eflow.model_analysis.outlier_analysis

Functions

check_if_directory_exists(directory_path)	Checks if the given directory path exists. Raises an error if doesn’t
convert_to_filename(filename[, …])	Attempts to make the filename string valid.
correct_directory_path(directory_path)	Attempts to convert the directory path to a proper one by removing
create_dir_structure(directory_path, …)	Creates required directory structures inside the parent
create_unique_directory(directory_path, …)	Creates a unique folder in the proper directory structure.
dict_to_json_file(dict_obj, directory_path, …)	Writes a dict to a json file.
get_all_directories_from_path(directory_path)	Gets directories names with the provided path.
get_all_files_from_path(directory_path[, …])	Gets all filenames with the provided path.
get_unique_directory_path(directory_path, …)	Iterate through directory structure until a unique folder name can be
json_file_to_dict(filepath)	Returns back the dictionary from of a json file.
load_pickle_object(file_path)
pickle_object_to_file(obj, directory_path, …)	Writes the object to a pickle file.
write_object_text_to_file(obj, …[, …])	Writes the object’s string representation to a text file.
zcore_remove_outliers(df, feature_name, …)	Any zscore that is between the negative and positive of the ‘zscore_val’ will be return as a pandas series object.

Classes

ModelAnalysis(dataset_name[, …])	All objects in model_analysis folder of eflow are related to this object.
OutlierAnalysis(dataset_name, model, …[, …])	Analyzes a classification model’s result’s based on the prediction function(s) passed to it.

class OutlierAnalysis(dataset_name, model, model_name, feature_order, df_features, project_sub_dir='Outlier Analysis', overwrite_full_path=None, save_models=True, notebook_mode=False)

Analyzes a classification model’s result’s based on the prediction function(s) passed to it. Creates graphs and tables to be saved in directory structure.