clusterConcepts: clusterConcepts

In jreps/patientCluster: Performs patient clustering for patient cohorts using OHDSI CDM v5

Description

Create topics by clusting condition_concept_ids based on ingredience counts

Usage

clusterConcepts(dbconnection, cdmDatabaseSchema = NULL, method = "kmeans",
  clusterSize = 10, topicSize = NULL, scale = T,
  covariatesToInclude = NULL, indications = T, dayStart = 1,
  dayEnd = 30, use_min_obs = TRUE, min_obs = 100,
  extraparameters = NULL, updateProgress = NULL, ...)

Arguments

dbconnection	using DatabaseConnector - connect to cdm database
cdmDatabaseSchema	- cdm schema used to extract data from
method	class:character - method used to do clustering (currently only supports kmeans)
clusterSize	class:numeric - number of clusters returned,
topicSize	class:numeric - number of topics in glrm
scale	class:boolean - whether to use ingredience percentage scale for clustering
covariatesToInclude	class:character vector - features to include: default NULL
indications	class:boolean extract drug indicator features;Default TRUE
dayStart	class:integer number of days relative to condition_concept_code to start looking for drugs
dayEnd	class:integer number of days relative to condition_concept_code to stop looking for drugs
use_min_obs	class:boolean whether to remove ingredient features that are rare
min_obs	clkass:integer threshold used when use_min_obs is TRUE to determine what is rare

Value

list contining definition data.frame containing columes for concept_id, covariate (cluster id)

Examples

clusterConcepts()

jreps/patientCluster documentation built on May 20, 2019, 10:46 a.m.