Measurements tables

Measurements tables are CSV or XLSX files which store Measurement data in tabular format. The content of the tables can be loaded into Julia environment as a DataFrame and added to the Platform object.

Format

The structure of tables corresponds to Measurements type properties. The first row is the header. The sequence of columns may vary.

• scenario : a String value with Scenario identifier.
• t : a Float64 value equal to the time point of measured value
• measurement : a Float64 measured value
• scope (optional): a String value which states a scope of simulation. Possible values are ode_ or event identifier if the value was saved after applying the event. Default value is ode_
• prob.<id> : a set of options to characterize the probability distribution. The supported ids depend on the distribution type. For normal and lognormal distributions the available headers are:
  • prob.type (optional) : a String declaring probability type. normal is default.
  • prob.mean : Float64 value or String representing @Const or @Record id in the model. The value represents mean parameter in normal distribution.
  • prob.sigma : Float64 value or String representing @Const or @Record id in the model. The value represents sigma (standard deviation) parameter in normal distribution.

Currently two probability types are available: normal, lognormal. This distributions can be used for the relevant types of error models.

prob.type: normal

Each row in the table will be transformed into the corresponding component of log-likelihood function $-2ln(L)$.

\[\Lambda = \sum_i \left( ln(<prob.sigma>_i^2) + \frac{(<prob.mean>_i - <measurement>_i)^2}{<prob.sigma>_i^2}\right)\]

prob.type: lognormal

Each row in the table will be transformed into the corresponding component of log-likelihood function $-2ln(L)$.

\[\Lambda = \sum_i \left( ln(<prob.sigma>_i^2) + \frac{(ln(<prob.mean>_i) - ln(<measurement>_i))^2}{<prob.sigma>_i^2}\right)\]

Loading to Platform

Measurement table can be loaded into Julia environment as a DataFrame using HetaSimulator.read_measurements method. This method reads the file, checks the content and formats the data to be used inside the Platform object.

measurements = read_measurements("measurements.csv")

32×7 DataFrame
 Row │ t        measurement  scope   prob.mean  prob.sigma  scenario  prob.type 
     │ Float64  Float64      Symbol  String     Float64     Symbol     Symbol    
─────┼───────────────────────────────────────────────────────────────────────────
   1 │     2.0     8.46154   ode_    a                1.0   dataone    normal
   2 │     4.0     7.33333   ode_    a                1.2   dataone    normal
   3 │     6.0     6.47059   ode_    a                2.2   dataone    normal
  ⋮  │    ⋮          ⋮         ⋮         ⋮          ⋮           ⋮          ⋮

The data frame can be loaded into platform using the HetaSimulator.add_measurements! method.

add_measurements!(platform, measurements)

Example

Loading file measurements.csv with the following content.

Read as DataFrame object.

measurements = read_measurements("./measurements.csv")

Add all measurements to Platform

add_measurements!(platform, measurements)

As a result the Platform will contain three measurements.

These operations are equivalent to manually created Measurement objects.

# dataone = Scenario(...)

m1 = NormalMeasurementPoint(2, 8.461539334, :ode, :a, 1)
m2 = NormalMeasurementPoint(4, 7.333333812, :ode, :a, 1.2)
m3 = NormalMeasurementPoint(6, 6.470591567, :ode, :a, 2.2)

push!(dataone.measurements, m1)
push!(dataone.measurements, m2)
push!(dataone.measurements, m3)