Description Usage Arguments Details Value References Examples
Fit seed viability/survival curve to estimate the seed lot constant (\mjseqnK_i) and the period to lose unit probit viability (\mjseqn\sigma). \loadmathjax
1 2 3 4 5 6 7 8 9 |
data |
A data frame with the seed viability data recorded periodically. It should possess columns with data on
|
viability.percent |
The name of the column in |
samp.size |
The name of the column in |
storage.period |
The name of the column in |
generalised.model |
logical. If |
use.cv |
logical. If |
control.viability |
The control viability (%). |
This function fits seed survival data to the following seed viability equation \insertCiteellis_improved_1980viabilitymetrics which models the relationship between probit percentage viability and time period of storage.
\mjsdeqnv=K_i-\fracp\sigma
or
\mjsdeqnv=K_i-\left ( \frac1\sigma \right )\cdot p
Where, \mjseqnv is the probit percentage viability at storage time \mjseqnp (final viability), \mjseqnK_i is the probit percentage viability of the seedlot at the beginning of storage (seedlot constant) and \mjseqn\frac1\sigma is the slope.
The above equation may be expressed as a generalized linear model (GLM) with a probit (cumulative normal distribution) link function as follows \insertCitehay_modelling_2014viabilitymetrics.
\mjsdeqny = \phi(v) = \phi\left ( K_i-\left ( \frac1\sigma \right )p \right )
Where, \mjseqny is the proportion of seeds viabile after time period \mjseqnp and the link function is \mjseqn\phi^-1, the inverse of the cumulative normal distribution function.
The parameters estimated are the intercept \mjseqnK_i, theoretical viability of the seeds at the start of storage or the seed lot constant, and the slope \mjseqn-\sigma^-1, where \mjseqn\sigma is the standard deviation of the normal distribution of seed deaths in time or the period of time to lose unit probit viability.
This function can also incorporate a control viability parameter into the model to fit the modified model suggested by \insertCitemead_prediction_1999viabilitymetrics. The modified model is as follows.
\mjsdeqny = C_v \times \phi(v) = C_v \times \phi\left (K_i-\left ( \frac1\sigma \right )p \right )
Where, \mjseqnC_v is the control viability parameter which is the proportion of respondent seeds. This excludes the bias due to seeds of the ageing population that have already lost viability at the start of storage and those non-respondent seeds that are not part of the ageing population due to several reasons.
A list of class FitSigma
with the following components:
data |
A data frame with the data used for computing the model. |
model |
The fitted model as an object of class |
parameters |
A data.frame of parameter estimates, standard errors and p value. |
fit |
A one-row data frame with estimates of model fitness such as log likelyhoods, Akaike Information Criterion, Bayesian Information Criterion, deviance and residual degrees of freedom. |
Ki |
The estimated seed lot constant from the model. |
sigma |
The estimated period of time to lose unit probit viability from the model. |
message |
Warning or error messages generated during fitting of model, if any. |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 | data(seedsurvival)
df <- seedsurvival[seedsurvival$crop == "Soybean" &
seedsurvival$moistruecontent == 7 &
seedsurvival$temperature == 25,
c("storageperiod", "rep",
"viabilitypercent", "sampsize")]
plot(df$storageperiod, df$viabilitypercent)
#----------------------------------------------------------------------------
# Generalised linear model with probit link function (without cv)
#----------------------------------------------------------------------------
model1a <- FitSigma(data = df, viability.percent = "viabilitypercent",
samp.size = "sampsize", storage.period = "storageperiod",
generalised.model = TRUE)
model1a
# Raw model
model1a$model
# Model parameters
model1a$parameters
# Model fit
model1a$fit
#----------------------------------------------------------------------------
# Generalised linear model with probit link function (with cv)
#----------------------------------------------------------------------------
model1b <- FitSigma(data = df, viability.percent = "viabilitypercent",
samp.size = "sampsize", storage.period = "storageperiod",
generalised.model = TRUE,
use.cv = TRUE, control.viability = 98)
model1b
# Raw model
model1b$model
# Model parameters
model1b$parameters
# Model fit
model1b$fit
#----------------------------------------------------------------------------
# Linear model after probit transformation (without cv)
#----------------------------------------------------------------------------
model2a <- FitSigma(data = df, viability.percent = "viabilitypercent",
samp.size = "sampsize", storage.period = "storageperiod",
generalised.model = FALSE)
model2a
# Raw model
model2a$model
# Model parameters
model2a$parameters
# Model fit
model2a$fit
#----------------------------------------------------------------------------
# Linear model after probit transformation (with cv)
#----------------------------------------------------------------------------
model2b <- FitSigma(data = df, viability.percent = "viabilitypercent",
samp.size = "sampsize", storage.period = "storageperiod",
generalised.model = FALSE,
use.cv = TRUE, control.viability = 98)
model2b
# Raw model
model2b$model
# Model parameters
model2b$parameters
# Model fit
model2b$fit
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.