21, 2016
Messages
1,276
Location
United Kingdon
#85
Why don't words like "The", "And", "It" and so on come up?
I managed to get a "her" in there on mine. You'd think those simple pronouns and conjunctions words would litter our posts more so than the other words we find.

Though why "piles" is in my cloud, I'm not entirely sure.
 

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#86
Why don't words like &quot;The&quot;, &quot;And&quot;, &quot;It&quot; and so on come up?
I managed to get a &quot;her&quot; in there on mine. You'd think those simple pronouns and conjunctions words would litter our posts more so than the other words we find.

Though why &quot;piles&quot; is in my cloud, I'm not entirely sure.
It's the TF-IDF algorithm. It deflates the weights of the words by how often they are used in the total collection of words, i.e. across all posts on the forum. For example since everyone uses "the" very frequently, its importance gets diminished.

It is surprising that "her" shows up in your cloud, but it just means you're using that word much more frequently than everyone else on the forum.
 

QuickTwist

Alive - Born Anew
Local time
Today, 06:00
Joined
Jan 24, 2013
Messages
6,922
Location
...
#87
What sort of format does it take as input? The word clouds are based on giving each word a numerical weight
I think all it does is take the words you use and put them into different categories. Therefore, I don't think I actually need to write sentences, but can just enter in how many times of each word from the word cloud I used. So having a rundown of how many times I used each word in the word cloud you did for me should suffice.
 

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#88
I think all it does is take the words you use and put them into different categories. Therefore, I don't think I actually need to write sentences, but can just enter in how many times of each word from the word cloud I used. So having a rundown of how many times I used each word in the word cloud you did for me should suffice.
Something like this? These are the weights of each word
 

Attachments

QuickTwist

Alive - Born Anew
Local time
Today, 06:00
Joined
Jan 24, 2013
Messages
6,922
Location
...
#89
Something like this? These are the weights of each word
Thanks, I appreciate that. Unfortunately, I really have no idea how many times this equates to with each word and that is what I need.
 

Serac

A menacing post slithers
Local time
Today, 12:00
Joined
Jun 7, 2017
Messages
1,629
Location
Stockholm
#90
Thanks, I appreciate that. Unfortunately, I really have no idea how many times this equates to with each word and that is what I need.
Well, I imagine you don't need the actual count, but just the relative count? In that case just multiply each weight with a scaling factor, say, 10000, and you get something akin to a relative count

Also note that these are not the actual frequencies but weights generated by TF-IDF algorithm.
 
Top Bottom