Society exists only as a mental concept; in the real world there are only individuals.
-- Oscar Wilde
Wearables: Watches, camera, …
I don’t care for money;I care for what money buys.
I don’t care for wearables;I care for what wearables do for me.
Disruption in Healthcare
When needed, I go to
the source of
healthcare.
Can healthcare come
to me in time?
Event MiningMachine Learning
Individual Model
‘Likely to get severe heart attack in 10 minutes – get him help immediately.’
• Individual Model from data.
• Health, social, personal.
• Actionable Predictive use.
• Better disease models.
Healthcare 2020
Objective Self
Important Revolution in Health
In the mid-20th century, the primary causes of death worldwide shifted from infections to chronic conditions.
Dream: Immortality
Problem: Longest people can live is 122 Yrs.
Realizable Dream:100+ Years of Happiness
Personal Health
What’s Lifestyle got to do with it?
High Blood Pressure
Heart Disease Allergies
Thyroid Imbalance
Diabetes
Chronic FatigueAuto Immune Disease
Cancer
StressPoor Diet
Lack of Sleep
Lack of ExerciseNutrition Deficiencies
GeneticsToxins
Needed: Medical Emancipation
Doctor Knows the Best.
Patient Knows the Best.
Smartphone + Wearables :=
Personalized 24/7 Recording Stethoscope
Financial
Physical
Environ-mental
Emotional
Sexual
Social
Intellectual
Spiritual
Lifestyle
Big 3: Lifestyle Factors
Physical Activities
Popular now: Output and state
FoodStarting to happen:Input
Environmental Factors
Coming soon: Surround
Healthcare Analytics is Data Rich Now
Past: Data is expensive and small• Input data is mostly from clinical trials• Models are small since data is limited• Personal information is anecdotal
Today: Data is cheap and large• Longitudinal data• Heterogeneous data• Diverse data from Electronic Health
Records and wearable/smartphone
Until recently, you
were a folder.
Now You are Your Data.
Disruption Time
“Shallow men believe in luck or in circumstance. Strong men believe in cause and effect.”
― Ralph Waldo Emerson
If you can't measure it, you can't control it!
Measure
Understand
Control
Improve
OBJECTIVE SELF (FUTURE WORK)OBJECTIVE SELF
Anecdotal
Diarizing data
Quantified Self
Data Streams to Objective Self
Daily Life Activities
Data Streams
Data Variety: From Data Streams to Abstracted Event Streams
Data Management
Event Streams
20
Event Stream
Daily Life
A
ctivity
ATUS: American Time Use Survey
Daily Life Activities: How we enter them?
Chronicle of Daily Life Activities
We collect diverse signals.
Running Example
25
Daily EventsWalkingMeeting Break A
rriv
e H
om
e
Exercise Asthma Attack
Pollution Events
91 4 6 10 15 19
Low
Temperature Events
Increase Suddenly High
Decrease SteadilyMedium Low
7 14
Leav
e H
om
e
Framework
Heart rate Events
NormalHigh Elevated
Definitions• Time Interval: [∂, ts, te] ∂+ = ts , ∂− = te
• Semi Interval: [∂+/−, t]
• Point Event (pE): e = (ν, [ E, t])
• Interval Event (iE): e = (v, [E, ts, te])
• Semi-interval Event (sE): e = (v, [E+/−, t])
26
Life EventsWalkingMeeting Break A
rriv
e H
om
e
Exercise Asthma Attack
91 4 6 10 15 197 14
Leav
e H
om
e
Framework
Definitions (Cont.)
• Event Stream: ES(i) = {e1(i), e2
(i), ..., en(i)}
• Multi-Event Stream: ES = {ES(1), ES(2), ..., ES(∣I∣)}
• Pattern: ρ = (X1 ⊙1 X2 ⊙2 ... ⊙k−1 Xk )
Xi ∈ { pE, iE, sE } and ⊙i ∈ { ; , ;ω∆t , , ∣ }
27
Life EventsWalkingMeeting Break A
rriv
e H
om
e
Exercise AsthmaAttack
91 4 6 10 15 197 14
Leav
e H
om
e
( Meeting ; Break)
( Exercise+ ;ω[5] AsthmaAttack)
( Walking ; ArriveHome)
Framework
Representations
Laleh Jalali [email protected]
Interactive Knowledge Discovery from Data Streams
28
Daily EventsWalkingMeeting Break A
rriv
e H
om
e
Exercise Asthma Attack
Pollution Events
91 4 6 10 15 19
1
Low
8 10
24
Temperature Events
Increase Suddenly High
1
Decrease SteadilyMedium Low
247 14
7 14
Leav
e H
om
e
(( Exercise+ Pollution.High) ;ω[5] AsthmaAttack)
(( Exercise+ Temp.Low ) ;ω[5] AsthmaAttack)
(( Exercise+ (Pollution.High | Temp.Low )) ;ω[5] AsthmaAttack)
Framework
Pattern Mining Operators
Sequential Co-occurrence
Concurrent Co-occurrence
29Laleh Jalali [email protected]
Interactive Knowledge Discovery from Data Streams
Framework
0.25 0.31 0.47 0.11 0.68 0
0.01 0.33 0.64 0.19 0 0.16
0.52 0.12 0.09 0 0.11
0.67 0.1 0 0.7 0.03 0.52
0.13 0 0.75 0.71 0.23 0.25
0 0.43 0.66 0.1 0.2 0.35
ΔtE1 E2 E3 E4 E5 E6
E1
E2
E3
E4
E5
E6
0.88
Co-occurrence Matrix Visualization
0.25
0.31 0.52 0.11 0.23 00.33
0.23 0.4 0.28 0 0.23
0.1 0.45 0 0.11
0.56 0.12 0 0.45 0.4 0.52
0.23 0 0.12 0.23 0.31
0 0.23 0.56 0.1 0.33 0.25
0.280.82
0.82
E1 E2 E3 E4 E5 E6
E1
E2
E3
E4
E5
E6
30
Cause - Effect Pattern Structure
event1 event2
event3
event4
Effect
Time lag between
events
Sequence of events
Events in parallel
Cause 1
Cause 2
Cause 3
(No medication ; Exercise) Asthma attack
(Exercise Pollen high) Asthma attack
(Exercise (Pollen high | pollution high)) Asthma attack
(Exercise Pollution high) Asthma attack
Exercise Asthma attackΔt
Formulate and query complex patterns:
Pattern Mining
High level Pattern
Formulation
Pattern Query
…
Data Streams Event Streams Semi-interval Event Sequences
Data-Driven Analysis
Hypothesis-Driven Analysis
While air pressure is high, pollution starts increasing
gradually, within T time units asthma outbreak happens.
((pollution_inc_steadily ;ωT asthma_outbreak ) ||
airpressure_stayshigh)
Example: Interactive Visualization
Interactive Event Mining
User Interface for Interactive Knowledge
Discovery and Model Building D
ata-
Dri
ven
Hyp
oth
esis
-
Dri
ven
Pollution and Meteorological Data
Asthma Risk Factor Recognition
1) Pollution increases
suddenly followed by high
wind while temperature
increases slightly will cause
an asthma
Outbreak within 2 days.
2) Thunderstorm followed
by temperature decreases
steadily will cause an
asthma outbreak within 1
day.
ResultsTemperature fluctuation has the most impact in the fall and winter seasons and it is not a risk factor during spring or summer
During spring and summer, when rain suddenly increases to a very high level, an asthma outbreak is more probable
The effect of PM2.5 is not noteworthy in the fall and winter seasons.
Results (Cont.)
• When PM2.5 increases followed by temperature stay high within 3 days,
then asthma outbreak is probable.
• When wind decreases followed by PM2.5 increases within 5 days, then
asthma outbreak is probable.
• When rain increases followed by PM2.5 stay low within 4 days then an
asthma outbreak is probable.
• Exposure to polluted air is a risk factor
of asthma attack within X hour?
Personicle case
Environmental factors
Spicy Indian food and 2 glasses of wine
result in severe acidity and sleepless nights.
Personicle
Food Stream
t1 t2 t3 t4 t5
40
Act Orient
Observe
Decide
Decision Making in Complex Situations
John BoydMilitary Strategist
OODA
Situation awareness is knowing what’s going on around you.
41
Observe + Orient = Situational Awareness
Orient: Baselines, Goals, and Action Plans
Situation awareness is required for every decision in life.
Act Orient
Observe
Decide
Dashboards Display Data and Information
42
Operators use relevant information to understand situation to decide relevant Action.
10/31/2016 43
Calendar PESi
FMB (Individual’s Feeling)
AccelerometerLocation
Fitness Data(Nike, Fitbit) Data
Ingestion & Aggregation
Heart Rate
Location (Move)
Food Log
FMB (People’s Feeling, Location)
ESOzoneCO2
SO2
PM 2.5
Pollen (Tree, Grass)
Air Quality Index
Data Ingestion & Aggregation
Social Media (News, Tweets)
Weather
Macro Situation Recognition
Predictive Analytics
PersonalSituation Recognition
Persona
Asthma Allergy App Server
Data Collection
Mac
ro S
itu
atio
nPe
rso
nal
Sit
uat
ion
Need and Resources Recommendation
Great Opportunity: Wearable
• Wearable building personal models.
• Using personal model to make people aware of their situation.
• Informing relevant people about the situation.
• Helping people make
Right Decision, Right Moment, Right Place.