| Type: | Package |
| Title: | Feature Selection in Highly Correlated Spaces |
| Version: | 0.1.1 |
| Description: | Feature selection algorithm that extracts features in highly correlated spaces. The extracted features are meant to be fed into simple explainable models such as linear or logistic regressions. The package is useful in the field of explainable modelling as a way to understand variable behavior. |
| License: | MIT + file LICENSE |
| URL: | https://allen-1242.github.io/TangledFeatures/ |
| Depends: | R (≥ 2.10) |
| Imports: | correlation, data.table, dplyr, fastDummies, ggplot2, igraph, janitor, Matrix, methods, purrr, ranger |
| Suggests: | knitr, R.rsp, rmarkdown, testthat (≥ 3.0.0) |
| VignetteBuilder: | knitr |
| Config/testthat/edition: | 3 |
| Encoding: | UTF-8 |
| LazyData: | true |
| RoxygenNote: | 7.2.3 |
| NeedsCompilation: | no |
| Packaged: | 2023-02-12 17:57:05 UTC; sunny |
| Author: | Allen Sunny [aut, cre] |
| Maintainer: | Allen Sunny <allensunny1242@gmail.com> |
| Repository: | CRAN |
| Date/Publication: | 2023-02-14 09:10:02 UTC |
Advertisement dataset
Description
Advertisement dataset
Automatic Data Cleaning
Description
Automatic Data Cleaning
Usage
DataCleaning(Data, Y_var)
Arguments
Data |
The imported Data Frame |
Y_var |
The X variable |
Value
The cleaned data.
Examples
DataCleaning(Data = TangledFeatures::Housing_Prices_dataset, Y_var = 'SalePrice')
Generalized Correlation function
Description
Generalized Correlation function
Usage
GeneralCor(df, cor1 = "pearson", cor2 = "polychoric", cor3 = "spearman")
Arguments
df |
The imported Data Frame |
cor1 |
The correlation metric between two continuous features. Defaults to pearson |
cor2 |
The correlation metric between one categorical feature and one cont feature. Defaults to biserial |
cor3 |
The correlation metric between two categorical features. Defaults to Cramers-V |
Value
Returns a correlation matrix containing the correlation values between the features
Examples
GeneralCor(df = TangledFeatures::Advertisement)
Housing prices dataset
Description
Housing prices dataset
The main TangledFeatures function
Description
The main TangledFeatures function
Usage
TangledFeatures(
Data,
Y_var,
Focus_variables = list(),
corr_cutoff = 0.7,
RF_coverage = 0.95,
plot = FALSE,
fast_calculation = FALSE,
cor1 = "pearson",
cor2 = "polychoric",
cor3 = "spearman"
)
Arguments
Data |
The imported Data Frame |
Y_var |
The dependent variable |
Focus_variables |
The list of variables that you wish to give a certain bias to in the correlation matrix |
corr_cutoff |
The correlation cutoff variable. Defaults to 0.8 |
RF_coverage |
The Random Forest coverage of explainable. Defaults to 95 percent |
plot |
Return if plotting is to be done. Binary True or False |
fast_calculation |
Returns variable list without many Random Forest iterations by simply picking a variable from a correlated group |
cor1 |
The correlation metric between two continuous features. Defaults to pearson correlation |
cor2 |
The correlation metric between one categorical feature and one continuous feature. Defaults to bi serial correlation correlation |
cor3 |
The correlation metric between two categorical features. Defaults to Cramer's V. |
Value
Returns a list of variables that are ready for future modelling, along with other metrics
Examples
TangledFeatures(Data = TangledFeatures::Advertisement, Y_var = 'Sales')