SAPs Predictive Analysis Library PAL Function Cheat Sheets

SAP’s Predictive Analysis Library (PAL) Function Cheat Sheets Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

Agenda What Will Be Covered: 1. Predictive Analysis Library (PAL) Overview 2. PAL Function Cheat Sheets 1. Clustering/Classification/Regression 2. Association/Time Series/Preprocessing 3. Statistics/Social Network Analysis/Miscellaneous 3. About us Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

Predictive Analysis Library (PAL ) Overview § Just the term “SAP Predictive Analysis” can make the best analyst tremble in fear, as it’s a complicated topic to comprehend. § The “why” versus the “what” are the easier topics to grasp. § The “how” is where people commonly hit unpassable roadblocks. § It can get overwhelming; SAP Predictive Analysis metrics & calculations are confusing, disjointed & constantly changing. § To help, SAP has grouped functions for particular topics together into the Application Function Library (AFL). For further information please refer to: http: //help. sap. com/saphelp_hanaplatform/helpdata/en/65/2 c 5 ada 587 d 4 a 018 aa 4 a 4 cde 6 ae 1 fc 4/frameset. htm Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

Predictive Analysis Library (PAL ) Overview § SAP’s Predictive functions were grouped together in the AFL’s Predictive Analysis Library (PAL). § Add-on set of application functions that implement analysis algorithms. § Makes executing clustering calculations w/ SAP HANA data easy & straightforward. § Leverages HANA’s in-memory & near-linear parallelism performance for scoring, training, & categorization; without data leaving the server. § PAL can be accessed by SAP HANA SQL Script, which is an extension to SQL. § § Adds control-flow capabilities & the ability to define complex application logic. Embeds predictive analytics into business applications. Contains universal predictive algorithms that execute directly against SAP HANA data. Enables up to 80% of the most use case common predictive scenarios. For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

Predictive Analysis Library (PAL ) Overview § PAL function algorithms are required for SAP HANA applications, based on market survey responses, & are usually available in other database products. § PAL includes classic and universal predictive analysis algorithms in nine datamining function categories: § Clustering/Classification/Regression § Association/Time Series/Preprocessing § Statistics/Social Network Analysis/Miscellaneous § Let’s now take a look at the PAL function algorithms in the following slides. § A full detailed list can be found via the link below: § http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Clustering § PAL Clustering Functions can be seen below: Category PAL Algorithm Built-in Function Name Clustering Affinity Propagation AP Agglomerate Hierarchical Clustering HCAGGLOMERATE Anomaly Detection ANOMALYDETECTION Cluster Assignment CLUSTERASSIGNMENT DBSCAN Gaussian Mixture Model (GMM) GMM K-Means KMEANS / VALIDATEKMEANS K-Medians KMEDIANS K-Medoids KMEDOIDS LDA Estimation and Inference LDAESTIMATE / LDAINFERENCE Self-Organizing Maps SELFORGMAP Slight Silhouette SLIGHTSILHOUETTE For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Classification § PAL Classification Functions can be seen below: Category PAL Algorithm Built-in Function Name Classification Area Under Curve (AUC) AUC Back Propagation Neural Network CREATEBPNN / PREDICTWITHBPNN C 4. 5 Decision Tree CREATEDTWITHC 45 CART Decision Tree CART CHAID Decision Tree CREATEDTWITHCHAID Confusion Matrix CONFUSIONMATRIX KNN Logistic Regression (w/ Elastic Net Regularization) LOGISTICREGRESSION / FORECASTWITHLOGISTICR Multi-Class Logistic Regression LRMCTR / LRMCTE Naive Bayes NBCTRAIN / NBCPREDICT Parameter Selection & Model Evaluation (PSME) PSME Predict w/ Tree Model PREDICTWITHDT Random Forest RANDOMFORESTTRAIN / RANDOMFORESTSCORING Support Vector Machine SVMTRAIN / SVMPREDICT For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Regression § PAL Regression Functions can be seen below: Category PAL Algorithm Built-in Function Name Regression Bi-Variate Geometric Regression GEOREGRESSION / FORECASTWITHGEOR Bi-Variate Natural Logarithmic Regression LNREGRESSION / FORECASTWITHLNR Exponential Regression EXPREGRESSION / FORECASTWITHEXPR Multiple Linear Regression LRREGRESSION / FORECASTWITHLR Polynomial Regression POLYNOMIALREGRESSION / FORECASTWITHPOLYNOMIALR For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Association § PAL Association Functions can be seen below: Category PAL Algorithm Built-in Function Name Association Apriori APRIORIRULE / LITEAPRIORIRULE / APRIORIRULE 2 FP-Growth FPGROWTH K-Optimal Rule Discovery (KORD) KORD For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Time Series § PAL Time Series Functions can be seen below: Category PAL Algorithm Built-in Function Name Time Series ARIMATRAIN / ARIMAFORECAST / ARIMAXFORECAST Auto ARIMA AUTOARIMA Brown Exponential Smoothing BROWNEXPSMOOTH Croston’s Method CROSTON Forecast Accuracy Measures ACCURACYMEASURES Forecast Smoothing FORECASTSMOOTHING Linear Regression w/ Damped Trend & Seasonal Adjust LRWITHSEASONALADJUST Single Exponential Smoothing SINGLESMOOTH Double Exponential Smoothing DOUBLESMOOTH Triple Exponential Smoothing TRIPLESMOOTH Seasonality Test SEASONALITYTEST Trend Test TRENDTEST White Noise Test WHITENOISETEST For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Preprocessing § PAL Preprocessing Functions can be seen below: Category PAL Algorithm Built-in Function Name Preprocessing Binning BINNING Binning Assignment BINNINGASSIGNMENT Convert Category Type to Binary Vector CONV 2 BINARYVECTOR Inter-Quartile Range Test IQRTEST Partition PARTITION Posterior Scaling POSTERIORSCALING Principal Component Analysis (PCA) PCA / PCAPROJECTION Random Distribution Sampling DISTRRANDOM Sampling SAMPLING Scaling Range SCALINGRANGE Substitute Missing Values SUBSTITUTE_MISSING_VALUES Variance Test VARIANCETEST For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Statistics § PAL Statistics Functions can be seen below: Category PAL Algorithm Built-in Function Name Statistics Chi-Squared Test for Fitness CHISQTESTFIT Chi-Squared Test for Independent CHISQTESTIND Cumulative Distribution Function DISTRPROB Distribution Fitting DISTRFIT / DISTRFITCENSORED Grubbs’ Test GRUBBSTEST Kaplan-Meier Survival Analysis KMSURV Multivariate Statistics MULTIVARSTAT Quantile Function DISTRQUANTILE Univariate Statistics UNIVARSTAT Variance Equal Test VAREQUALTEST For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Social Network Analysis § PAL Social Network Analysis Functions can be seen below: Category PAL Algorithm Built-in Function Name Social Network Analysis Link Prediction LINKPREDICTION For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

PAL Functions: Miscellaneous § PAL Miscellaneous Functions can be seen below: Category PAL Algorithm Built-in Function Name Miscellaneous ABC Analysis ABC Weighted Score Table WEIGHTEDTABLE For further information please refer to: http: //help. sap. com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en. pdf Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

Success – You Completed The Lesson! Congratulations, you completed the lesson! In this lesson you successfully learned: 1. Predictive Analysis Library (PAL) Overview 2. PAL Function Cheat Sheets 1. Clustering/Classification/Regression 2. Association/Time Series/Preprocessing 3. Statistics/Social Network Analysis/Miscellaneous 3. About us Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com 10/10

Require A Consultation? Want to learn more about SAP’s Predictive Capabilities? Contact us today for your FREE consultation with our experts. Email: info@blackvard. com Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

What We’ve Accomplished Key Achievements of Blackvard Management Consulting in Previous Projects n Technical project leads and ABAP architects responsible for quality in technical scope and budget in global roll-outs of SAP Logistics applications (SAP LE / LO) n Conducted multiple SAP ABAP, SAP HANA®, and SAP Fiori® trainings for various US companies n Implementation of a standard SAP software solution for Spend Management within SAP AG & ARIBA (annual spend volume 3 Bill. EUR) which can be used in all SAP systems n Improved claims management using SAP FS-CM which is generating annual savings of 15 Mio € for a huge German public healthcare organization n Implemented a global solution for procurement processes at BMW AG using SAP SRM / B 2 B n Blueprinting and implementation of SAP software for banking credit cancelations for VOLKSWAGEN Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com

Managing Director Short Bio: Lukas M. Dietzsch is managing director at Blackvard Management Consulting, LLC. He is holding a Master’s degree in Information Technology and is an experienced IT solution architect and project lead. His strong background in adapting to requirements and standards in different industries and on various platforms are valuable assets for Blackvard customers. Lukas M. Dietzsch lukas@blackvard. com Blackvard Management Consultants Copyright © Blackvard Management Consulting –All Allrightsreserved He is repeatedly commended by customers for driving efficient solutions for complex problems in globally distributed team environments and meeting tough deadlines. For further information please visit: www. blackvard. com

Customers That Recommend Blackvard An overview of current and previous customers: Copyright © Blackvard Management Consulting – All rights reserved www. blackvard. com
- Slides: 19