Features Engineering Art using R
•Download as PPTX, PDF•
3 likes•1,218 views
This document discusses feature engineering for machine learning models. It begins with an agenda and overview of what will be covered, including why feature engineering is needed, what feature engineering is, and the machine learning and feature engineering processes. It then discusses feature engineering techniques like manual construction, automatic construction, and feature selection. It provides examples of feature transformations, merging, splitting, and crafting features by hand. It notes that feature engineering is an art that requires turning data into representations models can understand. The document concludes with a workshop example using the Titanic dataset.
Recommended
More Related Content
Similar to Features Engineering Art using R
Similar to Features Engineering Art using R (20)
Recently uploaded
Recently uploaded (20)
Features Engineering Art using R
- 1. FEATURE ENGINEERING ART USING R MOHAMMED ALI MACHINE LEARNING ENG.
- 2. AGENDA What we will cover Why we need Feature Engineering What is Feature Engineering Art? Machine Learning Process Feature Engineering Process Manual Construction Automatic Construction Feature Selection Workshop
- 3. EXPECTATIONS What we will NOT cover The Iceberg Illusion What we will cover What we will NOT cover
- 5. OVERVIEW What is Feature Engineering Art? Feature Engineering Art
- 6. OVERVIEW What is Feature Engineering Art? Feature Feature is an individual measurable property or characteristic of a phenomenon being observed. Raw Data Feature Model Insight Feature Engineering Art
- 7. OVERVIEW What is Feature Engineering Art? Feature engineering is manually designing what the input x’s should be. (Tomasz Malisiewicz) You have to turn your inputs into things the algorithm can understand. (Shayne Miel) Feature Engineering Art
- 8. OVERVIEW What is Feature Engineering Art? Transformation Feature Engineering Art
- 9. OVERVIEW What is Feature Engineering Art? Transformation Feature Engineering Art
- 10. OVERVIEW What is Feature Engineering Art? Transformation Merging + = Feature Engineering Art
- 11. OVERVIEW What is Feature Engineering Art? Transformation Merging + = Feature Engineering Art
- 12. OVERVIEW What is Feature Engineering Art? MergingTransformation Splitting Feature Engineering Art Day Active Night Active
- 13. OVERVIEW What is Feature Engineering Art? Merging Crafting Transformation Splitting Feature Engineering Art
- 14. OVERVIEW What is Feature Engineering Art? Merging Crafting Transformation Splitting You have to turn your inputs into things the algorithm can understand. Shayne Miel Feature Engineering Art
- 15. OVERVIEW What is Feature Engineering Art? Feature Engineering Art Is Feature Engineering an ART ?
- 16. OVERVIEW What is Feature Engineering Art? Feature Engineering Art YES, you need to create the right question to get the right answer
- 17. OVERVIEW What is Feature Engineering Art? Feature Engineering Art
- 20. HOW TO ENGINEER Manual Construction + + + Base Data Frame Wednesday, March 12,2018, 12:20:59 am Features built one at a time by hand from given data set (S) using domain knowledge 123L - 12.5kg - 15$Simple Text
- 21. HOW TO ENGINEER Deep Feature Synthesis Features are derived from relationships between the data points in a dataset. Across datasets, many features are derived by using similar mathematical operations. New features are often composed from utilizing previously derived features. Problem specific. Error-prone. Limited both by human creativity and patience.
- 23. DISCUSSION What to use and when ………
- 24. FEATURE SELECTION Why Feature Selection ? Models Simplification. Shorter Training Times. Improve Model Accuracy. Enhanced Generalization by Reducing Overfitting.
- 25. FEATURE SELECTION Feature Selection Methods Embedded MethodsFilter Methods Wrapper Methods Feature Selection Methods Filter Methods considers the relationship between features and the target variable to compute the importance of features. In wrapper methods, we try to use a subset of features and train a model using them. Based on the inferences that we draw from the previous model, we decide to add or remove features from your subset.Forward Selection Backward Selection Recursive Feature Elimination Feature selection can also be acheived by the insights provided by some Machine Learning models.
- 26. WORKSHOP Titanic A Disaster to Learn From …
- 27. WORKSHOP Titanic
- 28. WORKSHOP Titanic Feature 1 What is in Name? Title Feature 2 Family swim together ? Maybe Feature 3 Can I buy new life ? Most probably NO
- 29. WORKSHOP Startups Chase the Vision not the Money …
- 30. QUESTIONS
Editor's Notes
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- transformation: a: invoice date season, Occasions. Weekend b. Age categorical (young, teenage, adult, senior) 2. Merging A: weight, height BMI b. Startup date, yearly profilt or loose - growth rate 3. Splitting a.Call records day active or night active 4. Crafting feature Scrapping social media for customer offers (card ahlawy)
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs
- Which technique to be applied and when You should have vision to extract features Know the right question to get the right answer 3awz azwd arba7, ytl3 mno el seasons w el holidays A3ml sales ll setat aw fe2a mo3yna A customize bona2n 3la customer needs