HS 405 Emergent Topic in Homeland- Exam 2

Cone of Uncertainty

-Delineate possibilities and extend them out from a particular moment.

-Most important point is to define the breadth, which defines the degree of uncertainty

Forecasting

- Cone of Uncertainty

- Look for the S curve

- Embrace the things that don't fit

- Hold Strong Opinions weakly

-Look back twice as far as you look forward

-Know when not to make a forecast

S- Curve

- The point in the middle is the inflection point

- The stuff to the left is chock full of potential indicators

- The stuff on the right is the future

- There is a tendency to overestimate the short-term and underestimate the long-term

Bayes Law

- Conditional Probabilities

- Every clue has some error attached

- Subjective, offering a measurement of the degree of belief an event will occur

Pierre Simon LaPlace

-If we don't have reliable data by which to assess the likelihood of two or more outcomes, we should assume the probability is equal for all outcomes.

-But . . . We are subject to biases: - we are prone to believe a favored outcome is more likely - we are prone to value evidence that supports a favored outcome - We are prone hold this belief despite contrary evidence

Frequentist Probability

-Events probability defined as its relative frequency in a series of trials

-Claimed to be objective and verifiable; directly applicable to scientific experiments

-Requires an infinite number of trials

-You are assuming it's a fair trial

Probabilities

Accurate calculation requires knowledge of all the possibilities and casual conditions

Mutually Exclusive

events which preclude one another (coin toss)

Conditionally dependent

events where one event's occurrence depends upon another; events that occur in sequence

Probability Tree

-Assess the whole from the perspective of probability

-Estimate the most likely scenario and least likely scenario and which are most and least likely

Sensitivity analysis: We can, if needed, manipulate the probabilities - arbitrarily changing them - to help us focus our attention and concentrate effort on those events and outcomes for which we would like to increase the likelihood of their outcome

Sensitivity analysis

We can, if needed, manipulate the probabilities - arbitrarily changing them - to help us focus our attention and concentrate effort on those events and outcomes for which we would like to increase the likelihood of their outcome

Probability Tree the Three inviolable Rules

- Events depicted must be mutually exclusive

- Events must be collectively exhaustive

- Probabilities of the branch at each node must equal 1

Probability Tree Steps

- Identify the problems

- Identify the major events, causes, decisions to assess

- Construct a scenario tree

- Assign probabilities to each event/decision

- Calculate conditional probabilites

- Calculate and assess the question related to events

Heuristic

- Rule of Thumb

- A mental model

Aggregate Prediction

the prediction of outcomes based on group characteristics

The five V of big Data

Volume, Velocity, Variety, Veracity, and Value

The Five V of the Big Data EXTRA

- Verbalize- most data is unstructured and uncategorized data

- Verbosity- there is a lot of redundancy

- Versatility- data can be useful in different ways in different scenarios

- Viscosity- ease or difficulty in moving data

- Visibility- access control we apply to it

Initial condition uncertainty

The short term factors that affect whatever it is we are looking at

Scenario uncertainty

The uncertainty that increases overtime

Structural uncertainty

The uncertainty that is inherent in the system itself

Scenarios

They are instruments for ordering our predictions about alternative futures.

When is it a good time to use scenarios

High Degree of uncertainty and When the impact is large

When is it not a good time to use scenarios

Low uncertainty and when the impact is low

How do you know when the scenario is good?

If it passes the sniff test AKA it is plausible

Scenario Development

1. Identify the topic, issue, or question of importance

2. Define the system under consideration; Define the parameters; define the context

3. Identify the driving forces of change (causes of events)

4. Cluster or group the change drivers

5. Rank the variables or change drivers in terms of their relative importance or impact on the system (small can have a big impact and vice versa)

6. Rank the variables or change drivers in terms of their relative uncertainty or predictability

7. Develop the scenarios

8. Test the scenarios

9. Evaluate the scenarios

Power Laws

-Change begins slowly, incrementally and putters along quietly, suddenly explodes, takes off, then eventually tapers off to a trickle, maybe even backing down

-Characterized by Scale invariance: in the magnitude of the event increases, the number of events at that scale decreases proportionally

Survivorship bias

Focus on the winners and ignore the losers

Satisfying

Accepting the less than the optimal

Weak Signals buried in the noise

- Initially appear insignificant

- Appear as outliers or contradictory evidence

- Take on much greater significance over time, and in response to other events

- Can provide the signposts for significant system disruption and change, which is frequently glaringly obvious in hindsight

Judging probabilities

- Option 1: wide margins, high confidence, and low accuracy

- Option 2: Narrow Margins, High Accuracy, and low confidence