Business Function Library (BFL) in SAP Hana

Business Function Library

It is the calculation library of applications business top of the SAP HANA database. If resides in the calculation engine and consists of many business Functions exciting at the database layer and is written in c

Design Goals

Significant performance improvement for SAP app Utilizing new hardware i.e. multi core, built vector engine Mastic parallel main memory processing Changing the dromedaries between application server and data management layer Simplification of application F gramming _ model Usage of extended SQL (SQL script) Rich functionless in the calculation engine Fast apps delivery

New Approach to Building Apps

1. Performance Improvement by Factors Applications need to push down the data-intensive calculation into the SAP HANA database Parallelization of algorithms

2. Shorten Development Life guide Rich business functions for different industries Extend and powerful SQL scripts for use.

3. Deliver – Customers’ real need application Quickly build business scenarios Effective communication on requirements

Predictive Analysis Algorithms in BFL

In SAP HANA SPUB, the following, will supported in the BFL, grouped into class syerred to as the predictive analysis library within the BFL Clustered (K-Means) Regression Association Analysis (approve) Decision Trce (c4.5) ABC clarification Weighted score tables The predictive analysis process is as follows Select the data for the predicative analysis Data could be in csr file, database table, universe….. Understand the data and prepare the data for analysis Various data preparation methods: Sample, Filter, Merge, Append. & Apply formula ……. Visualize the data  

Perform the analysis using predictive algorithms & visualize the results

Define the mode: clarification analysis, time series analysis Association analysis, segmentation analysis …………….. Training the model, testing the model, and visualizing the model.

Example: A decision tree viewer, cluster viewer  

Store the result or created model

Store the results in CSV, database, and universal store the model in PMML. Store the created “trained” model in the repository …  

Use the created model for further analysis

Use the model on the tire data to scours the model Export the created model in an Industry-standard model which can be used by SAP business applications.