In Vitro Growth, Linear Modeling, and Epidemiological Simulations

Flashcards covering R programming functions for linear modeling, statistical evaluation metrics, and epidemiological SIR model parameters.

PFU_per_ml

A measure of viral growth used in experiments looking at the in vitro growth of SARS-CoV-2 variants.

lm()

An R function used to fit a linear model, specified with a formula like $y \sim x$.

Coefficients

Components of a linear model object, such as the intercept or the slope of an explanatory variable, accessed via $model_fit\$coefficients$.

Residuals

The difference between the actual data and the model's prediction; a positive value means the model underestimated the real data, while a negative value means an overestimate.

Fitted values

The predicted values generated by a linear model based on the input data, accessed via $model_fit\$fitted.values$.

Wave

A period in a pandemic characterized by a distinct rise, peak, and fall in hospital admissions.

A model fits training data too closely and fails to predict new data; memorising the “noise“.

vaccine_sentiment

The percentage of people who were vaccinated, had an appointment, or would get vaccinated if given the opportunity, based on a Facebook survey.

mask_sentiment

The percentage of people who wore a mask most or all of the time in public, based on a Facebook survey.

Sum of squared error

A way of measuring model fit defined as the sum of the squares of the residuals.

Mean Squared Error (MSE)

A measure of model performance calculated as the mean of the squared residuals: $\text{mean}(\text{residuals}^2)$. Smaller values indicate a better fit to the data.

Dependent variable

Also called the response variable, this is the outcome of interest that a model tries to predict (e.g., $CovDp100K$).

Independent variable

Also called predictors, these are the factors used within a model to explain variation in the dependent variable (e.g., $pPop65$, $VacFullp100$).

Cross-validation

A process used to evaluate a model's predictive power by splitting data into a training set to fit the model and a test set to evaluate performance.

Training data

The subset of a dataset used to build and fit the parameters of a specific model.

Test data

The subset of data used to evaluate how well a model predicts new, unseen observations.

set.seed()

An R function that ensures random number generation (such as splitting data into test and training sets) is reproducible for all users.

predict()

An R function used to generate predicted values for a test dataset based on a model that was previously fitted to training data.

SIR model

A compartmental model used to simulate disease dynamics based on three groups: Susceptibles ($S$), Infecteds ($I$), and Recovereds ($R$).

Transmission rate (beta)

In an SIR model, the parameter $\beta$ that represents the rate at which susceptible individuals become infected.

Recovery rate (gamma)

In an SIR model, the parameter $\gamma$ that represents the rate at which infected individuals recover or are removed from the infectious pool.