intro to machine learning for web folks @ blendwebmix
TRANSCRIPT
![Page 1: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/1.jpg)
Intro to Machine Learningfor web folks
“Machine Learning, je m’y mets dès demain”
@louisdorard #blendwebmix 2015
![Page 2: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/2.jpg)
–Mike Gualtieri, Principal Analyst at Forrester
“Predictive apps are the next big thing
in app development.”
![Page 3: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/3.jpg)
Lars Trieloff
@trieloff
(see source)
![Page 4: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/4.jpg)
–Waqar Hasan, VISA
“Predictive is the ‘killer app’ for big data.”
![Page 5: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/5.jpg)
Amazon for David Jones (@d_jones, see source)
![Page 6: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/6.jpg)
Amazon for David Jones (@d_jones, see source)
![Page 7: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/7.jpg)
1. Machine Learning
2. Data
![Page 8: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/8.jpg)
TECH ??
![Page 9: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/9.jpg)
–Charles Parker, PhD, Allston Trading
“ML isn’t about ML”
![Page 10: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/10.jpg)
BIZ
DESIGN
RECH
code
code
code
![Page 11: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/11.jpg)
BLEND !!
![Page 12: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/12.jpg)
@louisdorard
![Page 13: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/13.jpg)
![Page 14: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/14.jpg)
“Where makers of Predictive APIs and apps meet”
![Page 15: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/15.jpg)
Machine Learning Use cases
Limitations Modern tools
Case study ML Canvas
![Page 16: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/16.jpg)
DemystifyingMachine Learning
![Page 17: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/17.jpg)
![Page 18: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/18.jpg)
“Which type of email is this?
— Spam/Ham”
![Page 19: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/19.jpg)
“Which type of email is this?
— Spam/Ham”
⇒ Classification
![Page 20: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/20.jpg)
I
O
“Which type of email is this?
— Spam/Ham”
![Page 21: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/21.jpg)
??
![Page 22: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/22.jpg)
![Page 23: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/23.jpg)
“How much is this house worth?
— X $”
-> Regression
![Page 24: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/24.jpg)
Bedrooms Bathrooms Surface (foot²) Year built Type Price ($)
3 1 860 1950 house 565,0003 1 1012 1951 house2 1.5 968 1976 townhouse 447,0004 1315 1950 house 648,0003 2 1599 1964 house3 2 987 1951 townhouse 790,0001 1 530 2007 condo 122,0004 2 1574 1964 house 835,0004 2001 house 855,0003 2.5 1472 2005 house4 3.5 1714 2005 townhouse2 2 1113 1999 condo1 769 1999 condo 315,000
![Page 25: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/25.jpg)
Bedrooms Bathrooms Surface (foot²) Year built Type Price ($)
3 1 860 1950 house 565,0003 1 1012 1951 house2 1.5 968 1976 townhouse 447,0004 1315 1950 house 648,0003 2 1599 1964 house3 2 987 1951 townhouse 790,0001 1 530 2007 condo 122,0004 2 1574 1964 house 835,0004 2001 house 855,0003 2.5 1472 2005 house4 3.5 1714 2005 townhouse2 2 1113 1999 condo1 769 1999 condo 315,000
![Page 26: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/26.jpg)
![Page 27: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/27.jpg)
ML is a set of AI techniques where “intelligence” is built by referring to
examples
![Page 28: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/28.jpg)
![Page 29: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/29.jpg)
Use cases
![Page 30: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/30.jpg)
• Real-estate
• Spam
• Priority inbox
• Crowd prediction
property price
email spam indicator
email importance indicator
location & context #people
Zillow
Gmail
Gmail
Tranquilien
![Page 31: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/31.jpg)
I. Get more customers • Reduce churn
• Score leads
• Optimize campaigns
customer churn indicator
customer revenue
customer & campaign interest indicator
![Page 32: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/32.jpg)
II. Serve customers better • Cross-sell
• Increase engagement
• Optimize pricing
customer & product purchase indicator
user & item interest indicator
product & price #sales
![Page 33: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/33.jpg)
III. Serve customers more efficiently • Predict demand
• Automate tasks
• Use predictive enterprise apps
context demand
credit application repayment indicator
![Page 34: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/34.jpg)
Predictive enterprise apps • Priority filtering
• Message routing
• Auto-configuration
message priority indicator
request employee
user & actions settings
RULES
![Page 35: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/35.jpg)
–Katherine Barr, Partner at VC-firm MDV
"Pairing human workers with machine learning and automation
will transform knowledge work and unleash new levels of human
productivity and creativity."
![Page 36: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/36.jpg)
Limitations
![Page 37: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/37.jpg)
![Page 38: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/38.jpg)
![Page 39: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/39.jpg)
![Page 40: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/40.jpg)
![Page 41: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/41.jpg)
Need examples of inputs AND outputs
![Page 42: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/42.jpg)
![Page 43: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/43.jpg)
What if not enough data points?
![Page 44: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/44.jpg)
![Page 45: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/45.jpg)
What if similar inputs have dissimilar outputs?
![Page 46: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/46.jpg)
![Page 47: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/47.jpg)
Bedrooms Bathrooms Price ($)
3 2 500,0003 2 800,0001 1 300,0001 1 800,000
![Page 48: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/48.jpg)
Bedrooms Bathrooms Surface (foot²) Year built Price ($)
3 2 800 1950 500,0003 2 1000 1950 800,0001 1 500 1950 300,0001 1 500 2014 800,000
![Page 49: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/49.jpg)
• Need examples of inputs AND outputs
• Need enough examples
• Need enough “features”
![Page 50: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/50.jpg)
–@louisdorard
“A model can only be as good as the data it was given to train on”
![Page 51: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/51.jpg)
–McKinsey & Co. (2011)
“A significant constraint on realizing value from big data will
be a shortage of talent, particularly of people with deep
expertise in statistics and machine learning.”
![Page 52: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/52.jpg)
MLaaS & Predictive APIs: ML for all
![Page 53: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/53.jpg)
![Page 54: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/54.jpg)
HTML / CSS / JavaScript
![Page 55: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/55.jpg)
HTML / CSS / JavaScript
![Page 57: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/57.jpg)
![Page 58: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/58.jpg)
![Page 59: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/59.jpg)
The two phases of machine learning:
• TRAIN a model
• PREDICT with a model
![Page 60: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/60.jpg)
The two methods of predictive APIs:
• TRAIN a model
• PREDICT with a model
![Page 61: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/61.jpg)
The two methods of predictive APIs:
• model = create_model(dataset)
• predicted_output = create_prediction(model, new_input)
![Page 62: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/62.jpg)
The two methods of predictive APIs:
• model = create_model(‘training.csv’)
• predicted_output = create_prediction(model, new_input)
![Page 63: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/63.jpg)
![Page 64: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/64.jpg)
![Page 65: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/65.jpg)
![Page 66: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/66.jpg)
From Large to Small & Medium Enterprises
• recommendations in e-commerce
• => 71% increase in revenue
• churn detection
• => 11% increase in retention
![Page 68: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/68.jpg)
Microsoft Azure ML
![Page 69: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/69.jpg)
PredictionIO
![Page 70: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/70.jpg)
![Page 71: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/71.jpg)
Case study:churn analysis
![Page 72: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/72.jpg)
• Who: SaaS company selling monthly subscription
• Question asked: “Is this customer going to leave within 1 month?”
• Input: customer
• Output: no-churn or churn
• Data collection: history up until 1 month ago
• Baseline: if no usage for more than 15 days then churn
![Page 73: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/73.jpg)
Learning: OK
but
• How to represent customers?
• What to do after predicting churn?
![Page 74: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/74.jpg)
Customer representation:
• basic info (age, income, etc.)
• usage of service (# times used app, avg time spent, features used, etc.)
• interactions with customer support (how many, topics of questions, satisfaction ratings)
![Page 75: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/75.jpg)
Taking action to prevent churn:
• contact customers (in which order?)
• switch to different plan
• give special offer
• no action?
![Page 76: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/76.jpg)
Measuring accuracy:
• #TP (we predict customer churns and he does)
• #FP (we predict customer churns but he doesn’t)
• #FN (we predict customer doesn’t churn but he does)
• Compare to heuristic/baseline
![Page 77: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/77.jpg)
Return On Investment:
• Taking action for each TP (and FP) has a cost
• For each TP we “gain”: (success rate of action) * (revenue /cust. /month)
• Imagine…
• perfect predictions
• revenue /cust. /month = 10€
• success rate of action = 20%
• cost of action = 2€
• Which ROI?
![Page 78: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/78.jpg)
Machine Learning Canvas
![Page 79: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/79.jpg)
![Page 80: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/80.jpg)
Machine Learning Canvas
PREDICTIONS OBJECTIVES DATA
Context
Who will use the predictive system / who will beaffected by it? Provide some background.
Value Proposition
What are we trying to do? E.g. spend less time onX, increase Y...
Data Sources
Where do/can we get data from? (internaldatabase, 3rd party API, etc.)
Problem
Question to predict answers to (in plain English)
Input (i.e. question "parameter")
Possible outputs (i.e. "answers")
Type of problem (e.g. classification, regression,recommendation...)
Baseline
What is an alternative way of making predictions(e.g. manual rules based on feature values)?
Performance evaluation
Domain-specific / bottom-line metrics formonitoring performance in production
Prediction accuracy metrics (e.g. MSE ifregression; % accuracy, #FP for classification)
Offline performance evaluation method (e.g.cross-validation or simple training/test split)
Dataset
How do we collect data (inputs and outputs)?How many data points?
Features
Used to represent inputs and extracted fromdata sources above. Group by types andmention key features if too many to list all.
Using predictions
When do we make predictions and how many?
What is the time constraint for making those predictions?
How do we use predictions and confidence values?
Learning predictive models
When do we create/update models? With which data / how much?
What is the time constraint for creating a model?
Criteria for deploying model (e.g. minimum performance value — absolute,relative to baseline or to previous model)
IDE
AS
PE
CS
DE
PLO
YM
EN
T
![Page 81: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/81.jpg)
BACKGROUND
ENGINE SPECS
INTEGRATION
![Page 82: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/82.jpg)
PREDICTIONS OBJECTIVES DATA
BACKGROUND
ENGINE SPECS
INTEGRATION
![Page 83: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/83.jpg)
PREDICTIONS OBJECTIVES DATA
BACKGROUND End-user Value prop Sources
ENGINE SPECS ML problem Perf eval Preparation
INTEGRATION Using pred Learning modelINTEGRATION Using pred Learning model
![Page 84: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/84.jpg)
Why fill in ML canvas?
• Target the right problem for your company
• Choose right algorithm, infrastructure, or ML solution
• Guide project management
• Improve team communication
![Page 86: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/86.jpg)
Recap
![Page 87: Intro to machine learning for web folks @ BlendWebMix](https://reader036.vdocument.in/reader036/viewer/2022062823/5878c8501a28ab26728b66cf/html5/thumbnails/87.jpg)
• ML to create value from data
• 2 phases: TRAIN and PREDICT
• MLaaS & Predictive APIs make it more accessible
• Good data is essential
• What do we do with predictions?
• Accuracy is not the objective! A/B test?
• Start with the ML Canvas
• Later: deploy, maintain, improve…