"Text-classification: Data from XML and multiple keywords"

tron42

Hi,

I'm new to Rapidminer and I want to use it for some text-classification. I've two questions:

1) All my data is stored in XML. Is it possible to import the data as XML (so I don't need to transform it into csv)?

2) I'm building a small csv-example for testing, something like:
   title;abstract;keyword;keyword;keyword;.....

As you can see I have multiple columns, each with one keyword. Is it possible to mark more than one column as an label? I tried, but when I change the next column, the previous is changing back.

I hope you understand my questions, my english is not the best.

Thanks and regards,
tron42

daktari999

Hi!

I just installed RapidMiner and wanted to start working on it with some test data that is also in XML. I also haven't found a way to import my files to RapidMiner. It it really not possible to do so?

el_chief

Generally speaking, RapidMiner works with flat-file data. The same as almost all other statistical software.

XML is hierarchical by nature, so it is hard to say how this would work.

You could try reading in the file as HTML and using XPATH to get the attribute values, but it is probably easiest to convert to CSV or Excel first.

As for the LABEL question, a LABEL in rapidminer is like a Y variable in a regression. It is the thing you are trying to predict.

Neil

tron42

Thanks Neil!

Transforming XML into CSV is not difficult, so ok. Thought that I could save a step.
Maybe in the next version of Rapidminer a build-in-plugin or something

Because of the label, I should paraphrase the question:

Is it possible to do MULTI-LABEL CLASSIFICATION with Rapidminer?
I want to predict more than one keyword for each text.

Thanks a lot and Regards!

Andrew2

Hello tron42

There is a multi label looping operator. Here's an example that uses it.

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<process version="5.0">
  <context>
    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="5.0.11" expanded="true" name="Process">
    <process expanded="true" height="656" width="815">
      <operator activated="true" class="generate_data" compatibility="5.0.11" expanded="true" height="60" name="Generate Data" width="90" x="45" y="30">
        <parameter key="target_function" value="sum classification"/>
      </operator>
      <operator activated="true" class="generate_attributes" compatibility="5.0.11" expanded="true" height="76" name="Generate Attributes" width="90" x="45" y="120">
        <list key="function_descriptions">
          <parameter key="label1" value="att1+att2"/>
          <parameter key="label2" value="att1+att3+att4"/>
          <parameter key="label3" value="abs(att1*att2)"/>
        </list>
      </operator>
      <operator activated="true" class="discretize_by_bins" compatibility="5.0.11" expanded="true" height="94" name="Discretize" width="90" x="45" y="210">
        <parameter key="attribute_filter_type" value="subset"/>
        <parameter key="attributes" value="label1|label2|label3"/>
        <parameter key="range_name_type" value="short"/>
      </operator>
      <operator activated="true" class="set_role" compatibility="5.0.11" expanded="true" height="76" name="Set Role" width="90" x="45" y="345">
        <parameter key="name" value="label1"/>
        <parameter key="target_role" value="label1"/>
      </operator>
      <operator activated="true" class="set_role" compatibility="5.0.11" expanded="true" height="76" name="Set Role (2)" width="90" x="45" y="435">
        <parameter key="name" value="label2"/>
        <parameter key="target_role" value="label2"/>
      </operator>
      <operator activated="true" class="set_role" compatibility="5.0.11" expanded="true" height="76" name="Set Role (3)" width="90" x="45" y="525">
        <parameter key="name" value="label3"/>
        <parameter key="target_role" value="label3"/>
      </operator>
      <operator activated="true" class="select_attributes" compatibility="5.0.11" expanded="true" height="76" name="Select Attributes" width="90" x="246" y="255">
        <parameter key="attribute_filter_type" value="subset"/>
        <parameter key="attributes" value="att1|att2|att3|att4|att5|label1|label2|label3"/>
        <parameter key="include_special_attributes" value="true"/>
      </operator>
      <operator activated="true" class="loop_labels" compatibility="5.0.11" expanded="true" height="76" name="Loop Labels" width="90" x="447" y="255">
        <process expanded="true" height="616" width="945">
          <operator activated="true" class="x_validation" compatibility="5.0.0" expanded="true" height="112" name="Validation" width="90" x="313" y="120">
            <description>A cross-validation evaluating a decision tree model.</description>
            <process expanded="true" height="654" width="466">
              <operator activated="true" class="naive_bayes" compatibility="5.0.11" expanded="true" height="76" name="Naive Bayes" width="90" x="188" y="30"/>
              <connect from_port="training" to_op="Naive Bayes" to_port="training set"/>
              <connect from_op="Naive Bayes" from_port="model" to_port="model"/>
              <portSpacing port="source_training" spacing="0"/>
              <portSpacing port="sink_model" spacing="0"/>
              <portSpacing port="sink_through 1" spacing="0"/>
            </process>
            <process expanded="true" height="654" width="466">
              <operator activated="true" class="apply_model" compatibility="5.0.0" expanded="true" height="76" name="Apply Model" width="90" x="45" y="30">
                <list key="application_parameters"/>
              </operator>
              <operator activated="true" class="performance" compatibility="5.0.0" expanded="true" height="76" name="Performance" width="90" x="179" y="30"/>
              <connect from_port="model" to_op="Apply Model" to_port="model"/>
              <connect from_port="test set" to_op="Apply Model" to_port="unlabelled data"/>
              <connect from_op="Apply Model" from_port="labelled data" to_op="Performance" to_port="labelled data"/>
              <connect from_op="Performance" from_port="performance" to_port="averagable 1"/>
              <portSpacing port="source_model" spacing="0"/>
              <portSpacing port="source_test set" spacing="0"/>
              <portSpacing port="source_through 1" spacing="0"/>
              <portSpacing port="sink_averagable 1" spacing="0"/>
              <portSpacing port="sink_averagable 2" spacing="0"/>
            </process>
          </operator>
          <connect from_port="example set" to_op="Validation" to_port="training"/>
          <connect from_op="Validation" from_port="averagable 1" to_port="out 1"/>
          <portSpacing port="source_example set" spacing="0"/>
          <portSpacing port="sink_out 1" spacing="0"/>
          <portSpacing port="sink_out 2" spacing="0"/>
        </process>
      </operator>
      <connect from_op="Generate Data" from_port="output" to_op="Generate Attributes" to_port="example set input"/>
      <connect from_op="Generate Attributes" from_port="example set output" to_op="Discretize" to_port="example set input"/>
      <connect from_op="Discretize" from_port="example set output" to_op="Set Role" to_port="example set input"/>
      <connect from_op="Set Role" from_port="example set output" to_op="Set Role (2)" to_port="example set input"/>
      <connect from_op="Set Role (2)" from_port="example set output" to_op="Set Role (3)" to_port="example set input"/>
      <connect from_op="Set Role (3)" from_port="example set output" to_op="Select Attributes" to_port="example set input"/>
      <connect from_op="Select Attributes" from_port="example set output" to_op="Loop Labels" to_port="example set"/>
      <connect from_op="Loop Labels" from_port="out 1" to_port="result 1"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
    </process>
  </operator>
</process>

Maybe this would help get you started.

regards

Andrew

tron42

Thank's a lot Andrew!

I will check the example out and report my progress when I'm done with my mid-term exams this week.

Thanks and regards!

daktari999

RapidMiner already has an built-in XPath-Batch-Worker. It's under Uitility -> Files-> Split File by Content. This plugin also supports Regular expressions.

Worked perfectly in my case

SebastianLoh

Hi all there,

I just wanted to add this to the discussion.

This is the basic setup to extract information from XML files. The Extract Information operator also allows to perform XPath queries on XML files. The result is then stored in the defined attribute.

It's just a matter of taste, whether you want to do this with RM or build a proper CSV/Excel/... beforehand. Althought with RM you can easily add another feature.

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<process version="5.0">
  <context>
    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="5.0.8" expanded="true" name="Process">
    <process expanded="true" height="370" width="547">
      <operator activated="true" class="text:process_document_from_file" compatibility="5.0.6" expanded="true" height="76" name="Process Documents from Files" width="90" x="179" y="120">
        <list key="text_directories"/>
        <parameter key="extract_text_only" value="false"/>
        <process expanded="true" height="590" width="912">
          <operator activated="true" class="text:extract_information" compatibility="5.0.6" expanded="true" height="60" name="Extract Information" width="90" x="246" y="75">
            <parameter key="query_type" value="XPath"/>
            <list key="string_machting_queries"/>
            <list key="regular_expression_queries"/>
            <list key="regular_region_queries"/>
            <list key="xpath_queries">
              <parameter key="att1" value="some xpath query"/>
              <parameter key="att2" value="another xpath query"/>
            </list>
            <list key="namespaces"/>
            <list key="index_queries"/>
          </operator>
          <connect from_port="document" to_op="Extract Information" to_port="document"/>
          <connect from_op="Extract Information" from_port="document" to_port="document 1"/>
          <portSpacing port="source_document" spacing="18"/>
          <portSpacing port="sink_document 1" spacing="0"/>
          <portSpacing port="sink_document 2" spacing="0"/>
        </process>
      </operator>
      <connect from_op="Process Documents from Files" from_port="example set" to_port="result 1"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
    </process>
  </operator>
</process>

Regarding the multi label problem:

You can learn a model for each label. Therefor you need to set the current label attribute to the role "label" and the other labels to the role "other" (just not regular, because then the current label is also learned on the other labels).

Ciao Sebastian

ps. thanks Matthias! I modified the example above.

colo

Hi,

just a little remark regarding the basic example process. The parameter "extract text only" ("Process Documents" operator) needs to be disabled if following XPath queries shall deliver something. This little details can be missed quickly and perhaps lead to some confusion about not working XPath statements. Shame on you, Sebastian

Regards,
Matthias

tron42

Sebastian Loh wrote:
Regarding the multi label problem: You can learn a model for each label. Therefor you need to set the current label attribute to the role "label" and the other labels to the role "other" (just not regular, because then the current label is also learned on the other labels). Ciao Sebastian ps. thanks Matthias! I modified the example above.

Hi,

thanks for your response!
I'm not sure that I understand that right. Lets say I have three keywords, therefore I have three labels. I would set the first keyword as a label and the other two as "other". I start the trainig and I will get a model. After that I would set the second keyword as a label and the other two as "other" (the first and the third keyword). Than I would restart the training and get a new model (a second)? Or will the first model be updated?

Sorry I'm new to rapidminer and I'm not really familiar with the tool yet.

Regards,
tron42

SebastianLoh

Hi tron42,

can you explain your intention again please? What I understood from

tron42 wrote:
2) I'm building a small csv-example for testing, something like:    title;abstract;keyword;keyword;keyword;..... As you can see I have multiple columns, each with one keyword. Is it possible to mark more than one column as an label? I tried, but when I change the next column, the previous is changing back.

is, that each keyword is an indicator/label. For example keyword1 indicates the sentiment good/bad review for quality, keyword2 indicates good/bad review for service, keyword3 for....

So then you learn on the attribute "abstract" (which you need to process with the Textprocessing operators, of course, Process Documents, and inside at least tokenization and possibly some Stopword Filter and Filter by legth) one classifiaction model for "quality", one for "service", and so on.

However, you seem to have something different in you mind.

Ciao Sebastian

tron42

Sebastian Loh wrote:
Hi tron42, can you explain your intention again please? What I understood from is, that each keyword is an indicator/label. For example keyword1 indicates the sentiment good/bad review for quality, keyword2 indicates good/bad review for service, keyword3 for....

Hi Sebastian,

each keyword is not only an indicator, it describes the text. For example I have a text about China, so the keywords are: china, asia, hongkong, north korea, ... and a lot more keywords which characterises the article. I want to train those relationships between the text and keywords, so that I can predict possible keywords for an unknown text.

Regards,
David

rakirk

I've been doing something similar to tron42 in that I want to process XML using XPath and Extract Information operator. I am using an XPath to query every node and it only returns the first result. My problem is that I want to extract all of the elements from a particular document and it seems like Extract Information terminates after first discovering an element that matches the XPath query. I will show a simple example of an XML file and then the process being used.

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<process version="5.1.002">
  <context>
    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="5.1.002" expanded="true" name="Process">
    <process expanded="true" height="431" width="547">
      <operator activated="true" class="text:process_document_from_file" compatibility="5.1.001" expanded="true" height="76" name="Process Documents from Files (3)" width="90" x="179" y="255">
        <list key="text_directories">
          <parameter key="SGML" value="C:\Users\Kirk\Desktop\tests"/>
        </list>
        <parameter key="extract_text_only" value="false"/>
        <parameter key="encoding" value="UTF-8"/>
        <parameter key="create_word_vector" value="false"/>
        <parameter key="prune_below_absolute" value="5"/>
        <parameter key="prune_above_absolute" value="1000000"/>
        <process expanded="true" height="650" width="710">
          <operator activated="true" class="text:extract_information" compatibility="5.1.001" expanded="true" height="60" name="Extract Information (2)" width="90" x="45" y="210">
            <parameter key="query_type" value="XPath"/>
            <list key="string_machting_queries">
              <parameter key="intro_m/d" value="&lt;intro_m\.*&gt;.&lt;/intro\.*&gt;"/>
            </list>
            <list key="regular_expression_queries"/>
            <list key="regular_region_queries"/>
            <list key="xpath_queries">
              <parameter key="Move 1" value="//title"/>
            </list>
            <list key="namespaces"/>
            <parameter key="assume_html" value="false"/>
            <list key="index_queries"/>
          </operator>
          <connect from_port="document" to_op="Extract Information (2)" to_port="document"/>
          <connect from_op="Extract Information (2)" from_port="document" to_port="document 1"/>
          <portSpacing port="source_document" spacing="0"/>
          <portSpacing port="sink_document 1" spacing="0"/>
          <portSpacing port="sink_document 2" spacing="0"/>
        </process>
      </operator>
      <connect from_op="Process Documents from Files (3)" from_port="example set" to_port="result 1"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
    </process>
  </operator>
</process>

Here is the XML example:

<?xml version="1.0" encoding="ISO-8859-1"?>

<bookstore>

<book category="COOKING">
  <title lang="en">Everyday Italian</title>
  <author>Giada De Laurentiis</author>
  <year>2005</year>
  <price>30.00</price>
</book>

<book category="CHILDREN">
  <title lang="en">Harry Potter</title>
  <author>J K. Rowling</author>
  <year>2005</year>
  <price>29.99</price>
</book>

<book category="WEB">
  <title lang="en">XQuery Kick Start</title>
  <author>James McGovern</author>
  <author>Per Bothner</author>
  <author>Kurt Cagle</author>
  <author>James Linn</author>
  <author>Vaidyanathan Nagarajan</author>
  <year>2003</year>
  <price>49.99</price>
</book>

<book category="WEB">
  <title lang="en">Learning XML</title>
  <author>Erik T. Ray</author>
  <year>2003</year>
  <price>39.95</price>
</book>

</bookstore>

Results:
<title lang="en">Everyday Italian</title>

Desired results: (4 separate examples)
<title lang="en">Everyday Italian</title>
<title lang="en">Harry Potter</title>
<title lang="en">XQuery Kick Star</title>
<title lang="en">Learning XML</title>

land

Hi,
you will have to use the Cut Document operator together with the XPath querry to get all matches as documents in the inner subprocess of Cut Document.

Greetings,
  Sebastian