Skip to Content

Difference in Prediction

So what is the difference of prediction

1.when being done at HANA DB using PAL

&

2.when its carried at SAP BO predictive analytics or any other tool like SAP Analytics cloud??

We have already a SAS(Software as service) developed in SAP analytics cloud tied with SAP HANA DB.

So to add predictive features to the existing product which is the right option to choose??

1.SAP Analytics cloud (with predictive) ?

2.SAP BO predictive analytics?

3.OR Doing the predictive modelling in HANA DB using PAL and then expose it through any of these kind of tools

what is the right direction to go with??

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

2 Answers

  • Best Answer
    Feb 20 at 06:52 AM

    Hello,

    It depends on your application use case and what you are trying to achieve there.

    One way to deal with this could be:

    1. Create predictive models using SAP Predictive Analytics on top of SAP HANA. Use Predictive Factory to regularly update the models and produce results onto the SAP HANA DB. By results I mean forecasts (time series), probabilities (classification), clusters (clustering), values (regression).

    2. Consume the results available in SAP HANA, using your SaaS application built on the basis of SAP Analytics Cloud.

    You can do similar things using PAL, but this will probably require more data science & SAP HANA skills to operate this.

    As I mentioned in my previous answer SAP Analytics cloud (with predictive) features are intended for the BI user or the Planning user and are deeply embedded into the workflows.

    Another thing you could use are the SAP Predictive Service in case you need specific insights like the key influencers (classification) or the forecasts.

    I hope this helps regarding the various options,

    kind regards

    Antoine

    Add comment
    10|10000 characters needed characters exceeded

  • Feb 20 at 06:54 AM

    On a separate note: Nice profile picture! Makes me think of Arcimboldo.

    Add comment
    10|10000 characters needed characters exceeded