[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

[xsl] A robot to do your job


Subject: [xsl] A robot to do your job
From: "Ihe Onwuka ihe.onwuka@xxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Tue, 15 Mar 2016 06:03:05 -0000

Somebody (NOT ME) posted this gig

Doesn't strike me as a great idea but given XSLT's homoiconicity one would
have thought the first thing to try would be genetic programming. In fact
someone has already tried it.

See Evolving XSLT stylesheets  - http://arxiv.org/pdf/0712.2630.pdf

But it wouldn't be a machine learning problem if the fashionable buzzwords
(Hadoop, Deep Learning) weren't chucked at it.


JOB DESCRIPTION
Use Machine Learning or Hadoop to find best mapping for a message

I need a tool to create mappings between two sets of XML documents.

I would like to see how fair this can be performed with Deep Learning,
Fuzzy logic, Hadoop or what you believe could help here.

The system will be trained with a set of in and output messages as the
example. It is then expected to create an XSLT 2.0 mapping that can create
the structure.
There will be use of concat, substring, datetransformation in the proof of
concept.
If successful there may be a bigger project involving nested structure and
other functions.

You can select your favorite tool for the project.

Input XML
<Customer>
    <Number>234</Number>
    <FirstName>daniel</FirstName>
    <LastName>Graversen</LastName>
    <Street>Ejbydalsvej </Street>
    <HouseNumber>15</HouseNumber>
    <Postalcode>DK26000</PostalCode>
    <Birthday>20160208</Birthday>
</Customer>

Output XML
<CustomerERP>
   <NO>234</NO>
   <Name>daniel graversen</Name>
   <Street>Ejbydalsvej 15</Street>
   <Postalcode>2600</Postalcode>
   <Country>DK</Country>
   <Birthday>0208</Birthday>
</CustomerERP>


Current Thread
Keywords