16. DSP1
● What do we know about the user?
● What are they interested in?
17. DSP1
● What do we know about the user?
● What are they interested in?
● What’s the probability of click and conversion?
18. DSP1
● What do we know about the user?
● What are they interested in?
● What’s the probability of click and conversion?
● How much are we willing to pay for it?
$0.10
19. DSP1
● What do we know about the user?
● What are they interested in?
● What’s the probability of click and conversion?
● How much are we willing to pay for it?
● Should we show any ad?
20. DSP1
● What do we know about the user?
● What are they interested in?
● What’s the probability of click and conversion?
● How much are we willing to pay for it?
● Should we show any ad?
24. Probability of conversion
● Device characteristics (e.g. OS)
● Geography (country, city)
● Demography (gender, age)
● History: visited pages, installed apps
● Features of the advertiser (how convenient the page is, etc)
39. Plan
● Use cases
○ Advertisement
○ Moderation in Online Classifieds
● Base skills
40. Base skills
● SQL, data manipulation
● Git
● Python
● NumPy, Pandas, Scikit-Learn
● Training and validating models
● Microservices, Flask, Docker
41. How to learn?
● Come up with a problem
● Look for solution (tools, libraries, tutorials)
● Solve the problem
● …
● Profit
42. How to learn?
● Come up with a problem ⇐ Important! Focus on the problem
● Look for solution (tools, libraries, tutorials)
● Solve the problem
● …
● Profit
43. mlbookcamp.com
● Learn ML by doing projects
● http://bit.ly/mlbookcamp
● Get 40% off with code “grigorevpc”
● Twitter: @Al_Grigor (book give-away
this Sunday!)
Machine Learning
Bookcamp