Machine Learning With Scikit-Learn - Pydata Strata NYC 2015

Introduction to machine learning and scikit-learn, including basic API, grid-search, pipelines, model complexity and in-depth review of some supervised models.

Transcript

1. Machine Learning with scikit-learn Andreas Mueller (NYU Center for Data Science, co-release manager scikit-learn)
2. 2 Me
3. 3 What is scikit-learn?
4. 4 Classification Regression Clustering Semi-Supervised Learning Feature Selection Feature Extraction Manifold Learning Dimensionality Reduction Kernel Approximation Hyperparameter Optimization Evaluation Metrics Out-of-core learning …...
5. 5 http://scikit-learn.org/
6. 6 What is machine learning?
7. 7 Hi Andy, I just received an email from the first tutorial speaker, presenting right before you, saying he's ill and won't be able to make it. I know you have already committed yourself to two presentations, but is there anyway you could increase your tutorial time slot, maybe just offer time to try out what you've taught? Otherwise I have to do some kind of modern dance interpretation of Python in data :-) -Leah Hi Andreas, I am very interested in your Machine Learning background. I work for X Recruiting who have been engaged by Z, a worldwide leading supplier of Y. We are expanding the core engineering team and we are looking for really passionate engineers who want to create their own story and help millions of people. Can we find a time for a call to chat for a few minutes about this? Thanks
8. 8 Hi Andy, I just received an email from the first tutorial speaker, presenting right before you, saying he's ill and won't be able to make it. I know you have already committed yourself to two presentations, but is there anyway you could increase your tutorial time slot, maybe just offer time to try out what you've taught? Otherwise I have to do some kind of modern dance interpretation of Python in data :-) -Leah Hi Andreas, I am very interested in your Machine Learning background. I work for X Recruiting who have been engaged by Z, a worldwide leading supplier of Y. We are expanding the core engineering team and we are looking for really passionate engineers who want to create their own story and help millions of people. Can we find a time for a call to chat for a few minutes about this? Thanks
9. 9 Doing Machine Learning With Scikit-Learn
10. 10 Representing Data X = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3
11. 11 Representing Data X = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3 one sample
12. 12 Representing Data X = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3 one sample one feature
13. 13 Representing Data X = y = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3 1.6 2.7 4.4 0.5 0.2 5.6 6.7 one sample one feature outputs / labels
14. 14 Training and Testing Data X = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3 y = 1.6 2.7 4.4 0.5 0.2 5.6 6.7
15. 15 Training and Testing Data X = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3 y = 1.6 2.7 4.4 0.5 0.2 5.6 6.7 training set test set
16. 16 Training and Testing Data X = 1.1 2.2 3.4 5.6 1.0 6.7 0.5 0.4 2.6 1.6 2.4 9.3 7.3 6.4 2.8 1.5 0.0 4.3 8.3 3.4 0.5 3.5 8.1 3.6 4.6 5.1 9.7 3.5 7.9 5.1 3.7 7.8 2.6 3.2 6.3 y = 1.6 2.7 4.4 0.5 0.2 5.6 6.7 training set test set from sklearn.cross_validation import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y)
17. 17 Supervised Machine Learning Training Data Training Labels Model
18. 18 Supervised Machine Learning Training Data Test Data Training Labels Model Prediction
19. 19 Supervised Machine Learning Training Data Test Data Training Labels Model Prediction Test Labels Evaluation
20. 20 Supervised Machine Learning Training Data Test Data Training Labels Model Prediction Test Labels Evaluation Training Generalization
21. 21 clf = RandomForestClassifier() clf.fit(X_train, y_train) Training Data Training Labels Model
22. 22 clf = RandomForestClassifier() clf.fit(X_train, y_train) Training Data Test Data Training Labels Model Prediction y_pred = clf.predict(X_test)
23. 23 clf = RandomForestClassifier() clf.fit(X_train, y_train) clf.score(X_test, y_test) Training Data Test Data Training Labels Model Prediction Test Labels Evaluation y_pred = clf.predict(X_test)
24. 24 Unsupervised Machine Learning Training Data Model
25. 25 Unsupervised Machine Learning Training Data Test Data Model New View
26. 26 pca = PCA() pca.fit(X_train) X_new = pca.transform(X_test) Training Data Test Data Model Transformation Unsupervised Transformations
27. 27 Basic API estimator.fit(X, [y]) estimator.predict estimator.transform Classification Preprocessing Regression Dimensionality reduction Clustering Feature selection Feature extraction
28. 28 Sample application: Sentiment Analysis
29. 29 Review: One of the worst movies I've ever rented. Sorry it had one of my favorite actors on it (Travolta) in a nonsense role. In fact, anything made sense in this movie. Who can say there was true love between Eddy and Maureen? Don't you remember the beginning of the movie ? Is she so lovely? Ask her daughters. I don't think so. Label: negative Training data: 12500 positive, 12500 negative IMDB Movie Reviews Data
30. 30 Bag Of Word Representations CountVectorizer / TfidfVectorizer
31. 31 Bag Of Word Representations “This is how you get ants.” CountVectorizer / TfidfVectorizer
32. 32 Bag Of Word Representations “This is how you get ants.” ['this', 'is', 'how', 'you', 'get', 'ants'] CountVectorizer / TfidfVectorizer tokenizer
33. 33 Bag Of Word Representations “This is how you get ants.” ['this', 'is', 'how', 'you', 'get', 'ants'] CountVectorizer / TfidfVectorizer tokenizer Build a vocabulary over all documents ['aardvak', 'amsterdam', 'ants', ... 'you', 'your', 'zyxst']
34. 34 Bag Of Word Representations “This is how you get ants.” [0, …, 0, 1, 0, … , 0, 1 , 0, …, 0, 1, 0, …., 0 ] ants get you aardvak zyxst ['this', 'is', 'how', 'you', 'get', 'ants'] CountVectorizer / TfidfVectorizer tokenizer Sparse matrix encoding Build a vocabulary over all documents ['aardvak', 'amsterdam', 'ants', ... 'you', 'your', 'zyxst']
35. 35 text_pipe = make_pipeline(CountVectorizer(), LinearSVC()) clf.fit(X_train, y_train) clf.score(X_test, y_test) > 0.85 Implementation and Results
36. 36 text_pipe = make_pipeline(CountVectorizer(), LinearSVC()) clf.fit(X_train, y_train) clf.score(X_test, y_test) > 0.85 Implementation and Results
37. 37 Model Complexity
38. 38 Overfitting and Underfitting Model complexity Accuracy Training
39. 39 Overfitting and Underfitting Model complexity Accuracy Training Generalization
40. 40 Overfitting and Underfitting Model complexity Accuracy Training Generalization Underfitting Overfitting Sweet spot
41. 41 Model Complexity Examples
42. 42 Linear SVM
43. 43 Linear SVM
44. 44 (RBF) Kernel SVM
45. 45 (RBF) Kernel SVM
46. 46 (RBF) Kernel SVM
47. 47 (RBF) Kernel SVM
48. 48 Decision Trees
49. 49 Decision Trees
50. 50 Decision Trees
51. 51 Decision Trees
52. 52 Decision Trees
53. 53 Decision Trees
54. 54 Random Forests
55. 55 Random Forests
56. 56 Random Forests
57. 57 Model Evaluation and Model Selection
58. 58 All Data Training data Test data
59. 59 All Data Training data Test data Fold 1 Fold 2 Fold 3 Fold 4 Fold 5
60. 60 All Data Training data Test data Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Split 1
61. 61 All Data Training data Test data Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Split 1 Split 2
62. 62 All Data Training data Test data Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Split 1 Split 2 Split 3 Split 4 Split 5
63. 63 Cross-Validation from sklearn.cross_validation import cross_val_score scores = cross_val_score(SVC(), X, y, cv=5) print(scores) >> [ 0.92 1. 1. 1. 1. ]
64. 64 SVC(C=0.001, gamma=0.001)
65. 65 SVC(C=0.001, gamma=0.001) SVC(C=0.01, gamma=0.001) SVC(C=0.1, gamma=0.001) SVC(C=1, gamma=0.001) SVC(C=10, gamma=0.001)
66. 66 SVC(C=0.001, gamma=0.001) SVC(C=0.01, gamma=0.001) SVC(C=0.1, gamma=0.001) SVC(C=1, gamma=0.001) SVC(C=10, gamma=0.001) SVC(C=0.001, gamma=0.01) SVC(C=0.01, gamma=0.01) SVC(C=0.1, gamma=0.01) SVC(C=1, gamma=0.01) SVC(C=10, gamma=0.01)
67. 67 SVC(C=0.001, gamma=0.001) SVC(C=0.01, gamma=0.001) SVC(C=0.1, gamma=0.001) SVC(C=1, gamma=0.001) SVC(C=10, gamma=0.001) SVC(C=0.001, gamma=0.01) SVC(C=0.01, gamma=0.01) SVC(C=0.1, gamma=0.01) SVC(C=1, gamma=0.01) SVC(C=10, gamma=0.01) SVC(C=0.001, gamma=0.1) SVC(C=0.01, gamma=0.1) SVC(C=0.1, gamma=0.1) SVC(C=1, gamma=0.1) SVC(C=10, gamma=0.1)
68. 68 SVC(C=0.001, gamma=0.001) SVC(C=0.01, gamma=0.001) SVC(C=0.1, gamma=0.001) SVC(C=1, gamma=0.001) SVC(C=10, gamma=0.001) SVC(C=0.001, gamma=0.01) SVC(C=0.01, gamma=0.01) SVC(C=0.1, gamma=0.01) SVC(C=1, gamma=0.01) SVC(C=10, gamma=0.01) SVC(C=0.001, gamma=0.1) SVC(C=0.01, gamma=0.1) SVC(C=0.1, gamma=0.1) SVC(C=1, gamma=0.1) SVC(C=10, gamma=0.1) SVC(C=0.001, gamma=1) SVC(C=0.01, gamma=1) SVC(C=0.1, gamma=1) SVC(C=1, gamma=1) SVC(C=10, gamma=1) SVC(C=0.001, gamma=10) SVC(C=0.01, gamma=10) SVC(C=0.1, gamma=10) SVC(C=1, gamma=10) SVC(C=10, gamma=10)
69. 69 All Data Training data Test data
70. 70 All Data Training data Test data Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Test data Split 1 Split 2 Split 3 Split 4 Split 5
71. 71 All Data Training data Test data Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Fold 1 Fold 2 Fold 3 Fold 4 Fold 5 Test data Finding Parameters Final evaluation Split 1 Split 2 Split 3 Split 4 Split 5
72. 72 Cross -Validated Grid Search from sklearn.grid_search import GridSearchCV from sklearn.cross_validation import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y) param_grid = {'C': 10. ** np.arange(-3, 3), 'gamma': 10. ** np.arange(-3, 3)} grid = GridSearchCV(SVC(), param_grid=param_grid) grid.fit(X_train, y_train) grid.score(X_test, y_test)
73. 73 Pipelines
74. 74 Training Data Training Labels Model
75. 75 Training Data Training Labels Model
76. 76 Training Data Training Labels Model Feature Extraction Scaling Feature Selection
77. 77 Training Data Training Labels Model Feature Extraction Scaling Feature Selection Cross Validation and Parameter selection
78. 78 Training Data Training Labels Model Feature Extraction Scaling Feature Selection Cross Validation and Parameter selection
79. 79 Pipelines from sklearn.pipeline import make_pipeline pipe = make_pipeline(StandardScaler(), SVC()) pipe.fit(X_train, y_train) pipe.predict(X_test)
80. 80 Pipelines from sklearn.pipeline import make_pipeline pipe = make_pipeline(StandardScaler(), SVC()) pipe.fit(X_train, y_train) pipe.predict(X_test)
81. 81 Combining Pipelines and Grid Search Proper cross-validation param_grid = {'svc__C': 10. ** np.arange(-3, 3), 'svc__gamma': 10. ** np.arange(-3, 3)} scaler_pipe = make_pipeline(StandardScaler(), SVC()) grid = GridSearchCV(scaler_pipe, param_grid=param_grid, cv=5) grid.fit(X_train, y_train)
82. 82 Combining Pipelines and Grid Search II Searching over parameters of the preprocessing step param_grid = {'selectkbest__k': [1, 2, 3, 4], 'svc__C': 10. ** np.arange(-3, 3), 'svc__gamma': 10. ** np.arange(-3, 3)} scaler_pipe = make_pipeline(SelectKBest(), SVC()) grid = GridSearchCV(scaler_pipe, param_grid=param_grid, cv=5) grid.fit(X_train, y_train)
83. 83 Do cross-validation over all steps jointly. Keep a separate test set until the very end.
84. 84 Scoring Functions
85. 85 Default: Accuracy (classification) R2 (regression) GridSeachCV cross_val_score
86. 86 Scoring with imbalanced data cross_val_score(SVC(), X_train, y_train) >>> array([ 0.9, 0.9, 0.9])
87. 87 Scoring with imbalanced data cross_val_score(SVC(), X_train, y_train) >>> array([ 0.9, 0.9, 0.9]) cross_val_score(DummyClassifier("most_frequent"), X_train, y_train) >>> array([ 0.9, 0.9, 0.9])
88. 88 Scoring with imbalanced data cross_val_score(SVC(), X_train, y_train) >>> array([ 0.9, 0.9, 0.9]) cross_val_score(DummyClassifier("most_frequent"), X_train, y_train) >>> array([ 0.9, 0.9, 0.9]) cross_val_score(SVC(), X_train, y_train, scoring="roc_auc") >>> array([ 1,0, 1.0, 1,0])
89. 89 Scoring with imbalanced data cross_val_score(SVC(), X_train, y_train) >>> array([ 0.9, 0.9, 0.9]) cross_val_score(DummyClassifier("most_frequent"), X_train, y_train) >>> array([ 0.9, 0.9, 0.9]) cross_val_score(SVC(), X_train, y_train, scoring="roc_auc") >>> array([ 1,0, 1.0, 1,0])
90. 90 Video Series Advanced Machine Learning with scikit-learn 50% Off Coupon Code: AUTHD
91. 91 Video Series Advanced Machine Learning with scikit-learn 50% Off Coupon Code: AUTHD
92. 92 CDS is hiring Research Engineers Work on your favorite data science open source project full time!
93. 93 Thank you for your attention. @t3kcit @amueller importamueller@gmail.com http://amueller.github.io