In this video, you will see how to build
a logistic regression model that assesses the likelihood that a customer of an outdoor equipment company will buy a tent. For this video, there’s already a Watson Studio project with the Watson Machine Learning, Apache Spark and Object Storage services associated. If you haven’t done that yet, first watch the video showing how to create the project in the IBM Watson and Cloud Platform Learning Center. Here you see a project that has all three associated services. For this video, you’ll use a data set called go sales transactions for logistic regression model which you’ll find in the IBM Watson community. Here you see the data preview and the column definitions. Add this data set to the Watson Machine Learning project so you can use it later. Now view the project, and on the assets
tab you’ll see the data set. Now you’re ready to create a new model. Name this model logistic regression and provide a description. The Machine Learning and Spark services are already associated. You can use the model builder to create your model or load an existing model from a file or use one of the sample models. In this case, select model builder
then select manual for the training method and create the model. After you load the data, you must train the data. This consists of choosing an appropriate technique and estimator to apply to the raw data. For this data set, you’ll use the ‘is tent’ label column and the logistic regression estimator type. To begin training, select the data set and click Next. For the label column, select is ‘is tent’
and keep the default feature columns. For the technique, select binary classification. Next add an estimator using logistic regression. Click Next and watch the progress as it
trains the data. After the training is complete save it. After you train and save the model,
you’re ready to deploy it. This is also a good time to do a check on the data and the results. On the deployments tab, add a deployment. This will be a web service deployment with the specified name. When you’re ready, save the new deployment. When the model deployment is complete,
view the deployment. Now test the model prediction. Use the
sample input data and click predict. This shows that there’s less than a 15%
chance that a 27 year old single professional male will buy a tent. Change the input data to test different predictions. On the implementation tab, you’ll find
the scoring end point for future reference. You’ll also find code snippets for
various programming languages to utilize this deployment from your application. You can also view the API specification from here. And of course this new model is available on the assets tab in the project. Find more videos in the IBM
Watson and Cloud Platform Learning Center.