"How to visualize textmining results"

maria_godric

Hi,
How to visualize text mining results.Which operator can be used for that
Thanks
Maria

land

Hi,
what exactly do you mean by "text mining results"? Usually you have only one outcome per text, which simply is noted in the table as prediction.

Greetings,
Sebastian

maria_godric

Hi,
Using other Text Mining operators i got TFIDF.Is there any other RM operators to graphically represent the words and its TFIDF.
Thanks
Maria

land

Hi,
in the new TextProcessing Extension of RapidMiner 5.0 there you will be able to visualize the word list with the values of occurrences with the rich selection of plotters available in RapidMiner.

If you want to see the source document of a word vector, you might use the TextExampleVisualizer. Whereever you click on an example in the plotter, you will see the original text.

Greetings,
Sebastian

derchief

Hi,

"you will be able to visualize the word list with the values of occurrences with the rich selection of plotters available"

I tried a lot of combinations but couldn´t get any list not to mention a visualisation out of the wordlist I tried to create with the following code:

<参数键=“记录文件”值= " C: \ Dokumente和Einstellungen\cniemann\Eigene Dateien\bla.log"/>

<过程扩展= " true "高度= " 521" width="681">

Where is the failure?

Regards, chris

land

Hi Chris,
you are definitively right. I forgot that the count's aren't shown when you don't have a label. In fact this is a small bug that I have fixed now. Unfortunately the latest update of the Text Extension was already rolled out, so it didn't made it in time. But there's a workaround:
Simply add a label by inserting a generate attribute operator and setting this attribute to role label:

The following process will show you, how you can turn the word list into an example set and use arbitrary plotter:

<参数键=“记录文件”值= " C: \ Dokumente和Einstellungen\cniemann\Eigene Dateien\bla.log"/>

<过程扩展= " true "高度= " 521" width="748">

With the next update the displaying bug will be removed and the table shows all columns that are there but are currently not shown.
Following the normal update cycle, it will be update in around a month again, if no enterprise customer demands an earlier update.

Greetings,
Sebastian

derchief

Hello again,

you said, "you can turn the word list into an example set and use arbitrary plotter". In my results workplace, there is just one tab (result overview). Only the metadata of the example set from the "wordlisttodata" is shown there with: Role, Name, Type, Range etc., but there is no wordlist. In screenshots of RM I´ve seen tabs like "ExampleSet(Retrieve)". I guess I could find the results in such tabs normally, but in my case they are not shown. In your example you said I should "add a label by inserting a generate attribute operator and setting this attribute to role label". The generate attribute operator demands a list entry with an attribute name (for what?) and a function expression. I don´t know what I should paste here?!

best regards
chris

land

Hi Chris,
sorry I forgot to add example values to the parameter list:

The Attribute name will be the name of the attribute that is added to your example set. The Expression is evaluated for each example. Please take a look at the operator documentation for details on this. (In the Help view) Here the expression comes down to just enter a constant: For example something like I entered in my example above. This String will be set as value of each example in the new attribute column "label".

Greetings,
Sebastian