FitSigma: Seed viability curve fitting to estimate \mjseqnK_i and...
In aravind-j/viabilitymetrics: Seed Viability Calculations and Curve Fitting
Description
Usage
Arguments
Details
Value
References
Examples
Description
Fit seed viability/survival curve to estimate the seed lot constant
(\mjseqnK_i) and the period to lose unit probit viability
(\mjseqn\sigma). \loadmathjax
Usage
1
2
3
4
5
6
7
8
9
Arguments
data
A data frame with the seed viability data recorded periodically.
It should possess columns with data on
Viability percentage
(to be indicated by the argument viability.percent),
Sample
size (to be indicated by the argument samp.size) and
Storage
period (to be indicated by the argument storage.period).
viability.percent
The name of the column in data with the
viability percentages as a character string.
samp.size
The name of the column in data with the sample size
used for calculating viability percentages as a character string.
storage.period
The name of the column in data with the time
periods at which the viabilty percentages was recorded as a character
string.
generalised.model
logical. If TRUE (recommended), the seed
viability curve is fitted as a generalised linear model with a probit link
function. If TRUE, it is fitted as a linear model with probit
transformed viability percentages.
use.cv
logical. If TRUE, then the percentage value specified in
the control.viabilty argument is incorporated as the control
viability parameter into the seed viability equation for fitting. Default
is FALSE.
control.viability
The control viability (%).
Details
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.
Value
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 glm (if
generalised.model = TRUE) or lm (if generalised.model =
FALSE).
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.
References
\insertAllCited
Examples
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
aravind-j/viabilitymetrics documentation built on May 15, 2021, 9:10 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Description Usage Arguments Details Value References Examples
Description
Fit seed viability/survival curve to estimate the seed lot constant (\mjseqnK_i) and the period to lose unit probit viability (\mjseqn\sigma). \loadmathjax
Usage
1 2 3 4 5 6 7 8 9 |
Arguments
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 (%). |
Details
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.
Value
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. |
References
\insertAllCitedExamples
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
|
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.