Escolar Documentos
Profissional Documentos
Cultura Documentos
Public
Speakers
Las Vegas, Sept 19 - 23
Bangalore, October 5 - 7
Barcelona, Nov 8 - 10
Public
Disclaimer
The information in this presentation is confidential and proprietary to SAP and may not be disclosed without the permission of
SAP. Except for your obligation to protect confidential information, this presentation is not subject to your license agreement or
any other service or subscription agreement with SAP. SAP has no obligation to pursue any course of business outlined in this
presentation or any related document, or to develop or release any functionality mentioned therein.
This presentation, or any related document and SAP's strategy and possible future developments, products and or platforms
directions and functionality are all subject to change and may be changed by SAP at any time for any reason without notice.
The information in this presentation is not a commitment, promise or legal obligation to deliver any material, code or functionality.
This presentation is provided without a warranty of any kind, either express or implied, including but not limited to, the implied
warranties of merchantability, fitness for a particular purpose, or non-infringement. This presentation is for informational
purposes and may not be incorporated into a contract. SAP assumes no responsibility for errors or omissions in this
presentation, except if such damages were caused by SAPs intentional or gross negligence.
All forward-looking statements are subject to various risks and uncertainties that could cause actual results to differ materially
from expectations. Readers are cautioned not to place undue reliance on these forward-looking statements, which speak only
as of their dates, and they should not be relied upon in making purchasing decisions.
Public
Agenda
Introduction to SAP HANA Smart Data Streaming
Streaming Analytics & Complex Event Processing
Machine Learning in HANA Smart Data Streaming
Adaptive Hoeffding Tree
DenStream Clustering
Demo
Example Use Cases
Integration with SAP Predictive Analytics
Resources
Public
Immediate response
Act in Real-time
Public
Public
Secure
Rapid Development
Predictive Analytics
Scalable
Fault Tolerant
Public
Public
Sensors
Click streams
Social media
Transactions
Market prices
Public
Database Queries
Step 1:
Step 2:
Continuous Queries
Step 1:
Step 2:
Define the
continuous
queries and the
dataflow
Public
10
Virtually no useful
information in a
single isolated event
history
e.g. Compare
variance of trends
across multiple
sensors against
historical norms
Alert
Public
11
Input Streams
HANA DB
Alerts
adapters
Dashboards
Streaming Server
Applications
Public
12
Spot significant
changes
Monitor correlations
Compare current
values to historical
norms
Alerts
Examples:
Examples:
Automatically adjust
prices based on
market conditions
Dispatch a technician
for urgent preventive
maintenance
Tailor an offer to a
user based on current
activity
Alert a supervisor
when a machine
needs adjustment
before quality is
affected
Alert IT staff as soon
as a security threat is
detected
Apply predictive
models to anticipate
whats coming
Live Dashboards
Immediate Response
Continuously compute
and stream summary
data to live
dashboards
Public
13
Data reduction
Options include:
Sample high frequency data to reduce the number of data
points
Only record changes
Data tiering:
Store high value data in HANA in-memory tables
Lower value data in HANA Dynamic Tiering or in Hadoop
High speed HANA database loading with support for parallel
writing to multiple partitions
Public
14
Machine Learning
Public
Public
16
Bayesian networks
Reinforcement learning
Representation learning
Similarity and metric learning
Sparse dictionary learning
Genetic algorithms
Public
17
DenStream (Clustering)
An incremental clustering algorithm which
uses the concept of micro-clusters to
summarize clusters of arbitrary shapes and
a pruning technique to detect outliers.
Public
18
Machine Learning Algorithms that learn on the fly are always kept up-todate while utilizing less hardware resources and generating predictions
with a single pass of data
2016 SAP SE or an SAP affiliate company. All rights reserved.
Public
19
Demo
Implementing machine learning algorithms in HANA Smart Data Streaming
Public
Demo Notes
Demo will show:
HANA Studio
Minimal streaming project for Hoeffding Training
Streaming project for Hoeffding Scoring
Streaming project for DenStream Clustering
Model Definition in HANA
Discuss interpreting the output
Public
21
Demo Notes
Public
22
Public
Public
24
Public
25
Public
26
Purchasing behavior using either debit or credit will likely show a unique pattern for individual users
Common dimensions to define the behavior may be
Geographic location
Purchase amount
Time of day
Others?
The DenStream Clustering algorithm can be used to identify clusters of normal purchasing
behavior for individual users.
Outliers to the normal clustered purchasing behavior would be indicators of potential theft or fraud
Combining multiple models for different dimensions can increase the strength of the prediction
Example: Combine geographic location and purchase amount
If both dimensions show an outlier to the normal clustered behavior then you have a stronger
indicator of possible fraud than either dimension in isolation
Further combining with clustered behavior of larger groups can help to eliminate false alarms
Public
27
If you further identified that a normal cluster of behavior is for customers from England who make purchases in a
foreign country have also purchased a plane ticket in the past X months then you have an even stronger indicator
that this is or isnt a normal practice
Which may be best handled by combining with a Hoeffding Tree to make a decision
based on the combination of factors
Public
28
Public
Automated Analytics
A few data mining functions for answering various business questions
Classification / Scoring
Who will churn, fraud or buy next week, next month ?
Regression
How many products will a customer buy next month, next quarter ?
Segmentation / Clustering
What are the groups of customers with similar behavior or profile ?
Forecasting
How much will be the monthly revenue or number of churners next year ?
Recommendations
What is the best offer or action for a customer or internet user ?
Public
30
Prepare data:
Clustering
Forecasting
Association rules
Public
31
Automated Analytics
Automated Data preparation
Provide business
analysts with a
fully automated
process
Public
32
Expert Analytics
Perform statistical analysis on
your data to understand
trends & detect outliers in
your business
Provide data
scientists with
sophisticated
algorithms to take
the next step in
understanding
their business
and modeling
outcomes
Breadth of connectivity to
access almost any data
Optimized for SAP HANA to
support huge data volumes &
in-memory processing
Expert Analytics
Public
33
Demo
Integrating SAP Predictive Analytics with HANA Smart Data Streaming
Public
Resources
Public
Resources
Smart Data Streaming Developer Center
http://scn.sap.com/community/developer-center/streaming
http://scn.sap.com/community/developer-center/streaming/blog/2016/03/11/table-of-contents
http://scn.sap.com/docs/DOC-71415
http://scn.sap.com/docs/DOC-71407
CIO Guide to Using the SAP HANA Platform for Big Data
https://jam4.sapjam.com/wiki/show/0reRAeek9m48mfnFYv7Dm7
2016 SAP SE or an SAP affiliate company. All rights reserved.
Public
36
Further information
Related SAP TechEd sessions:
DMM167 - Implementing Streaming Analytics for Real-Time Alerting and Response
Public
37
Keynotes
Demo Jam
SAP TechEd live interviews
Select lecture sessions
Hands-on sessions
Public
38
Feedback
Please complete your
session evaluation for
DMM105.
Contact information:
Rob Waywell
HANA Product Management
robert.waywell@sap.com
Public
39
Public
40