Type: Package
Title: Integrating Morphological Modeling and Machine Learning for Decision Support
Version: 0.1.0
Description: Integrating morphological modeling with machine learning to support structured decision-making (e.g., in management and consulting). The package enumerates a morphospace of feasible configurations and uses random forests to estimate class probabilities over that space, bridging deductive model exploration with empirical validation. It includes utilities for factorizing inputs, model training, morphospace construction, and an interactive 'shiny' app for scenario exploration.
License: MIT + file LICENSE
URL: https://github.com/theogrost/MLmorph
BugReports: https://github.com/theogrost/MLmorph/issues
Encoding: UTF-8
Depends: R (≥ 4.3.0)
Imports: caret (≥ 6.0.94), jsonlite (≥ 1.8.8), magrittr, openxlsx (≥ 4.2.5.2), randomForest (≥ 4.7.1.1), shiny (≥ 1.10.0), stats (≥ 4.3.0), tidyr (≥ 1.3.1), utils (≥ 4.3.0)
RoxygenNote: 7.3.2
Suggests: testthat (≥ 3.0.0)
Config/testthat/edition: 3
Language: en-US
NeedsCompilation: no
Packaged: 2025-08-27 19:06:23 UTC; theog
Author: Oskar Kosch ORCID iD [aut, cre, cph]
Maintainer: Oskar Kosch <contact@oskarkosch.com>
Repository: CRAN
Date/Publication: 2025-09-02 06:20:02 UTC

MLmorph: Integrating Morphological Modeling and Machine Learning for Decision Support

Description

Integrating morphological modeling with machine learning to support structured decision-making (e.g., in management and consulting). The package enumerates a morphospace of feasible configurations and uses random forests to estimate class probabilities over that space, bridging deductive model exploration with empirical validation. It includes utilities for factorizing inputs, model training, morphospace construction, and an interactive 'shiny' app for scenario exploration.

Author(s)

Maintainer: Oskar Kosch contact@oskarkosch.com (ORCID) [copyright holder]

See Also

Useful links:

Launch the MLmorph shiny app

Description

Launch the MLmorph shiny app

Usage

MLmorph(
  host = "127.0.0.1",
  port = NULL,
  launch.browser = TRUE,
  maxUploadSize = 200 * 1024^2
)

Arguments

host

Host interface to bind (default "127.0.0.1").

port

Integer port or NULL to auto-select.

launch.browser

Logical; open in a browser. Default TRUE.

maxUploadSize

Maximum request size in bytes; sets options(shiny.maxRequestSize = ...). Default 200 * 1024^2.

Value

The value returned by runApp.

See Also

runApp

Examples

if(interactive()){
MLmorph()
}

Launch MLmorph from the source tree (development helper)

Description

Launch MLmorph from the source tree (development helper)

Usage

MLmorph_live()

Create a morphospace of predictor combinations with class probabilities

Description

Create a morphospace of predictor combinations with class probabilities

Usage

create_morphospace(the_data, model, shiny = FALSE)

Arguments

the_data

A data.frame used to derive unique values of predictors.

model

A classification model fitted via a formula interface that supports predict(model, newdata, type = "prob") (e.g., from create_rf_model).

shiny

Logical; if TRUE, compute predictions in batches with shiny progress. Default FALSE.

Value

A list with components:

Examples

n  <- 60
y  <- factor(sample(letters[1:3], n, TRUE))
x1 <- factorize_numeric_vector(runif(n, 10, 20), breaks_no = 3)
x2 <- factorize_numeric_vector(runif(n,  1,  2), breaks_no = 3)
df <- data.frame(y, x1, x2)
fit <- create_rf_model(df, dependent = "y", ntree = 50)$model
ms  <- create_morphospace(df, fit)
names(ms)

Create a random forest classification model

Description

Create a random forest classification model

Usage

create_rf_model(
  data,
  dependent = colnames(data)[ncol(data)],
  independent = setdiff(colnames(data), dependent),
  train_validate_split = 0.8,
  shiny = FALSE,
  ...
)

Arguments

data

A data.frame containing predictors and the outcome.

dependent

Character scalar; the name of the outcome (must be a factor for classification). Defaults to the last column of data.

independent

Character vector; names of predictor variables. Defaults to all columns except dependent.

train_validate_split

Numeric in (0, 1); proportion of rows used for training. Default is 0.8.

shiny

Logical; if TRUE, trains incrementally and reports progress via MLmorph app. Default FALSE.

...

Additional arguments passed to randomForest (e.g., ntree).

Value

A named list with components:

Examples

n <- 60
y  <- factor(sample(letters[1:3], n, TRUE))
x1 <- factorize_numeric_vector(runif(n, 10, 20), breaks_no = 3)
x2 <- factorize_numeric_vector(runif(n,  1,  2), breaks_no = 5)
df <- data.frame(y, x1, x2)
fit <- create_rf_model(df, dependent = "y", ntree = 50)
names(fit)

Turn binary vector into a factor

Description

Turn binary vector into a factor

Usage

factorize_binary_vector(data_vector, custom_labels = NULL)

Arguments

data_vector

Logical vector.

custom_labels

Optional length-2 character vector: first for TRUE, second for FALSE.

Value

A factor with two levels in TRUE, FALSE order.

Examples

factorize_binary_vector(c(TRUE, FALSE, TRUE))

Turn character vector into a factor

Description

Turn character vector into a factor

Usage

factorize_character_vector(data_vector, custom_labels = NULL)

Arguments

data_vector

Character vector.

custom_labels

Optional named character vector where names are original values and values are labels.

Value

A factor with labeled levels.

Examples

factorize_character_vector(c("A First", "B Second", "C Third"))

Identity factorization for numbered strings

Description

Identity factorization for numbered strings

Usage

factorize_identity(data_vector)

Arguments

data_vector

Character vector where values are already labeled (e.g., "1. A").

Value

A factor with levels == labels.

Examples

factorize_identity(c("1. First", "2. Second", "3. Third"))

Heuristic factorization for all columns of a data frame

Description

Heuristic factorization for all columns of a data frame

Usage

factorize_nicely_dataframe(data_frame)

Arguments

data_frame

A data frame.

Value

A data frame with all columns converted to factors.

Examples

df <- data.frame(x = runif(20), y = rep(c(TRUE, FALSE, TRUE, TRUE), 5))
factorize_nicely_dataframe(df)

Heuristic factorization for a single vector

Description

Heuristic factorization for a single vector

Usage

factorize_nicely_vector(data_vector)

Arguments

data_vector

A vector (numeric, logical, or character).

Value

A factor (ordered for numeric inputs with many distinct values).

Examples

factorize_nicely_vector(c("a", "b", "a"))

Turn numeric vector into an ordered factor

Description

Turn numeric vector into an ordered factor

Usage

factorize_numeric_vector(
  data_vector,
  method = c("equal_bins", "equal_distance", "custom_breaks"),
  breaks_no = 5,
  custom_breaks = NULL,
  custom_labels = NULL
)

Arguments

data_vector

Numeric vector.

method

Factorization rule: one of "equal_bins", "equal_distance", "custom_breaks".

breaks_no

Integer \ge 2; number of intervals when method != "custom_breaks".

custom_breaks

Optional numeric vector of cut points (strictly increasing) used when method = "custom_breaks".

custom_labels

Optional character vector of labels. If supplied, its length should equal length(custom_breaks) - 1.

Value

An ordered factor with interval labels.

Examples

factorize_numeric_vector(runif(10))

Zero-padded ordinal labels

Description

Zero-padded ordinal labels

Usage

get_label_numbers(vec)

Arguments

vec

A vector; its length determines the padding width.

Value

A character vector of zero-padded ordinals (e.g., "01", "02", …).

Load tabular data (xlsx, csv, or json)

Description

Load tabular data (xlsx, csv, or json)

Usage

load_data(data_path)

Arguments

data_path

Character scalar; path to a .xlsx, .csv, or .json file.

Value

A base data.frame with the imported data.

Examples

tmp_csv <- tempfile(fileext = ".csv")
utils::write.csv(data.frame(a = 1:2, b = c("x", "y")), tmp_csv, row.names = FALSE)
load_data(tmp_csv)

tmp_json <- tempfile(fileext = ".json")
jsonlite::write_json(list(a = 1:2, b = c("x","y")), tmp_json, auto_unbox = TRUE)
load_data(tmp_json)

tmp_xlsx <- tempfile(fileext = ".xlsx")
openxlsx::write.xlsx(data.frame(a = 1:2, b = c("x","y")), tmp_xlsx)
load_data(tmp_xlsx)