Automated decision making with predictive applications – Big Data Frankfurt
13 likes3,251 views
The document discusses the challenges and dynamics of automated decision-making and big data usage in businesses, highlighting that a significant portion of decisions still relies on qualitative factors rather than data-driven methodologies. It emphasizes the importance of predictive applications and machine learning in improving decision-making accuracy and efficiency. Additionally, the text outlines the intricacies of business rules, common cognitive biases affecting human decisions, and the need for reliable data integration in predictive modeling.
Automated decision making with predictive applications – Big Data Frankfurt
- 2. Automated Decision Making with Big Data Lars Trieloff | @trieloff
- 3. Automated Decision Making with Big Data Predictive Applications Lars Trieloff | @trieloff
- 4. — Holger Kisker, Forrester Research “Even after more than 20 years of using BI, they still base nearly 45% of business decisions on qualitative decision factors instead of quantitative, fact-based evidence.“
- 5. If data is not used for decision making, what is used then?
- 8. 4%Worldwide average profit margin in retail: 4%
- 9. 4‰German average profit margin in retail: 4‰
- 10. Your Customer gives you this
- 11. All you got to keep is that
- 12. — –Libby Rittenberg “Economic profits in a system of perfectly competitive markets will, in the long run, be driven to zero in all industries.”
- 14. — Abraham Maslov – probably never said this. It’s true anyway. “Data has Human Needs, too”
- 18. — W. Edward Deming “In God we trust, all others bring data”
- 19. How Data-Driven Decisions should work Computer Collects Computer Stores Human Analyzes Human Predicts Human Decides
- 20. — Daniel Kahneman “Prejudice against algorithms is magnified when the decisions are consequential.”
- 21. How Data-Driven Decisions REALLY work Computer Collects Computer Stores Human Analyzes C O M M U N I C AT I O N B R E A K D O W N Human Decides
- 22. — Led Zeppelin Communication Breakdown, It's always the same, I'm having a nervous breakdown, Drive me insane!
- 23. • Drill-down analysis … misunderstood or distorted • Metrics dashboards … contradictory and confusing • Monthly reports … ignored after two iterations • In-house analyst teams … overworked and powerless How Data-Driven Decisions REALLY work CO M M U N I C AT I O N B R E A K D O W N
- 24. How Data-Driven Decisions REALLY work http://dilbert.com/strips/comic/2007-05-16/
- 25. How Decisions REALLY should work Computer Collects Computer Stores Computer Analyzes Computer Predicts CO M P U T E R D E C I D E S
- 26. — Everyone at Blue Yonder, all the time 99.9% of all business decisions can be automated
- 27. How Decisions are Being Made
- 28. 90% No Decision is made
- 29. — Robin Sharma “Making no decision is a decision. To do nothing. And nothing always brings you nowhere..”
- 30. Business Rules for Beginners Not doing anything is the simplest business rule in the world – and also the most popular
- 31. 90% No Decision is made
- 32. 9% Decision Follows Rule
- 33. Advanced Business Rules Computers are machines following rules. This means business rules are programs.
- 34. • Business rules are like programs – written by non-programmers • Business rules can be contradictory, incomplete, and complex beyond comprehension • Business rules have no built-in feedback mechanism:“It is the rule, because it is the rule” Business rules are Programs, just not very good ones.
- 35. — Mark Twain “It ain’t what we don’t know that causes trouble, it’s what we know for sure that just ain’t so”
- 36. 1% Human Decision making
- 37. Human Decision Making has two systems – and only one is rational.
- 38. Not quite Almost there That’s it.
- 39. Quick: What do you see here?
- 41. — Steven Pinker, describing Moravec’s Paradox “The hard problems are easy and the easy problems are hard.”
- 42. Quick: Add all even numbers 65 7 1 0 60 63 18 80 547039100
- 43. 94 39 37 31 92 70 100 67 4956080 69 20 26 73
- 44. 51 60 23 22 5 48 43 14 9525669 23 67 1 43
- 45. Correct Result:
- 48. — Daniel Kahneman “All of us would be better investors if we just made fewer decisions.”
- 50. How we are making decisions (Like the big apes we are) Anchoring effect IKEA effect Confirmation bias Bandwagon effect Substitution Availability heuristic Texas Sharpshooter Fallacy Rhyme as reason effect Over-justification effect Zero-risk bias Framing effect Illusory correlation Sunk cost fallacy Overconfidence Outcome bias Inattentional Blindness Benjamin Franklin effect Hindsight bias Gambler’s fallacy Anecdotal evidence Negativity bias Loss aversion Backfire effect
- 52. K-Means Clustering Naive Bayes Support Vector Machines Affinity Propagation Least Angle Regression Nearest Neighbors Decision Trees Markov Chain Monte Carlo Spectral clustering Restricted Bolzmann Machines Logistic Regression Computers making decisions (cold, fast, cheap, rational)
- 53. • A machine learning algorithm is a system that derives a set of rules based on a set of data • It is based on systematic observation, double- checking and cross-validation • There is no magic, just data – and without data there is no magic either Machine Learning means Programs that write Programs
- 54. Better Decisions through Predictive Applications
- 55. How Predictive Applications Work Collect & Store Analyze Correlations Build Decision Model Decide & Test Optimize
- 56. — Warren Buffett “I checked the actuarial tables, and the lowest death rate is among six-year-olds, so I decided to eat like a six-year- old.”
- 57. More than half of the apps on a typical iPhone home screen are predictive applications.
- 58. Building Predictive Applications Machine Learning ModelPredictive Application Enterprise Integration
- 59. Story Time (Not safe for vegetarians)
- 60. The Ground Beef Dilemma
- 61. How much ground beef are we going to sell on Friday?
- 62. How much ground beef are we going to sell on Friday? And how much on Saturday?
- 63. Challenge #1 Accurately predict demand
- 64. Great. But how much do we need to order each day?
- 65. Great. But how much do we need to order each day? Let’s reduce the risk of running out of stock to 20%
- 66. Sales Forecasts for Friday SalesProbability 0 0,01 0,02 0,03 0,04 0 4 8 12 16 Friday Sales Amount
- 67. Sales Forecasts for Saturday SalesProbability 0 0,01 0,02 0,03 0,04 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Saturday Sales Amount
- 68. Great. But how much do we need to order each day? Let’s reduce the risk of running out of stock to 20% So it’s 3 on Friday and 5,5 on Saturday.
- 69. Sales Forecasts for Both Days SalesProbability 0 0,01 0,02 0,03 0,04 0 4 8 12 16 Friday Sales Amount Saturday Sales Amount
- 70. Bad news…
- 71. Bad news… We need to skip the Saturday delivery.
- 72. Bad news… We need to skip the Saturday delivery. How big should we make the Friday delivery instead?
- 73. If you need 3 on Friday and 5,5 on Saturday to fulfill 80% of the demand, how much do you need to fulfill 80% of the combined demand?
- 74. 3 + 5,5 = 8,5 Common Sense isn’t it?
- 75. — Albert Einstein Common sense is what tells us the world is flat.
- 76. Combined Sales Forecasts SalesProbability 0 0,01 0,02 0,03 0,04 0 4 8 12 16 Combined Sales Amount
- 77. If you ordered 8,5 cases, you would waste a lot of meat, the ideal order amount is 8 cases.
- 78. Predictive Apps in a Nutshell Batch and streaming data ingestion, batch and streaming delivery (with real-time option) Reduce risk and cost » increase revenue and profit Trend Estimation Classification Event Prediction Optimize Returns Collect Data Predict Results Drive Decisions
- 79. — John Maynard Keynes “When my information changes, I alter my conclusions. What do you do, sir?”
- 80. One Common Platform for Predictive Applications Your own and third- party data, easily integrated via API Link Build Machine Learning and application code Build Automatically run and scale ML models and applications Run Monitor and inspect resource usage and model quality View Your data stored in high-performance database as a service Store
- 81. — Kevin Kelly “The business plans of the next 10,000 startups are easy to forecast: Take X and add AI”
- 82. Lars Trieloff @trieloff (this guy is hiring)