Slides from a talk delivered during 4Developers conference in Warsaw covering basic machine learning concepts and possibilities

1. @SrcMinistry @MariuszGil
Holistic approach to Machine Learning
Data processing

2. @SrcMinistry

3. We are developers

4. We love to…

5. Write code

6. Write tests

7. Use DDD/OOP/AOP/
SOLID/GRASP/XYZ

8. What for?

9. Write code

10. Make money

11. Make users happy

12. Solve problems

13. Solve problems by writing
code, to make users happy
and make money

14. Solve problems by writing
code, to make users happy
and make money
Solve problems

15. Solve problems by writing
code, to make users happy
and make money
Solve

16. Solve problems by writing
code, to make users happy
and make money
problems

17. Mapping all problems to
DDD/OOP/AOP/SOLID/
GRASP/XYZ

18. Test first

19. Understand the
problem first

20. Domain
knowledge

21. Ask expert

22. Real problems

23. Data classification

24. Bot detection

25. Minimize
risk of error

27. + value estimator

28. + chance of sell

29. + $ optimization

30. Tens of thousands
historical transactions

31. Tens of data
components

32. Hundreds of
data components

34. IF-Unsolveable

35. Machine Learning

36. The theory

37. A computer program is said to learn from experience E
with respect to some class of tasks T and performance
measure P if its performance at tasks in T, as measured
by P, improves with experience E
Tom M. Mitchell

38. Task

39. Typical ML techniques
Classification
Regression
Clustering
Dimensionality reduction
Association learning

40. o
oo
o
oo
o
oo
o o
o
o oo
o
o
o
o oo
o
oo
o
o
o
feature 1
feature2

41. o
oo
o
oo
o
oo
o o
o
o oo
o
o
o
o oo
o
oo
o
o
o
feature 1
feature2

42. o
oo
o
oo
o
oo
o o
o
o oo
o
o
o
o oo
o
oo
o
o
o
feature 1
feature2

43. Experience

44. Typical ML paradigms
Supervised learning
Unsupervised learning
Reinforcement learning

45. Accuracy

46. The practice

48. data + algo = result

49. +-------+--------+------+--------+---------+-------+
| brand | model | year | milage | service | price |
+-------+--------+------+--------+---------+-------+
| ford | mondeo | 2005 | 123000 | 9900 | 67000 |
+-------+--------+------+--------+---------+-------+
| ford | mondeo | 2005 | 175000 | 9900 | 30000 |
+-------+--------+------+--------+---------+-------+
| ford | focus | 2010 | 45000 | 6700 | 30000 |
+-------+--------+------+--------+---------+-------+
…

50. Learning Data
Algorithm Learning
Classifier ModelReal Data Classification

51. Failure recipe

52. +-------+--------+------+--------+---------+-------+
| brand | model | year | milage | service | price |
+-------+--------+------+--------+---------+-------+
| ford | mondeo | 2005 | 123000 | 9900 | 67000 |
+-------+--------+------+--------+---------+-------+
| ford | mondeo | 2005 | 175000 | 9900 | 30000 |
+-------+--------+------+--------+---------+-------+
| ford | focus | 2010 | 45000 | 6700 | 30000 |
+-------+--------+------+--------+---------+-------+
…

53. +-------+--------+------+--------+---------+--------+-------+
| brand | model | year | milage | service | repair | price |
+-------+--------+------+--------+---------+--------+-------+
| ford | mondeo | 2005 | 123000 | 9000 | 900 | 67000 |
+-------+--------+------+--------+---------+--------+-------+
| ford | mondeo | 2005 | 175000 | 900 | 9000 | 30000 |
+-------+--------+------+--------+---------+--------+-------+
| ford | focus | 2010 | 45000 | 3700 | 3000 | 30000 |
+-------+--------+------+--------+---------+--------+-------+
…

54. +-------+--------+------+--------+---------+--------+-------+
| brand | model | year | milage | service | repair | price |
+-------+--------+------+--------+---------+--------+-------+
| ford | mondeo | 2005 | 123000 | 9000 | 900 | 67000 |
+-------+--------+------+--------+---------+--------+-------+
| ford | mondeo | 2005 | 175000 | 900 | 9000 | 30000 |
+-------+--------+------+--------+---------+--------+-------+
| ford | mondeo | 2005 | 175000 | 900 | 9000 | 45000 |
+-------+--------+------+--------+---------+--------+-------+
| ford | focus | 2010 | 45000 | 3700 | 3000 | 30000 |
+-------+--------+------+--------+---------+--------+-------+
…

55. +-------+--------+-----+------+--------+---------+--------+-------+
| brand | model | gen | year | milage | service | repair | price |
+-------+--------+-----+------+--------+---------+--------+-------+
| ford | mondeo | 4 | 2005 | 123000 | 9000 | 900 | 67000 |
+-------+--------+-----+------+--------+---------+--------+-------+
| ford | mondeo | 3 | 2005 | 175000 | 900 | 9000 | 30000 |
+-------+--------+-----+------+--------+---------+--------+-------+
| ford | mondeo | 4 | 2005 | 175000 | 900 | 9000 | 45000 |
+-------+--------+-----+------+--------+---------+--------+-------+
| ford | focus | 4 | 2010 | 45000 | 3700 | 3000 | 30000 |
+-------+--------+-----+------+--------+---------+--------+-------+
…

56. +-------+--------+-----+------+--------+---------+--------+------+---------------+-------+
| brand | model | gen | year | milage | service | repair | igla | crying German | price |
+-------+--------+-----+------+--------+---------+--------+------+---------------+-------+
| ford | mondeo | 4 | 2005 | 123000 | 9000 | 900 | 0 | 0 | 67000 |
+-------+--------+-----+------+--------+---------+--------+------+---------------+-------+
| ford | mondeo | 3 | 2005 | 175000 | 900 | 9000 | 1 | 1 | 30000 |
+-------+--------+-----+------+--------+---------+--------+------+---------------+-------+
| ford | mondeo | 4 | 2005 | 175000 | 900 | 9000 | 0 | 0 | 45000 |
+-------+--------+-----+------+--------+---------+--------+------+---------------+-------+
| ford | focus | 4 | 2010 | 45000 | 3700 | 3000 | 1 | 0 | 30000 |
+-------+--------+-----+------+--------+---------+--------+------+---------------+-------+
…

57. Understand your
data first

58. Exploratory
analysis

59. http://blogs.adobe.com/digitalmarketing/wp-content/uploads/2013/08/aq2.jpg

60. ML pipeline

61. Raw Data Collection
Pre-processing
Sampling
Training Dataset
Algorithm Training
Optimization
Post-processing
Final model
Pre-processingFeature Selection
Feature Scaling
Dimensionality Reduction
Performance Metrics
Model Selection
Test Dataset
CrossValidation
Final Model
Evaluation
Pre-processing Classification
Missing Data
Feature Extraction
Data
Split
Data

62. Raw Data Collection
Pre-processing
Sampling
Training Dataset
Algorithm Training
Optimization
Final model
Pre-processingFeature Selection
Feature Scaling
Dimensionality Reduction
Performance Metrics
Model Selection
Test Dataset
CrossValidation
Final Model
Evaluation
Pre-processing Classification
Missing Data
Feature Extraction
Data
Split
Post-processing
Data

63. Classification algorithms
Linear Classification
Logistic Regression
Linear Discriminant Analysis
PLS Discriminant Analysis
Non-Linear Classification
Mixture Discriminant Analysis
Quadratic Discriminant Analysis
Regularized Discriminant Analysis
Neural Networks
Flexible Discriminant Analysis
Support Vector Machines
k-Nearest Neighbor
Naive Bayes
Decission Trees for Classification
Classification and Regression Trees
C4.5
PART
Bagging CART
Random Forest
Gradient Booster Machines
Boosted 5.0

64. Regression algorithms
Linear Regiression
Ordinary Least Squares Regression
Stepwise Linear Regression
Prinicpal Component Regression
Partial Least Squares Regression
Non-Linear Regression /
Penalized Regression
Ridge Regression
Least Absolute Shrinkage
ElasticNet
Multivariate Adaptive Regression
Support Vector Machines
k-Nearest Neighbor
Neural Network
Decission Trees for Regression
Classification and Regression Trees
Conditional Decision Tree
Rule System
Bagging CART
Random Forest
Gradient Boosted Machine
Cubist

65. Algorithm is only
element in the ML chain

66. Everything may be
important for ML

68. Testing

69. Test datasets

70. 60% 20% 20%

71. Andrew NG
rule of ML

72. Does it do well on
the training data?
Does it do well on
the test data?
Better features /
Better parameters
More data
Done!
No No
Yes
by Andrew Ng
Yes

73. Calculate, measure,
apply later

74. The code

75. import org.apache.spark.mllib.classification.{SVMModel, SVMWithSGD}
import org.apache.spark.mllib.evaluation.BinaryClassificationMetrics
import org.apache.spark.mllib.util.MLUtils
// Load training data in LIBSVM format.
val data = MLUtils.loadLibSVMFile(sc, "data/mllib/sample_libsvm_data.txt")
// Split data into training (60%) and test (40%).
val splits = data.randomSplit(Array(0.6, 0.4), seed = 11L)
val training = splits(0).cache()
val test = splits(1)
// Run training algorithm to build the model
val numIterations = 100
val model = SVMWithSGD.train(training, numIterations)
// Clear the default threshold.
model.clearThreshold()
// Compute raw scores on the test set.
val scoreAndLabels = test.map { point =>
val score = model.predict(point.features)
(score, point.label)
}
// Get evaluation metrics.
val metrics = new BinaryClassificationMetrics(scoreAndLabels)
val auROC = metrics.areaUnderROC()
println("Area under ROC = " + auROC)
// Save and load model
model.save(sc, "myModelPath")
val sameModel = SVMModel.load(sc, "myModelPath")

76. Art of asking right
questions related to
right data

77. @SrcMinistry
Thanks!
@MariuszGil