Title: Bayesian%20Network%20Anomaly%20Pattern%20Detection%20for%20Disease%20Outbreaks
1Bayesian Network Anomaly Pattern Detection for
Disease Outbreaks
- Weng-Keen Wong (Carnegie Mellon University)
- Andrew Moore (Carnegie Mellon University)
- Gregory Cooper (University of Pittsburgh)
- Michael Wagner (University of Pittsburgh)
2Motivation
Suppose we have real-time access to Emergency
Department data from hospitals around a city
(with patient confidentiality preserved)
Primary Key Date Time Hospital ICD9 Prodrome Gender Age Home Location Work Location Many more
100 6/1/03 912 1 781 Fever M 20s NE ?
101 6/1/03 1045 1 787 Diarrhea F 40s NE NE
102 6/1/03 1103 1 786 Respiratory F 60s NE N
103 6/1/03 1107 2 787 Diarrhea M 60s E ?
104 6/1/03 1215 1 717 Respiratory M 60s E NE
105 6/1/03 1301 3 780 Viral F 50s ? NW
106 6/1/03 1305 3 487 Respiratory F 40s SW SW
107 6/1/03 1357 2 786 Unmapped M 50s SE SW
108 6/1/03 1422 1 780 Viral M 40s ? ?
3The Problem
From this data, can we detect if a disease
outbreak is happening?
4The Problem
From this data, can we detect if a disease
outbreak is happening?
Were talking about a non-specific disease
detection
5The Problem
From this data, can we detect if a disease
outbreak is happening? How early can we detect
it?
6The Problem
From this data, can we detect if a disease
outbreak is happening? How early can we detect
it?
The question were really asking
Whats strange about recent events?
7Traditional Approaches
- What about using traditional anomaly detection?
- Typically assume data is generated by a model
- Finds individual data points that have low
probability with respect to this model - These outliers have rare attributes or
combinations of attributes
- Need to identify anomalous patterns not isolated
data points
8Traditional Approaches
What about monitoring aggregate daily counts of
certain attributes?
- Weve now turned multivariate data into
univariate data - Lots of algorithms have been developed for
monitoring univariate data
- Time series algorithms
- Regression techniques
- Statistical Quality Control methods
- Need to know apriori which attributes to form
daily aggregates for!
9Traditional Approaches
- What if we dont know what attributes to monitor?
- What if we want to exploit the spatial, temporal
and/or demographic characteristics of the
epidemic to detect the outbreak as early as
possible?
10Traditional Approaches
- We need to build a univariate detector to monitor
each interesting combination of attributes
Diarrhea cases among children
Number of cases involving people working in
southern part of the city
Respiratory syndrome cases among females
Number of cases involving teenage girls living
in the western part of the city
Viral syndrome cases involving senior citizens
from eastern part of city
Botulinic syndrome cases
Number of children from downtown hospital
And so on
11Traditional Approaches
- We need to build a univariate detector to monitor
each interesting combination of attributes
Diarrhea cases among children
Number of cases involving people working in
southern part of the city
Respiratory syndrome cases among females
Number of cases involving teenage girls living
in the western part of the city
Youll need hundreds of univariate detectors! We
would like to identify the groups with the
strangest behavior in recent events.
Viral syndrome cases involving senior citizens
from eastern part of city
Botulinic syndrome cases
Number of children from downtown hospital
And so on
12One Possible Approach
Primary Key Date Time Gender Age Hospital Many more
100 8/24/03 912 M 20s 1
101 8/24/03 1045 F 40s 1
2243 8/17/03 1107 M 60s 2
2244 8/17/03 1215 M 60s 1
12567 8/24/02 1305 F 40s 3
12568 8/24/02 1357 M 50s 2
Todays Records
Yesterdays Records
Last Years Records
13One Possible Approach
Primary Key Date Time Gender Age Hospital Many more
100 8/24/03 912 M 20s 1
101 8/24/03 1045 F 40s 1
2243 8/17/03 1107 M 60s 2
2244 8/17/03 1215 M 60s 1
12567 8/24/02 1305 F 40s 3
12568 8/24/02 1357 M 50s 2
Todays Records
Yesterdays Records
Last Years Records
Idea Can use association rules to find patterns
in todays records that werent there in past data
14One Possible Approach
Recent records ( from today )
Primary Key Date Time Gender Age
100 8/24/03 912 M Child
101 8/24/03 1045 M Senior
Primary Key Date Time Source
100 8/24/03 912 Recent
101 8/24/03 1045 Recent
2164 8/17/03 1305 Baseline
2165 8/17/03 1357 Baseline
Baseline records ( from 7 days ago )
Primary Key Date Time Gender Age
2164 8/17/03 1305 F Senior
2165 8/17/03 1357 F Senior
Find which rules predict unusually high
proportions in recent records when compared to
the baseline eg. 52/200 records from recent
have Gender Male AND Age Senior 90/180
records from baseline have Gender Male AND
Age Senior
15Which rules do we report?
- Search over all rules with at most 2 components
- For each rule, form a 2x2 contingency table eg.
- Perform Fishers Exact Test to get a p-value for
each rule (call this the score) - Report the rule with the lowest score
CountRecent CountBaseline
Home Location NW 48 45
Home Location ? NW 86 220
16Problems with the Approach
- Multiple Hypothesis Testing
- 2. A Changing Baseline
17Problem 1 Multiple Hypothesis Testing
- Cant interpret the rule scores as p-values
- Suppose we reject null hypothesis when score lt ?,
where ? 0.05 - For a single hypothesis test, the probability of
making a false discovery ? - Suppose we do 1000 tests, one for each possible
rule - Probability(false discovery) could be as bad as
- 1 ( 1 0.05)1000 gtgt 0.05
18Randomization Test
Aug 16, 2003 C2
Aug 17, 2003 C3
Aug 17, 2003 C4
Aug 17, 2003 C5
Aug 17, 2003 C6
Aug 17, 2003 C7
Aug 21, 2003 C8
Aug 21, 2003 C9
Aug 22, 2003 C10
Aug 22, 2003 C11
Aug 23, 2003 C12
Aug 23, 2003 C13
Aug 24, 2003 C14
Aug 24, 2003 C15
Aug 16, 2003 C2
Aug 17, 2003 C3
Aug 24, 2003 C4
Aug 17, 2003 C5
Aug 24, 2003 C6
Aug 17, 2003 C7
Aug 21, 2003 C8
Aug 21, 2003 C9
Aug 22, 2003 C10
Aug 22, 2003 C11
Aug 23, 2003 C12
Aug 23, 2003 C13
Aug 17, 2003 C14
Aug 17, 2003 C15
- Take the recent cases and the baseline cases.
Shuffle the date field to produce a randomized
dataset called DBRand - Find the rule with the best score on DBRand.
19Randomization Test
Repeat the procedure on the previous slide for
1000 iterations. Determine how many scores from
the 1000 iterations are better than the original
score.
If the original score were here, it would place
in the top 1 of the 1000 scores from the
randomization test. We would be impressed and an
alert should be raised.
Corrected p-value of the rule is better scores
/ iterations
20Problem 2 A Changing Baseline
From Goldenberg, A., Shmueli, G., Caruana, R.
A., and Fienberg, S. E. (2002). Early
statistical detection of anthrax outbreaks by
tracking over-the-counter medication sales.
Proceedings of the National Academy of Sciences
(pp. 5237-5249)
21Problem 2 A Changing Baseline
- Baseline is affected by temporal trends in health
care data - Seasonal effects in temperature and weather
- Day of Week effects
- Holidays
- Etc.
- Choosing the wrong baseline distribution can
affect the detection time and false positives rate
22Generating the Baseline
- Taking into account that today is a public
holiday - Taking into account that this is Spring
- Taking into account recent heatwave
- Taking into account recent flu levels
- Taking into account that theres a known natural
Food-borne outbreak in progress
23Generating the Baseline
- Taking into account that today is a public
holiday - Taking into account that this is Spring
- Taking into account recent heatwave
- Taking into account recent flu levels
- Taking into account that theres a known natural
Food-borne outbreak in progress
Use a Bayes net to model the joint probability
distribution of the attributes
24Obtaining Baseline Data
All Historical Data
- Learn Bayesian Network using Optimal Reinsertion
Moore and Wong 2003
Todays Environment
Baseline
2. Generate baseline given todays environment
25Environmental Attributes
- Divide the data into two types of attributes
- Environmental attributes attributes that cause
trends in the data eg. day of week, season,
weather, flu levels - Response attributes all other non-environmental
attributes
26Environmental Attributes
- When learning the Bayesian network structure, do
not allow environmental attributes to have
parents. - Why?
- We are not interested in predicting their
distributions - Instead, we use them to predict the distributions
of the response attributes - Side Benefit We can speed up the structure
search by avoiding DAGs that assign parents to
the environmental attributes
Season
Day of Week
Weather
Flu Level
27Generate Baseline Given Todays Environment
Suppose we know the following for today
Season Day of Week Weather Flu Level
Today Winter Monday Snow High
Day of Week Monday
Flu Level High
Season Winter
Weather Snow
We fill in these values for the environmental
attributes in the learned Bayesian network
We sample 10000 records from the Bayesian network
and make this data set the baseline
Baseline
28Generate Baseline Given Todays Environment
Suppose we know the following for today
Season Day of Week Weather Flu Level
Today Winter Monday Snow High
Day of Week Monday
Flu Level High
Season Winter
Weather Snow
We fill in these values for the environmental
attributes in the learned Bayesian network
Sampling is easy because environmental attributes
are at the top of the Bayes Net
We sample 10000 records from the Bayesian network
and make this data set the baseline
Baseline
29Generate Baseline Given Todays Environment
Suppose we know the following for today
Season Day of Week Weather Flu Level
Today Winter Monday Snow High
Day of Week Monday
Flu Level High
Season Winter
Weather Snow
We fill in these values for the environmental
attributes in the learned Bayesian network
An alternate possible technique is to use
inference
We sample 10000 records from the Bayesian network
and make this data set the baseline
Baseline
30Whats Strange About Recent Events (WSARE) 3.0
2. Search for rule with best score
- Obtain Recent and Baseline datasets
All Data
Recent Data
- Determine p-value of best scoring rule
Baseline
4. If p-value is less than threshold, signal alert
31Simulator
32Simulation
- 100 different data sets
- Each data set consisted of a two year period
- Anthrax release occurred at a random point during
the second year - Algorithms allowed to train on data from the
current day back to the first day in the
simulation - Any alerts before actual anthrax release are
considered a false positive - Detection time calculated as first alert after
anthrax release. If no alerts raised, cap
detection time at 14 days
33Other Algorithms used in Simulation
1. Standard algorithm
Signal
2. WSARE 2.0 Create baseline using historical
data from 7, 14, 21 and 28 days ago 3. WSARE 2.5
Use all past data but condition on
environmental attributes
34Results on Simulation
35Results on Simulation
36Results on Actual ED Data from 2001
- 1. Sat 2001-02-13 SCORE -0.00000004 PVALUE
0.00000000 - 14.80 ( 74/500) of today's cases have Viral
Syndrome True and Encephalitic Prodome False - 7.42 (742/10000) of baseline have Viral
Syndrome True and Encephalitic Syndrome False - 2. Sat 2001-03-13 SCORE -0.00000464 PVALUE
0.00000000 - 12.42 ( 58/467) of today's cases have
Respiratory Syndrome True - 6.53 (653/10000) of baseline have
Respiratory Syndrome True - 3. Wed 2001-06-30 SCORE -0.00000013 PVALUE
0.00000000 - 1.44 ( 9/625) of today's cases have 100 lt
Age lt 110 - 0.08 ( 8/10000) of baseline have 100 lt Age
lt 110 - 4. Sun 2001-08-08 SCORE -0.00000007 PVALUE
0.00000000 - 83.80 (481/574) of today's cases have
Unknown Syndrome False - 74.29 (7430/10001) of baseline have Unknown
Syndrome False - 5. Thu 2001-12-02 SCORE -0.00000087 PVALUE
0.00000000 - 14.71 ( 70/476) of today's cases have Viral
Syndrome True and Encephalitic Syndrome False - 7.89 (789/9999) of baseline have Viral
Syndrome True and Encephalitic Syndrome False
37Related Work
- Deviations between models induced by two datasets
Ganti, Gehrke and Ramakrishnan - Emerging Patterns Dong and Li
- Mining Surprising Patterns using Temporal
Description Length Chakrabarti, Sarawagi and
Dom - Contrast sets Bay and Pazzani
- Association Rules and Data Mining in Hospital
Infection Control and Public Health Surveillance
Brossette et. al. - Spatial Scan Statistic Kulldorff
38Conclusion
- One approach to biosurveillance one algorithm
monitoring millions of signals derived from
multivariate data - instead of
- Hundreds of univariate detectors
- WSARE is best used as a general purpose safety
net in combination with other detectors - Careful evaluation of statistical significance
- Modeling historical data with Bayesian Networks
to allow conditioning on unique features of today
Software http//www.autonlab.org/