4. Hype Cycle for Emerging Technologies
Source: http://www.gartner.com/newsroom/id/2819918
5. Trends, Customers, Business
Source: The Data Warehousing Institute, TDWI Best Practices Report: Predictive Analytics for Business Advantage, 2014 First Quarter
https://www.sap.com/bin/sapcom/en_us/downloadasset.2014-03-mar-17-21.predictive-analytics-for-business-advantage-pdf.html
6. Data-Rush
Source: The Data Warehousing Institute, TDWI Best Practices Report: Predictive Analytics for Business Advantage, 2014 First Quarter
https://www.sap.com/bin/sapcom/en_us/downloadasset.2014-03-mar-17-21.predictive-analytics-for-business-advantage-pdf.html
7. Talent, Adoption, Integration
Source: The Data Warehousing Institute, TDWI Best Practices Report: Predictive Analytics for Business Advantage, 2014 First Quarter
https://www.sap.com/bin/sapcom/en_us/downloadasset.2014-03-mar-17-21.predictive-analytics-for-business-advantage-pdf.html
8. Simple is Popular
Source: The Data Warehousing Institute, TDWI Best Practices Report: Predictive Analytics for Business Advantage, 2014 First Quarter
https://www.sap.com/bin/sapcom/en_us/downloadasset.2014-03-mar-17-21.predictive-analytics-for-business-advantage-pdf.html
9. They just need a start
Source: The Data Warehousing Institute, TDWI Best Practices Report: Predictive Analytics for Business Advantage, 2014 First Quarter
https://www.sap.com/bin/sapcom/en_us/downloadasset.2014-03-mar-17-21.predictive-analytics-for-business-advantage-pdf.html
12. Analytics Value Chain: Customer Purchase
Decision for Ecommerce
Source: http://www.slideshare.net/valiancesolutions/predictive-analytics-in-ecommerce
13. Segmentation & Personalization
Segmentation is Product Centric
“Customers who viewed/bought this also viewed/bought those”
Better than a “lucky guess”, but not enough for Ecommerce
True Personalization goes way beyond Segmentation
Taking note of everything about the customer
Every search query, every click, every add to cart, and every purchase, along with all the attributes associated with each
Continuously learning about the customer and delivering an experience that responds directly to his or her
intrinsic interests and immediate needs with each returned visit.
Over time, this experience becomes smarter and more intuitive. The customer feels like their online
engagement with a brand is natural, pain free, and even special.
By combining individualized historical data with real time relevancy, we can better anticipate
immediate needs
Source: http://blog.reflektion.com/?p=17
14. If You are BIG, You can Predict HUGE
Source: Amazon’s Patent on Anticipatory Package Shipping
Notes de l'éditeur
Linear regression: Predict future price by using past data
Decision tress: supervised learning to model specific target variables or outcomes of interest.
Clustering is very useful in market segmentation, and marketing and sales are popular areas for predictive analytics among current users.
Time series analysis is used for time-dependent data, popular for forecasting
Logistic regression transforms info about the binary dependent variable into an unbounded continuous variable
Neural networks are used when the exact nature of the relationship between inputs and output is not known. Nonlinear modelling.
naive Bayes classifiers are a family of simple probabilistic classifiers based on applying Bayes' theorem with strong (naive) independenceassumptions between the features.
Support Vector Machines (SVM) are used to detect and exploit complex patterns in data by clustering, classifying and ranking the data. They are learning machines that are used to perform binary classifications and regression estimations. They commonly use kernel based methods to apply linear classification techniques to non-linear classification problems. There are a number of types of SVM such as linear, polynomial, sigmoid etc.
Survival analysis is another name for time to event analysis. These techniques were primarily developed in the medical and biological sciences, but they are also widely used in the social sciences like economics, as well as in engineering (reliability and failure time analysis).
Ensemble learning is the process by which multiple models, such as classifiers or experts, are strategically generated and combined to solve a particular computational intelligence problem. Ensemble learning is primarily used to improve the (classification, prediction, function approximation, etc.) performance of a model, or reduce the likelihood of an unfortunate selection of a poor one.
Model Management
Data Governance
In-memory computing refers to data processing where data is stored in memory
to reduce disk I/O. Models can run faster with in-memory computing, which can be good for
iterative models. In-memory computing is also useful for interactive work such as visualization and
data discovery. In-database analytics embeds analytics in the database. When the amount of data is
large, it can be cheaper when computation is closer to the data