TODO.md

In decisionpatterns/tidyimpute: Impute the Tidyverse Way

TODO

• [ ] Add impute_mode_*

• [x] Add impute_false_* and impute_true_*

• [ ] Support easybake(TM) "better recipes"

• [ ] Should impute_max / impute_min should not return -Inf and +Inf for all NA rows or be consistent with na.mean and na.median

• [x] Create Logo Circle Ban of NA

• [x] na[._]impute as alias for na[._]replace respectively.

• [x] na.explicit and na_explicit apply to factors only

• [ ] Row-based imputation does not need to calculate every-value values for all observations, only the missing ones. This is different than column-based imputations which need values from all observations. There might be some efficiency gains from doing this.

• [ ] There is a generalized imputation that uses both rows and columns and might automatically consider by-groups (how does the values of )

• [x] Is recall a part of

• na.tools::na.replace only deals with vectors and references to multivariate replacement might be misplaced.
• [x] tidyimpute::impute deals with tables and therefore is the pro
• How to implement recall?
• Recall should be able to take the same object. Identify the input. A recollection should:
  • test for the form of the input,
  • apply one or more
  • recall:
  • test for input: digest(.tbl[0,])
  • a functions with arguments modified except the input. impute( . , .na, ... )

coerce_safe

• [ ] Move coerce_safe to the coercion package. [ ] Import coercion.

• _all, _if, _at

  • na_replace_all( .tbl, .funs, ... )
  • na_replace_if( .tbl, .predicate, .funs, ... )
  • na_replace_at( .tbl, .predicate, .funs, ... )

So .tbl %>% na_replace_all( iris, 3, ...) na_replace( iris, mean, na.rm=TRUE ) na_replace_if( iris, is.cont, mean, na.rm=TRUE ) na_replace_at( iris, ! Species, mean, na.rm=TRUE )

• [ ] Support atomics with impute
• na_replace and na_explicit
• [x] replace by scalar? (Low-level)
• [x] replace by vector? (Other)
• [x] replace by unary function? ()
• [x] replace by function of multiple args: impute_*
• [ ] replace by model/formula? ()
• [ ] store .na value (if scalar)

• [ ] store replaced idxs --- like na.omit

• Implement slow functions with Rcpp

• Might there be a clever way to allow something like: NA_explicit_ <- . %>% mean(., na.rm=TRUE) This will not work

• [ ] Consider having an option for values for the na_level, e.g. options( na_explicit = "(Missing)" ) or, options( na_explicit = mean )

• [x] Categorical and continuous variables must be different.

• [ ] Explicit value might depend on the class, type (cat vs. cont), or on a specific attribute, na_explicit or na of the specific vari`able.

• [ ] Consider how explicit NA will be treated in sorting

• [ ] Do we want exceptional values first or last?

• [ ] Use catcont package?

• [x] na_replace and na_explicit are getting very similar and should probably be made aliases

• [ ] Devise syntax of related to list-like/recursive objects

• applying a function to an entire vs.
• applying to each element object See na_explicit and na_implicit

Completed

• [x] na_drop_rows, na_drop_cols for table-like objects
  • Remove rows/cols with all NAs
• [x] Make na_replace vectorized, e.g. na_replace( value=... )
• [x] na_ifelse for na_replace

decisionpatterns/tidyimpute documentation built on May 28, 2019, 4 p.m.