xml - Java Support for PMML -


I'm new to PMML: predictive model markup language () and I was wondering if some kind of Java support (open source / professional) for creating / parsing of PMML files

Initially I have the possibility to create a program from Java environment only for PMML files.

I am "Goggle" and I have many possibilities:

Open Source:

  • (PMML 3.2) / Ul>

    From Java

    • javax.datamining Does it seem dead?

    • Use an XML Java Library and Automatically Make parser / author of PMML files

      I appreciate all your thoughts.

      Oscar

      You should realize that this answer may depend on the model element That's what you want to work with. It is also very likely that your best choice to make PMML and parsing PMML will come from different software packages. I am going to think that 'PMML's creation' means you document, not model. I've never heard of anyone who integrates automatic model fitting with execution, but maybe it already exists. Certainly a PMML model can be passed using SOAP.

      I can not talk to other projects, but the product offered by Zintins, which is the execution of "PMML". This product assumes that there is a model fitting application which will be prepared by exporting the fit model in PMML. There are already quite a well developed models of fitting applications, so I think this is a reasonable assumption.

      The version I used (3.6) was generally faster but it could not handle normal random forest shapes (500+ trees) of the world, especially without a large pile. I think they can fix it in new versions. Although it has not been advertised, the Xintines does not appear to present some models, such as text models, sequences, baseline models, or time series (for which PMML standard is currently only desirable smoothing). In my version K-closest neighbors are not even neighbors, but I have heard what recent versions do.

      As long as you are considering integrated fitting and execution (in that case you should consider learning online) my advice will be to consider these questions:

      1. What is the model I'm interested in using?
      2. What model do I want to model in?
      3. How do I execute it and what are the requirements (web-services, cloud, display etc.) in this regard?

        If you look at the list of DMG groups, you will find many commercial vendors who will be on the supply side (e.g. SAS, SPSS, TOWWARE, RAPID-I) or the demand side (so many in the list) all).

        You have not mentioned Vaka in your list but there are also and therefore you can execute in Java environment (but you can also execute R).

        Finally, if you have a very specific model in your mind and you understand that in order to 'execute' it mathematically it means that what you need to do to make it is not difficult Should be

Comments

Popular posts from this blog

mysql - BLOB/TEXT column 'value' used in key specification without a key length -

c# - Using Vici cool Storage with monodroid -

python - referencing a variable in another function? -