ass.tile: Assign Tiles to Train and Test
In Ehsan-F/R-Mixtape: Mixtape: Enhancing the model making process

Description Usage Arguments Details

By using the model output (Propensity or otherwise), this function will create n tiles on train and test. The function creates the tiles on train, applies the to the whole dataset, and then creates new train and test datasets. This ensures consistency across all of the datasets.

1	ass.tile(df = ds, df_train = ds1, df_test = ds2, prob = "p_", tile_name = "p_Tile", model = NULL, model_var = "X1", assign_data = "all", n = 10, TrainIndex = "TrainIndex")

`df`	The entire data.
`df_train`	Train data.
`df_test`	Test data.
`prob`	The model output variable; this could be the Propensity variable and is used to rank order and create the tiles.
`tile_name`	The desired name for the tile variable; "p_Tile" is a good option.
`model`	The name of the model in focus. If not NULL, then the prob argument becomes irrelevant and predictions are made directly through this function.
`model_var`	If model prediction is of type "prob", then the name of the required predicted variable. By default this is "X1" which indicates the probability of the event.
`assign_data`	The type of assignment that will be taking place. By default this is set to "all". "all" assigns the tiles to the entire data, train, and test. "main_only" assigns the tiles to the entire data only (this does not create new train and test datasets). "none" only outputs a vector of tiles for the entire data. Note: this option does not assign tiles to any datasets.
`n`	The number of desired tiles.
`TrainIndex`	The name of the variable indicating which observations within the entire data belong to the `train` data. By default this is set to "TrainIndex".