Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
cancel
Showing results for 
Search instead for 
Did you mean: 
abdel_dadouche
Active Contributor
In case you are catching the train running, here is the link to the introduction blog of the Machine Learning in a Box series which allow you to get the series from the start. At the end of this introduction blog you will find the links for each elements of the series.




Before we get started, a quick recap from last time


Last time, we looked at how to use TensorFlow from within SAP HANA, express edition. This allows you to surface your TensorFlow ModelServer models inside your instances and use them as regular stored procedure.

This allows, for example, to process image or documents stored as blobs with an image classification model or something as simple as a classification on the Iris dataset.

Let's now look at this new post.




Welcome to part 9 of Machine Learning in a Box!


Build your first Machine Learning application






For those following here or on Twitter, you may have seen that I ran a webinar on July the 17th with the SAP HANA International Focus Group named Dive in SAP HANA, Express Edition Machine Learning Capabilities.



You can register for the recording here.

During this session, I presented some of the new tutorials I was working on to help you build your first Machine Learning application using SAP HANA, express edition.

So, let's get started!




The Scenarios


During the webinar, I presented 2 scenarios in 2 different flavors, so 4 possible tracks to follow.

One is about Recommendation engines (or association rules for the purist) and the other about Time Series.

In each scenario, you will have the ability to either:

  • train models and use them to score using SQL script leveraging SAP HANA libraries like APL and PAL


Completing this flavor requires SAP HANA, express edition "server only"




  • build a complete application using the XS Advanced development model to generate models and score your data


Completing this flavor requires SAP HANA, express edition "server + applications"







The Datasets


The MovieLens dataset

The GroupLens Research © group have collected and made available movie rating data sets from the MovieLens web site which were collected over various periods of time.

It also includes details about genres or some attached tags.

This dataset is mainly used for education or research purpose related to recommendation engines.
SAP Predictive Analytics samples for Time Series

When installing SAP Predictive Analytics, you can get access to sample datasets that can be used to demonstrate product features.

The time series dataset includes both real life data like a US Census or Ozone layer reading for the Los Angeles region, as well as dataset that highlight certain signal phenomenon like trends, cycles or white noise.




Where to get started


As usual, I built some new tutorials to help you get started and replicate what was demonstrated during the webinar.

In each tutorial group, you should find details about the required environment (SAP HANA, express edition "server only" or "server + applications") and some context about what you will be trying to achieve.

Here are the groups:

Here is a screenshot of the Time Series application you will build:

Applications

 

For the SQL oriented scenarios, I highly encourage you to use Jupyter notebook as I provided code snippets to execute and get some nice visualizations.

Jupyter




Conclusion

Now, you should have all the pieces to leverage the SAP HANA predictive libraries for your next projects and application! No more excuse!

I might soon expand the groups to add examples of using the open source R integration in these scenario as well as TensorFlow, but for the moment let's keep it simple.

Also, if you are planning to enter a Machine Learning contest and plan on using SAP HANA, express edition, let us know, I will support you (personally)!



(Remember sharing && giving feedback is caring!)


UPDATE: Here are the links to all the Machine Learning in a Box weekly blogs: