Example of how Jaccard coefficient is computed when using sentences as unit of analysis.
A: Number of sentences that contain both word X and Y.
B: Number of sentences that contain word X only.
C: Number of sentences that contain word Y only.
Jaccard coef. of X-Y = A / ( A + B + C )
N 47, E 60, D .056 - these appeared at the bottom of the co-occurrence network
N: number of words in the network
E: number of edges in the network
D: number of edges / possible max number of edges (density of the network)
stress = 0.360 - this appeared at the bottom of the MDS
The stress is MDS specific value that indicates one aspect of how MDS is performed. When you show your MDS results, it would be better to show stress value too.
Thanks. It helps.
In regard to stress = 0.360, what does 0.360 mean? Hows is it different for example stress = 0.20, if ever there is stress like this.
Thanks.
Dan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
View and moderate all "Open Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Discussion"
Dear Ko-ichi,
Since you are very receptive to questions, I's like to take this opportunity to ask more questions and seek further assistance.
2.1 N 47, E 60, D .056 - these appeared at the bottom of the co-occurrence network of my text;
unit of analysis was "sentence";
2.2 stress = 0.360 - this appeared at the bottom of the MDS graph?
Can I obtain the Jaccard Matrix of the words using the software. If yes, how?
How to obtain the frequency matrix of co-occurring words, for example:
bomb - threat = 30
bomb - hoaxe = 20
crime- map = 11?
Thank you.
Dan
Dear Dan,
A: Number of sentences that contain both word X and Y.
B: Number of sentences that contain word X only.
C: Number of sentences that contain word Y only.
Jaccard coef. of X-Y = A / ( A + B + C )
N: number of words in the network
E: number of edges in the network
D: number of edges / possible max number of edges (density of the network)
The stress is MDS specific value that indicates one aspect of how MDS is performed. When you show your MDS results, it would be better to show stress value too.
Does this thread help?
https://sourceforge.net/p/khc/discussion/222396/thread/3a7e094d/
Best,
Last edit: HIGUCHI Koichi 2014-06-03
View and moderate all "Open Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Discussion"
Koichi,
Thanks. It helps.
In regard to stress = 0.360, what does 0.360 mean? Hows is it different for example stress = 0.20, if ever there is stress like this.
Thanks.
Dan
Dear Dan,
The stress value is the smaller the better.
For more details, please consult literature.
BTW, can anybody recommend a good introductory literature about multidimensional scaling written in English??
Last edit: HIGUCHI Koichi 2014-06-14
View and moderate all "Open Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Discussion"
A simple on regarding numbers found on graphs, could you please clarify what is the meaning of the number to the left:
http://khc.sourceforge.net/en/gallery/pages/image/imagepage_ctp1.html
Is it percentage of all words or sentences? If a sentence includes a "positive" and "aggression" code how would it count it?
Thank you
Last edit: Anonymous 2014-08-18
In chapter I, about 15% of the sentences contain "agression."
The line graph in the screenshot represents the table in this one.
http://khc.sourceforge.net/en/gallery/pages/image/imagepage21.html
The sentence which contains both codes will be counted for both codes. It will increase both percentages.
Last edit: HIGUCHI Koichi 2015-06-25