snip: Run entire SNIP deduplication algorithm
In bayesmendel/snipR: Pedigree deduplication

Description Usage Arguments Value

View source: R/snip.R

Run entire SNIP deduplication algorithm

snip(
  pedigrees,
  requestID,
  isProband,
  keyVars,
  keyVars.male = NULL,
  keyVars.female = NULL,
  keyWt = NULL,
  blockVar = NULL,
  repSN = 1,
  windowSN = 10,
  keyLength = length(keyVars),
  method = "intersection",
  thresh = 1:7,
  priority,
  dateFormat = NULL,
  printRuntime = TRUE,
  seed = NULL
)

`pedigrees`	Pedigree data to deduplicate.
`requestID`	Column that has the ID for the family.
`isProband`	Column that indicates the proband.
`keyVars`	Character vector of column names for the variables in the sort key
`keyVars.male`	Optional character vector of column names for the variables in the sort key that are specific to males
`keyVars.female`	Optional character vector of column names for the variables in the sort key that are specific to females
`keyWt`	Numeric vector of weights assigned to variables in the sort key, corresponding to `keyVars`. If `NULL`, the standard deviations of the variables in the data will be used as weights.
`blockVar`	Vector of column names for the blocking variables, where families in different blocks will not be considered when searching for duplicates.
`repSN`	Number of iterations when sorting neighbors according to the sort key
`windowSN`	Integer representing the size of the sliding window to use during sorted neighbors.
`keyLength`	Numeric vector representing the number of key variables (out of `c(keyVar.bin, keyVar.cont)`) to concatenate per sort key. If missing, the key lengths will be randomly generated. The length of the vector should be `repSN`.
`method`	If "intersection", we use the intersection score. If "greedy", we use the greedy match score. If "both", we use both.
`thresh`	Vector of thresholds. If `method = "intersection"`, then a pair is considered neighbors if the intersection match score is greater than the threshold. If `method = "greedy"`, then the threshold is treated as a percentile, and a pair is considered neighbors if the greedy match score is greater than the percentile. If `method = "both"`, then the user should provide a list, such as `list(intersection = 1:7, threshold = c(0.8, 0.9))`.
`priority`	A list of structure (var = 'Varx', min = TRUE) with 'Varx' being a character value corresponding to a column in rawData. This parameter determines how to sort the duplicates. If `min = TRUE`, then we use the minimum value of 'Varx' for each duplicate entity. Otherwise, we use the maximum value.
`dateFormat`	Character string of the format of the date. This is only used if the priority variable is a date. The format should match the formats of class `POSIXlt` used in the `base::strptime` function.
`printRuntime`	If TRUE, will print the runtime
`seed`	Seed

An object of class Duplicates containing the duplicate entities and representatives for each duplicate entity (including singletons without duplicates).

bayesmendel/snipR documentation built on Jan. 25, 2022, 12:33 a.m.