tag cloud

[quant]

Amazon added "tag cloud" feature as well ...

http://www.amazon.co.uk/Volatility-S...-3#concordance

[janrif]

Originally posted by quant
Amazon added "tag cloud" feature as well ...

http://www.amazon.co.uk/Volatility-S...-3#concordance

Just curious, what practical use do you see for this in URp?

[quant]

Originally posted by janrif
Just curious, what practical use do you see for this in URp?

just what wordmuse already mentioned. Plus, it could be used for ranking search results by relevance (I hope)
http://www.kinook.com/Forum/showthre...&threadid=1133
As I already said, I think simple list with frequency would be enough (and much easier to implement), the graphical output doesn't add any further value to this feature imho, this is where I disagree with wordmuse who states that "there is something very powerful in the tag cloud"

[janrif]

Originally posted by quant
just what wordmuse already mentioned. Plus, it could be used for ranking search results by relevance (I hope)
http://www.kinook.com/Forum/showthre...&threadid=1133
As I already said, I think simple list with frequency would be enough (and much easier to implement), the graphical output doesn't add any further value to this feature imho, this is where I disagree with wordmuse who states that "there is something very powerful in the tag cloud"

OK, but here's where I get tripped up on this request. Mind you, I am nor against it, per se. I just don't get it.

In search engine results, the relevance is expressed in the order the results are displayed.

If I understand this request this gives user keyword statistics within each item returned by the search criteria.

So why do I care how many times the keyword is expressed in each item?

The only thing I would care about is if the list is ordered correctly, i..e the most relevant item displayrd @ the top of the list with some relevance statistic attribute which can be seen in a column in the search pane.

What am I missing?

[zargron]

Originally posted by janrif
OK, but here's where I get tripped up on this request ...
So why do I care how many times the keyword is expressed in each item? ...
What am I missing?

Well janrif, I suggest you take a step back from the hard core precise quantitive analysis stable for a moment and think of tag clouds as giving a fuzzy furry landscape picture of revealing. You are pondering where you might like to direct your future time and effort. Well, what have you been doing up until now? You could nut out a heavily numeric analysis to aid in your decision making. Or, you could embrace a more abstract approach. Imagine querying a tag cloud associated with a particular topic that you have stored in your UR database and sitting back after a couple of glasses of a nice red and a big fat doob staring at it. Comparing the big fat words with the little dimmed words. Wondering whether you like what it's saying to you. Do you like what the big fat bold words are saying about where you have been spending your time? Does a little dimmed word tickle your fancy? Does it sneak subtly into your subconscious as something you wish you were spending more time on? Whilst not listening to the words you are hearing come out from the mouth of a person you wish you weren't with, an explosion of ideas rambles through your brain as it reveals from whence you have been and where you should head.

I hope this helps.
Lord Zargron.

[janrif]

Originally posted by zargron [snip] I hope this helps.[/snip] Lord Zargron. [/B]

Lord Zargron, I'm a bit confused by the 'doob' reference.

As for the rest of your post, I've come to see the reason for our divergence of interest in this subject.

It boils down to perspective.

You see, at my age, I altready spend a great deal of time in the fuzzy, furry landscapes of my past, trying to uderstand all the dimmed words flowing thru my consciousness.

So at the end of the day, I appreciate the clarity of "hard core precise quantitive analysis" when I can find it.

[quant]

Originally posted by janrif
So why do I care how many times the keyword is expressed in each item?

The only thing I would care about is if the list is ordered correctly, i..e the most relevant item displayrd @ the top of the list with some relevance statistic attribute which can be seen in a column in the search pane.

I don't understand, you ask a question in the first paragraph, and answer yourself in the second.

"Tag cloud" and "relevance" features could be implemented at the same time.

Example:
Imagine you imported several documents dealing with some financial analysis, and put them in the "new documents" directory. Now you create search, tick "limit search to siblings" and tick the newly created "keyword frequency analysis" option. The search returns this list:

keyword frequency
dollar 1000
volatility 600
bond 300
market 200
price 100
...

By this search, I know what is the main topic the documents deal with. If I wanted to learn more about bonds, I'd click on "bond", this would perform the search on the same subset of documents, but with search string "bond", and the result would look like:

item frequency
Bond analysis.pdf 78
Capital markets.pdf 23
Fixed income products.pdf 20
...

So if I want to learn about bonds, the above documents would be probably the best candidates to start with ... without frequency, the search would return 20 documents that contain keyword "bond" maybe once, and I'd waste time opening each of them trying to see if I that is the document I need to learn about bonds

This is very primitive frequency analysis, easily implemented, ... maybe enough for the start. For the better search results, I decided to only link documents to UR, on which I can perform much more with SearchInform that includes morphology, exact, containing, phrase, ... search

[janrif]

Originally posted by quant
I don't understand, you ask a question in the first paragraph, and answer yourself in the second.

"Tag cloud" and "relevance" features could be implemented at the same time.

Example:
Imagine you imported several documents dealing with some financial analysis, and put them in the "new documents" directory. Now you create search, tick "limit search to siblings" and tick the newly created "keyword frequency analysis" option. The search returns this list:

keyword frequency
dollar 1000
volatility 600
bond 300
market 200
price 100
...

By this search, I know what is the main topic the documents deal with. If I wanted to learn more about bonds, I'd click on "bond", this would perform the search on the same subset of documents, but with search string "bond", and the result would look like:

item frequency
Bond analysis.pdf 78
Capital markets.pdf 23
Fixed income products.pdf 20
...

So if I want to learn about bonds, the above documents would be probably the best candidates to start with ... without frequency, the search would return 20 documents that contain keyword "bond" maybe once, and I'd waste time opening each of them trying to see if I that is the document I need to learn about bonds

This is very primitive frequency analysis, easily implemented, ... maybe enough for the start. For the better search results, I decided to only link documents to UR, on which I can perform much more with SearchInform that includes morphology, exact, containing, phrase, ... search

I understand the concept but backing up to the step before importing to UR, when I search the internet, the results can be presented by relevance, i.e. 99% listed first, etc.

I'd just link/import to the first few documents assuming these would be the most relevant documents to import into UR.

If, however, what I get from the internet is not sorted by relevance then your relevance list is extremely useful.

Again, I repeat, I am not aginst this improvement. I'm just having an 'intellectual' discussion here.

But I would say in the scheme of things that UR needs, I don't think clouds should be higher in priority than other things..... IMO a global search ability across all open databases like Zoot would be one those things.

[zargron]

Originally posted by janrif
I'm a bit confused by the 'doob' reference.

Herbs + paper + light = clouds.

Again, I repeat, I am not aginst this improvement. I'm just having an 'intellectual' discussion here.

But I would say in the scheme of things that UR needs, I don't think clouds should be higher in priority than other things..... IMO a global search ability across all open databases like Zoot would be one those things.

That's the main point isn't it janrif? I misinterpreted, or rather came down a bit heavy on your "... I just don't get it..." post. Tag clouds are good - it's simply a matter of scarce resource allocation. I agree with you, I would vote for plenty other suggestions before tag cloud.

[janrif]

Originally posted by zargron
Herbs + paper + light = clouds.
That's the main point isn't it janrif? I misinterpreted, or rather came down a bit heavy on your "... I just don't get it..." post. Tag clouds are good - it's simply a matter of scarce resource allocation. I agree with you, I would vote for plenty other suggestions before tag cloud.

No problem, Lord Zargron, I'm thoroughly enjoying this thread.

BTW, my first Bouvier des Flanders pup was named "Bogart" as in Don't bogart my doob....

[zargron]

Originally posted by janrif
BTW, my first Bouvier des Flanders pup was named "Bogart" as in Don't bogart my doob....

[wordmuse]

I've been away awhile... (grumble, grumble...)

As the originator of this thread, I'd like to weigh in again...

If all the information I have is self-created, then a tag cloud isn't all that useful to me.

But...

... if I import 20 pdf files, a dozen Word documents, etc. from a bunch of different sources, and I want to get the common threads without having to even scan the documents at all, a tag cloud seems to be the best way that I've seen available.

If someone has a way, using URP3 as it's currently developed, to let me have an idea of the information I've imported without doing visual inspections, I'd be grateful.

Again - my idea is to give me a very fast gander at the major themes of my imports.

I don't much care how I get to do this; I just want to do it; and a tag cloud seems like it could do the trick.

Savvy?

Regards,
Bal

[zargron]

Originally posted by wordmuse
I've been away awhile... (grumble, grumble...)

Me to - flat out like a lizard drinking at the moment...

If someone has a way, using URP3 as it's currently developed, to let me have an idea of the information I've imported without doing visual inspections, I'd be grateful.

Can't help you directly. However, I've had this idea in my someday maybe tray for ages where I export a branch from UR then run the exported data through a tool (custom or otherwise) that generates a Tag Cloud. If anyone wants to explore such an idea before I get to it then be my guest. It would take a lot of pressure off Kinook 'cause they could say "tag clouds from UR are possible - just a few clicks even if not built-into the product".

[janrif]

Originally posted by wordmuse
[snip] If someone has a way, using URP3 as it's currently developed, to let me have an idea of the information I've imported without doing visual inspections, I'd be grateful.[/snip]

Not automated but how about using the note pane for an identifying sentence or two.

[wordmuse]

Hi janrif,

nope - that won't let me do what I asked for, i.e., to get the basic themes of my imported documents without having to visually inspect them.

If I already knew the content and the themes, then I wouldn't need or ask for a tag cloud or something that would let me do what I'm after.

It's the visual inspection that's the time killer. I want to automate this, and I'd like to do so in the URP3 environment if possible.

Regards,
Bal