McMasterPandemic
1
Fast and Flexible Modelling with McMasterPandemic
1.1
History and Motivation
1.1.1
COVID-19 Forecasts
1.1.2
Calibration to data
1.1.3
Speed
1.1.4
Model Extensibility
1.2
Installation
1.3
Dependencies
1.4
Generalized Model at a Glance
1.5
Vision and Direction
2
Model Initialization
2.1
Initial Parameter Vector
2.2
Initial State Vector
2.3
Start and End Date
2.4
Next Steps
3
Flow Between States
3.1
State Flows
3.2
Flow Matrix
3.3
Rate Matrix
3.4
Rate Matrix Dependence on State Variables and Parameters
3.5
Connections with Classic McMasterPandemic
3.6
Topological Sort
4
Simulation
4.1
Observation Error
5
Calibration
5.1
Calibrating with Observation Error
5.2
Loss Function Theory
5.2.1
Negative Binomial
6
Convergence
7
Time Varying Parameters
7.1
Model of Piece-Wise Time-Variation
7.2
Calibrating Time-Variation Schedules
8
Other Variables
8.1
Intermediate Results
8.1.1
Sums of State Variables and Parameters
8.1.2
Factrs
8.1.3
Power Laws
8.2
Additional Variables in the Simulation History
8.2.1
Simulation History Expressions
8.2.2
Lagged Differencing
8.2.3
Convolutions
9
Ensemble Forecasts
9.1
Time-Varying Ensemble Forecasts
10
Vectors
11
Hazard Smoothing
12
State Initialization
13
Outflows
13.1
Accumulators
14
TMB Engine
15
Examples
15.1
Hello World: Simulating an SIR Model
15.2
SI
15.3
SEIR
15.4
Structure: Two-Strain SIR
15.5
Erlang SEIR
15.6
SIRV
15.7
Variolation model
15.8
SEIRD
15.9
Covid SEIR
15.10
BC Covid Omicron
15.11
Classic McMasterPandemic
15.12
Granich HIV Model
16
Troubleshooting
16.1
NaN-Valued Objective Function
16.1.1
Simulated time-series close to zero
16.1.2
Negative rate matrix elements
16.1.3
Optimizer tries very large dispersion parameters
16.2
Non-Positive Definite Covariance Matrix
Published with bookdown
Using McMasterPandemic
12
State Initialization