This tool uses classifcation algorithms to predict penguin species (Adelie, Chinstrap, Gentoo) based on bill length (mm) and bill (mm) depth. Models include multinomial regression, k-nearest neighbors, random forest, xgboost (default hyperparameters settings were used). Accuracy of prediction for the four models on the test set were as follows:
The script utilizes the argparser package in R for parsing command-line arguments. The following examples show the incantation for running the Rscript with the following arguments: model, bill length and bill depth.
Ex 1: Rscript penguin_sh_tool.R model=mn_mod_fit bill_length=39 bill_depth=18
Ex 2: Rscript penguin_sh_tool.R knn_mod_fit 50 18
Ex 3: Rscript penguin_sh_tool.R rf_mod_fit 50 18
Ex 4. Rscript penguin_sh_tool.R xg_mod_fit 50 14
The output of Ex 3 should include the following messages:
And return the predicted species based on the model and inputs.
