generate_networks: Return a JSON array of network data of a fitting model
In roqua/autovar: Automated Vector Autoregression Model Creation

generate_networks

R Documentation

Return a JSON array of network data of a fitting model

Description

This function uses repeated calls to var_main to find a fitting model for the data provided and then calls convert_to_graph to return a JSON representation of best valid model found.

Usage

generate_networks(
  data,
  timestamp,
  always_include = NULL,
  pairs = NULL,
  positive_variables = NULL,
  negative_variables = NULL,
  pick_best_of = NULL,
  incident_to_best_of = NULL,
  labels = list(),
  measurements_per_day = 3,
  max_network_size = 6,
  include_model = FALSE,
  second_significances = c(0.05, 0.01, 0.005)
)

Arguments

`data`	a data frame. Each row is a measurement. Each column is an endogenous variable.
`timestamp`	the date of the first measurement in the format `'yyyy-mm-dd'`. If multiple measurements are taken per day, the first measurement (row) in the data frame should correspond to the first measurement on that day.
`always_include`	a vector of variable names that should always be included in the network if possible. Can also be `NULL`, in which case no variables are included by default.
`pairs`	a vector of variable names in the form `c('pair1a','pair1b','pair2a','pair2b','pair3a',...)`. In other words: it is a vector of even length in which each two subsequent positions are seen as a pair. The pairs are treated such that only one pair or one variable of the pairs is included in the networks.
`positive_variables`	a vector of names of variables that measure a positive affect (e.g., happiness). Variable names not occurring in the `positive_variables` or `negative_variables` list are considered neutral.
`negative_variables`	a vector of names of variables that measure a negative affect (e.g., sadness). Variable names not occurring in the `positive_variables` or `negative_variables` list are considered neutral.
`pick_best_of`	a vector of variable names of which one should always be included in the network. Can also be `NULL` (the default value), in which case no variable is always included (unless specified by `always_include`).
`incident_to_best_of`	if `pick_best_of` is not NULL, a choice is made between the networks generated at each iteration based on which has the most incoming edges that originate from nodes listed in the vector of variable names `incident_to_best_of`.
`labels`	a list where keys are variable names and values are labels.
`measurements_per_day`	an integer in [1,16] denoting the number of measurements per day. Defaults to 3.
`max_network_size`	an integer in [2,6] denoting the number of nodes to include in the networks initially. Defaults to 6.
`include_model`	determines whether the imputed data set and coefficients of the best model should be returned in the network JSON. Defaults to `FALSE`.
`second_significances`	is the vector of significance levels to be used after the first attempt.

Value

This function returns a string representing a json array of two networks and an array of the top links.

Examples

## Not run: 
GN_COLUMNS <- c('ontspanning', 'opgewektheid', 'hier_en_nu', 'concentratie',
                'beweging', 'iets_betekenen', 'humor', 'buiten_zijn',
                'eigenwaarde', 'levenslust', 'onrust', 'somberheid',
                'lichamelijk_ongemak', 'tekortschieten', 'piekeren', 'eenzaamheid',
                'uw_eigen_factor')
d<-load_file("../data/input/DataDndN_nonimputed_voorAndo.sav",log_level=3)
d<-d$raw_data[,GN_COLUMNS]
timestamp <- '2014-03-01'
cat(generate_networks(data = d,
                  timestamp = timestamp,
                  always_include = 'uw_eigen_factor',
                  pairs = c('opgewektheid','onrust',
                            'somberheid','ontspanning',
                            'somberheid','onrust'),
                  positive_variables = c('opgewektheid','ontspanning','hier_en_nu',
                                         'concentratie', 'beweging','iets_betekenen',
                                         'humor', 'buiten_zijn','eigenwaarde', 'levenslust'),
                  negative_variables = c('onrust','somberheid','lichamelijk_ongemak',
                                         'tekortschieten','piekeren','eenzaamheid'),
                  labels = list(ontspanning = "Ontspanning",
                                opgewektheid = "Opgewektheid",
                                hier_en_nu = "In het hier en nu leven",
                                concentratie = "Concentratie",
                                beweging = "Beweging",
                                iets_betekenen = "Iets betekenen",
                                humor = "Humor",
                                buiten_zijn = "Buiten zijn",
                                eigenwaarde = "Eigenwaarde",
                                levenslust = "Levenslust",
                                onrust = "Onrust",
                                somberheid = "Somberheid",
                                lichamelijk_ongemak = "Lichamelijk ongemak",
                                tekortschieten = "Tekortschieten",
                                piekeren = "Piekeren",
                                eenzaamheid = "Eenzaamheid",
                                uw_eigen_factor = "Mijn eigen factor"),
                  measurements_per_day = 3,
                  max_network_size = 6))

## End(Not run)

roqua/autovar documentation built on Jan. 21, 2023, 7:37 p.m.