Hyderabad (Indien) Budapest/Szeged (Ungarn) Magdeburg Stuttgart Ingolstadt München Stuttgart, München, Ingolstadt, Magdeburg Madrid (Spanien) Locations in different continents Global services
THE “INTELLIGENT ENTERPRISE” Clustering Classification Regression Association Time Series Miscellaneous Statistics Machine learning methods, also applicable for data from ERP systems and data lakes. Special challenge: Analysis across system boundaries (virtual or with replication) • Natural Language Processing based on neural networks incl. • Incident classification with semantic normalization • Real-time analysis of the evolution of problem classes Integration of Deep Learning (i.e. Tensorflow) in different environments, see #sitBCN 2018
AI/ML/PA techniques with non-integrated functionality requires Extraction and import of externally determined results to the ML environments Productive use requires continuous verification of results determined with the point above and production-ready model adjustments Heterogeneous technologies/programming environments based on different knowledges
the shown scenario require HANA 2.0.2 or higher) SAP ABAP 7.5.x (i.e. ECC 6.0 EHP, S/4HANA, BW on HANA, BW/4HANA) Abap Development Tools in Eclipse Understanding of the basics of ML/AI/PA Documentation of ABAP, ABAP CDS, SQL Script and PAL REQUIRED COMPONENTS FOR OUR SCENARIO
classification, consumable by the end user For today’s cooking, we will need 3 ABAP CDS Views 1 AMDF 1 SAP HANA PAL procedure 1 BW on HANA (i.e. ABAP NW 7.50) 1 HANA DB 2.0.2 or higher Data: SBOOK, SCUSTOM,… 1 ADT in Eclipse 1 Analytical Frontend
CLASSIFICATION” (Sales) Data Parameters PERCENT_A = e.g. 70% PERCENT_B = e.g. 20% PERCENT_C = e.g. 10% • All customer which are responsable for 70% of the revenue are class A, • the next group of customers responsible for the next 20% are class B, • the rest is class C ABC classification Result Customer ID Class A|B|C