Developing for WML for z/OS

IBM Watson® Machine Learning for z/OS helps you build machine learning models to extract value from your mission critical data on IBM® Z systems. Use the samples and scenarios to get started with developing comprehensive machine learning algorithms and models in the WML for z/OS user interface.

Before you begin

Tip: WML for z/OS supports the standard view (or desktop version) of Mozilla Firefox and Google Chrome. Make sure that you run the WML for z/OS UI in one of these supported browsers. See Prerequisites and maintenance for IBM Machine Learning for z/OS for browser support details.
  • Install and configure WML for z/OS as described in Roadmap for installing and configuring WML for z/OS.
  • Understand the following terms:
    Pipeline
    A sequence that loads data, transforms it, and runs it through an estimator to produce an algorithm, or model, that can generate a prediction for each data record.
    Estimator
    Code that receives data as input and identifies patterns to generate the prediction algorithm or model.
    Model
    The algorithm that can detect patterns in a data record to generate a prediction.

Procedure

Complete the following steps to generate machine learning algorithms and deploy them to the production environment where the data resides:

  1. Create a pipeline that generates a model:
    1. Select data from a data source.
      For more information about the data sources that WML for z/OS supports, see Supported algorithms, data sources, data types, and model types.
    2. Transform and clean the data, to prepare it for use as input to the estimator.
    3. Define an estimator to prepare the machine learning algorithm.
    4. Train the machine learning algorithm by using a portion of the prepared data as input the estimator.
    5. Save the machine learning algorithm as a model.
  2. Evaluate the model with a larger set of prepared data.
  3. Deploy the model to the production environment.
  4. Automate a feedback loop to retrain the model in the pipeline to address accuracy degradation as the data changes.

What to do next

Prepare the sample data that is provided with WML for z/OS, and use the following scenarios to learn how to develop with WML for z/OS.