RS-class: Class RS - An S4 class to represent a raw signal inputs

In 4301350/sm4sd: Statistical Modeling for Sensor Data

Description

The RS class contains 3 equal length vector inputs: σ_b, ε_b and ε_p.

Slots

sigb

A n-dimensional numeric vector represents the electrical conductivity of bulk soil, σ_b.

epsb

A n-dimensional numeric vector represents the electrical permitivity of bulk soil, ε_b.

epsp

A n-dimensional numeric vector represents the electrical conductivity of pore water, ε_p.

Details

Before the class is created, setValidity will be launched to check the length of 3 input vectors. It will return "Error" if the lengths are not equal.

Methods

Available methods for this class are:

• show: the class display is simplyfied, each slot is limited to 25;

• print: displays the whole class;

• get ('['): returns the value of a slot;

• plot: plots 3 slots as "timeseries" objects in the same graphic;

• plotCor: plots the correlation between ε_b and σ_b;

• buildClass: builds classes SM.dlm and SM.HL;

• buildRNN: training a RNN model in order to build clase SM.rnn.

Check the example below for more details.

Examples

n <- 50
# An example to create the RS class.
my_first_class <- RS(sigb = rnorm(n, mean = 0, sd = 1),
                     epsb = rnorm(n, mean = 0, sd = 2),
                     epsp = rnorm(n, mean = 1, sd = 1))

# Show the class
my_first_class
# print the class
print(my_first_class)
# Extract the slot "sigb"
sigb <- my_first_class["sigb"]
### o  sigb <- `[`(my_first_class, "sigb)

# plot the class as timeseries object
plot(my_first_class)
# plot the correlation
plotCor(my_first_class, 7)

# build class
## build dlm model
dlm_class <- buildClass(object = my_first_class,
                        method = "dlm", freq = 1,
                        ind = list(), lagMax = 3,
                        verify = TRUE, parallel = TRUE)
## build HL model
mp_class <- buildClass(object = my_first_class,
                       method = "HL", ind = 9.5)
### ind in this case is offset

4301350/sm4sd documentation built on June 14, 2020, 4:22 p.m.