README.md
In jmpaon/ScatterR-gui: A GUI for Scatter-R, an implementation of scatter-algorithm in R
ScatterR-gui
ScatterR-gui is a graphical user interface for the basic use of Scatter-R.
Scatter-R is an implementation of Scatter-algorithm in R.
Description
The ScatterR-gui provides a graphical user interface for Scatter-R.
Detailed description about the Scatter-R project can be found in the project wiki.
Scatter is an algorithm useful to determine if the dataset or parts of it has such information that can be successfully used for classification and class prediction. [1] It can also used in dimensionality reduction as shown by Saarikoski et al. in [2].
It works simply by traversing the dataset from a randomly chosen starting case to always closest neighbour recording the class label. In this way a collection vector is produced and the label change count is calculated from the collection vector. Then, scatter value is calculated as a proportion of the label changes v and theoretical maximum number of label changes w, thus the equation for Scatter value S = v / w. [1]
Scatter value is also used to calculate separation power, which is the difference between random situation, i.e. the labels of the current dataset is randomly distributed, and the current situation. The equation for separation power is F = z - s, where z is the scatter value for random situation.[1]
ScatterR-gui() function summons a graphical user interface
to assist in using the scatter algorithm
and the associated preprocessing function scatter.preprocess.
ScatterR-gui() provides an easy-to-use graphical interface for
reading data files,
making selections related to preprocessing
and the scatter algorithm calculation,
calling the scatter.preprocess and scatter functions
with the selected arguments
and ultimately saving the results.
Installation
See installation instructions here.
Usage
Calling the GUI
ScatterR-gui() opens the ScatterR-gui main window.
Using the GUI
In the following, the GUI elements and their functionality
is described, starting from the top-most elements.
- Read CSV datafile... button
- Opens a dialog to select a data file to read
and the selections concerning the format of the data file.
- Select file... opens the file selection dialog;
user selects a CSV or TXT file containing the data to be
processed with Scatter algorithm.
- Does the file have column headers? (Yes/No): Select "Yes" if
your data has variable/column/attribute names on the first row,
otherwise select "No".
- Select value separator: Select the character that is
used as value separator in your data file.
Comma and semicolon are common options.
- Read file: reads the file you have selected
with the header and value separator selections you have made.
- Select class variable... button
- Shows a list of the column/variable names in the data you have read
using the Read CSV datafile button.
- Select the name of the column that contains the class label information
in your data.
- When you read the datafile, the initial guess for the column containing
class variable is going to be the rightmost factor-like column.
Remember to select the correct column if your dataset's class variable column
is not the rightmost factor-like column.
- Select included classes... button
- Shows a list of the different class labels that the column you have selected
as class label containing column with the Select class variable button
- Check/Uncheck the classes you want included/excluded in/from the
scatter calculation.
- Select included variables... button
- Shows a list of the column/variable names in the data you have read
using the Read CSV datafile button.
- Check/Uncheck the columns/variables you want included
in the scatter calculation.
- Select variables to scale... button
- Shows a list of the column/variable names in the data you have read
using the Read CSV datafile button, that are numeric and
can be scaled by the preprocessing function.
- Check/Uncheck the columns/variables you want to scale during preprocessing.
- Select variables to binarize... button
- Shows a list of the column/variable names in the data you have read
using the Read CSV datafile button, that are factor-like or integer-like
and can be binarized by the preprocessing function.
- Check/Uncheck the columns/variables you want to binarize during preprocessing.
- Handling of missing values radio button set
- This option selects how missing values are handled during preprocessing.
The preprocessing function always removes the rows where classvar is missing.
- Do nothing: No action; Missing values are not estimated or removed.
- Replace with class median/mode: Replaces missing values in each column
with the median or mode (depending on the nature of the variable) of the
class of the case with the missing value. This is the recommended
missing value handling strategy.
- Replace with column median/mode: Replaces missing values in each column
with the column median/mode, depending on the nature of the data in the column.
- Remove rows: Removes the rows that have missing values.
- Select distance measure radio button set
- Select which distance measure is to be used in scatter calculation.
- Euclidean: Euclidean distance measure.
- Manhattan: Manhattan distance measure.
- HEOM: Heterogenous Euclidean-Overlap Metric.
- Iterations spinbox
- Selects the number of iterations for the Scatter calculation.
Consult the Scatter algorithm documentation for details.
- Baseline iterations spinbox
- Selects the number of baseline iterations for the Scatter calculation.
Consult the Scatter algorithm documentation for details.
- Select calculation radio button set
- Selects the calculation procedure for the Scatter algorithm.
The different options are described
in the Scatter algorithm documentation.
- Single
- Variables
- Classes
- All
- Calculate button
- Starts the Scatter calculation with the selected data and options.
Data is first preprocessed according to the preprocessing options
and the preprocessed dataframe is then passed to the Scatter algorithm.
- The results are printed in the R console
- When the calculation finishes,
scatter.gui
shows a dialog where it is possible to save the result.
- Save result to TXT file button: Opens a save dialog to save the results
in the format they are printed in the R console.
- Save result object as text representation button:
Opens a save dialog to save the results
as an R object text representation, that can be read back to
R environment with the
dget() function.
Further reading and references
[1] Juhola, M., & Siermala, M. (2012). A scatter method for data and variable importance evaluation. Integrated Computer-Aided Engineering, 19 (2), 137–149. http://doi.org/10.3233/ICA-2011-0385
[2] Saarikoski, J., Laurikkala, J., Järvelin, K., Siermala, M., & Juhola, M. (2014). Dimensionality reduction in text classification using scatter method. International Journal of Data Mining, Modelling and Management, 6 (1), 1. http://doi.org/10.1504/IJDMMM.2014.059978
jmpaon/ScatterR-gui documentation built on May 19, 2019, 1:55 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
ScatterR-gui
ScatterR-gui is a graphical user interface for the basic use of Scatter-R. Scatter-R is an implementation of Scatter-algorithm in R.
Description
The ScatterR-gui provides a graphical user interface for Scatter-R. Detailed description about the Scatter-R project can be found in the project wiki.
Scatter is an algorithm useful to determine if the dataset or parts of it has such information that can be successfully used for classification and class prediction. [1] It can also used in dimensionality reduction as shown by Saarikoski et al. in [2].
It works simply by traversing the dataset from a randomly chosen starting case to always closest neighbour recording the class label. In this way a collection vector is produced and the label change count is calculated from the collection vector. Then, scatter value is calculated as a proportion of the label changes v and theoretical maximum number of label changes w, thus the equation for Scatter value S = v / w. [1]
Scatter value is also used to calculate separation power, which is the difference between random situation, i.e. the labels of the current dataset is randomly distributed, and the current situation. The equation for separation power is F = z - s, where z is the scatter value for random situation.[1]
ScatterR-gui() function summons a graphical user interface
to assist in using the scatter algorithm
and the associated preprocessing function scatter.preprocess.
ScatterR-gui() provides an easy-to-use graphical interface for
reading data files,
making selections related to preprocessing
and the scatter algorithm calculation,
calling the scatter.preprocess and scatter functions
with the selected arguments
and ultimately saving the results.
Installation
See installation instructions here.
Usage
Calling the GUI
ScatterR-gui() opens the ScatterR-gui main window.
Using the GUI
In the following, the GUI elements and their functionality is described, starting from the top-most elements.
- Read CSV datafile... button
- Opens a dialog to select a data file to read and the selections concerning the format of the data file.
- Select file... opens the file selection dialog; user selects a CSV or TXT file containing the data to be processed with Scatter algorithm.
- Does the file have column headers? (Yes/No): Select "Yes" if your data has variable/column/attribute names on the first row, otherwise select "No".
- Select value separator: Select the character that is used as value separator in your data file. Comma and semicolon are common options.
- Read file: reads the file you have selected with the header and value separator selections you have made.
- Select class variable... button
- Shows a list of the column/variable names in the data you have read using the Read CSV datafile button.
- Select the name of the column that contains the class label information in your data.
- When you read the datafile, the initial guess for the column containing class variable is going to be the rightmost factor-like column. Remember to select the correct column if your dataset's class variable column is not the rightmost factor-like column.
- Select included classes... button
- Shows a list of the different class labels that the column you have selected as class label containing column with the Select class variable button
- Check/Uncheck the classes you want included/excluded in/from the scatter calculation.
- Select included variables... button
- Shows a list of the column/variable names in the data you have read using the Read CSV datafile button.
- Check/Uncheck the columns/variables you want included in the scatter calculation.
- Select variables to scale... button
- Shows a list of the column/variable names in the data you have read using the Read CSV datafile button, that are numeric and can be scaled by the preprocessing function.
- Check/Uncheck the columns/variables you want to scale during preprocessing.
- Select variables to binarize... button
- Shows a list of the column/variable names in the data you have read using the Read CSV datafile button, that are factor-like or integer-like and can be binarized by the preprocessing function.
- Check/Uncheck the columns/variables you want to binarize during preprocessing.
- Handling of missing values radio button set
- This option selects how missing values are handled during preprocessing. The preprocessing function always removes the rows where classvar is missing.
- Do nothing: No action; Missing values are not estimated or removed.
- Replace with class median/mode: Replaces missing values in each column with the median or mode (depending on the nature of the variable) of the class of the case with the missing value. This is the recommended missing value handling strategy.
- Replace with column median/mode: Replaces missing values in each column with the column median/mode, depending on the nature of the data in the column.
- Remove rows: Removes the rows that have missing values.
- Select distance measure radio button set
- Select which distance measure is to be used in scatter calculation.
- Euclidean: Euclidean distance measure.
- Manhattan: Manhattan distance measure.
- HEOM: Heterogenous Euclidean-Overlap Metric.
- Iterations spinbox
- Selects the number of iterations for the Scatter calculation. Consult the Scatter algorithm documentation for details.
- Baseline iterations spinbox
- Selects the number of baseline iterations for the Scatter calculation. Consult the Scatter algorithm documentation for details.
- Select calculation radio button set
- Selects the calculation procedure for the Scatter algorithm.
The different options are described
in the Scatter algorithm documentation.
- Single
- Variables
- Classes
- All
- Selects the calculation procedure for the Scatter algorithm.
The different options are described
in the Scatter algorithm documentation.
- Calculate button
- Starts the Scatter calculation with the selected data and options. Data is first preprocessed according to the preprocessing options and the preprocessed dataframe is then passed to the Scatter algorithm.
- The results are printed in the R console
- When the calculation finishes,
scatter.guishows a dialog where it is possible to save the result.- Save result to TXT file button: Opens a save dialog to save the results in the format they are printed in the R console.
- Save result object as text representation button:
Opens a save dialog to save the results
as an R object text representation, that can be read back to
R environment with the
dget()function.
Further reading and references
[1] Juhola, M., & Siermala, M. (2012). A scatter method for data and variable importance evaluation. Integrated Computer-Aided Engineering, 19 (2), 137–149. http://doi.org/10.3233/ICA-2011-0385
[2] Saarikoski, J., Laurikkala, J., Järvelin, K., Siermala, M., & Juhola, M. (2014). Dimensionality reduction in text classification using scatter method. International Journal of Data Mining, Modelling and Management, 6 (1), 1. http://doi.org/10.1504/IJDMMM.2014.059978
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.