MALLET can be downloaded at http://mallet.cs.umass.edu/download.php/. There are instructions for installation on the page linked above. MALLET will be downloaded into your downloads folder. Once downloaded you should move the folder(which should be named mallet-2.0.8) onto your desktop. The next step will be transforming your XML files into a format that can be inputted into MALLET.
Prior to Topic Modeling I would advise having a corpus of of XML that can all be validated by a single schema(so that you can write XSLT that can target the text that you want to analyze easily). MALLET cannot work with XML files; it only works with an input of .txt documents. In order to save time you will use Oxygen to run a batch transformation on your XML files in order to convert them all into indivdiual .txt files inside of a single directory. To do this we will write XSLT that targets the XML node that stores the text that you want to extract. I will show the process that I used to break down my input files.
<letter>
<head><location>Nuenen</location>, on or about Sunday, <date>20 July 1884</date>.</head>
<salutation>My dear Theo,</salutation>
<body>
<p><good_health>I was delighted to learn from your letter to Pa and Ma that you plan to go to London</good_health>
on 4 Aug. and then to come on here from there. I’m again looking forward very much
to your arrival and to finding out what you’ll think of the work that I’ve done
since. <work origin="existing" stage="finished">The last things I did are a couple of rather large studies of oxcarts, a
black ox and a red and white one.</work></p>
<p><work origin="new" stage="middle">And have also been working again on the old tower in the fields in the evening; I’ve
made a larger study of it than my previous ones — with the wheatfields around
it.</work></p>
<p>Rappard sent me back the little book by Vosmaer that belongs to you — I started to
read it but — is it just me? — find it almighty boring and actually written in an
academic, sermonizing tone. Perhaps you will too when you look at it again. Have
you read Sapho by Daudet? <unstress reason="environment">It’s very beautiful, and so vigorous, and so close to
life that the female figure lives, breathes, and one can hear, literally hear the
voice, and forgets that one is reading.</unstress></p>
<p>You’ll also <work stage="finished" origin="new">see a couple more new weavers </work>when you come.</p>
<p>Nature is certainly pure here — I’m still very pleased with the studio, too.We must
visit some farms and weavers together when you come.</p>
<p>Rappard’s plan is to come back again in October; he’s probably in Drenthe again
now.</p>
<p>Well, I write in some haste because I’m hard at work. <good_health>I work a good deal early in the
morning or in the evening, and then sometimes everything is so inexpressibly
beautiful.</good_health></p>
<p>Regards, believe me</p>
</body>
<close>Ever yours,</close>
<sig>Vincent</sig>
</letter>All of the information that I wish to extract is nested inside of the <p> element. So to extract that infromation I wrote
XSLT to target that node and strip it of all markup.
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:xs="http://www.w3.org/2001/XMLSchema" exclude-result-prefixes="xs" version="3.0">
<xsl:output method="text" indent="yes" doctype-system="about:legacy-compat"/>
<xsl:template match="/">
<html>
<head>
<title></title>
</head>
<body>
<xsl:apply-templates select="//body"></xsl:apply-templates>
</body>
</html>
</xsl:template>
<xsl:template match="body">
<p><xsl:apply-templates select="p"></xsl:apply-templates></p>
</xsl:template>If I were to run the just this one file and this XSLT(and set the output document to a .txt file) it would generate this:
I was delighted to learn from your letter to Pa and Ma that you plan to go to London
on 4 Aug. and then to come on here from there. I’m again looking forward very much
to your arrival and to finding out what you’ll think of the work that I’ve done
since. The last things I did are a couple of rather large studies of oxcarts, a
black ox and a red and white one.And have also been working again on the old tower in the fields in the
evening; I’ve made a larger study of it than my previous ones — with the wheatfields around
it.Rappard sent me back the little book by Vosmaer that belongs to you — I started to
read it but — is it just me? — find it almighty boring and actually written in an
academic, sermonizing tone. Perhaps you will too when you look at it again. Have
you read Sapho by Daudet? It’s very beautiful, and so vigorous, and so close to
life that the female figure lives, breathes, and one can hear, literally hear the
voice, and forgets that one is reading.You’ll also see a couple more new weavers when you come.Nature is
certainly pure here — I’m still very pleased with the studio, too.We must
visit some farms and weavers together when you come.Rappard’s plan is to come back again in October; he’s
probably in Drenthe again now. Well, I write in some haste because I’m hard at work. I work a good deal
early in the morning or in the evening, and then sometimes everything is so inexpressibly beautiful.This output is something MALLET could use. In order to do this quickly for your whole corpus of XML files you must run a batch transfromation. Detailed instructions on how to perform batch transformations can be found here.
After your batch transformation you should have a subdirectory containing all of your doccuments as .txt files. If you didn't orginally save it to the MALLET directory(the mallet-2.0.8 folder on your desktop) you should do that now. In order to run MALLET requires the input be stored in the MALLET Directory. For information on how to run your first TOPIC Model refer to firstTopicModel.md in this repository.