data_preprocessing: Data preprocessing

In pat-s/2019-feature-selection: Research Compendium package for the publication "Monitoring forest health using hyperspectral imagery: Does feature selection improve the performance of machine-learning techniques?"

Description

clean_single_plots():

• Removes columns with NA values and cols containing "ID"

• Removes the sf geometry column.

Extracts coordinates from a list of data.frames containing X and Y coordinates

Standardizes all variables of a list of data.frames in parallel

Mutates the "defoliation" variables so that no absolut zeros occur. These might cause problems when standardizing variables.

Performs a log transformation on the response variable.

Performs a boxcox transformation on the response variable.

Splits data into feature sets.

Usage

clean_single_plots(data, cols_to_drop, remove_coords = FALSE)

extract_coords(data)

standardize(data, cores)

mutate_defol(data)

log_response(data, response)

boxcox_response(data, response)

split_into_feature_sets(data, feature_set)

Arguments

data	(data.frame) data.frame.
cols_to_drop	(character(()) Columns to remove from the data
remove_coords	(logical(1)) Whether to remove coordinates from the data (i.e. columns named x and y) and their sf related geom list column
cores	(integer) Number of cores
response	(character) Name of response.
feature_set	(character) Name of feature set.

Details

Standardization applies from 2:length(names(data)), so it is expected that the response is sorted first in the data.

Value

List of data.frames with X/Y information

List of data.frames with standardized variables

data.frame

data.frame

data.frame

data.frame

pat-s/2019-feature-selection documentation built on Dec. 24, 2021, 8:40 a.m.