Assocs: Rules Mining Tool

Overview

Assocs is a rules mining tool designed to mine associations from transaction data. It runs on the R interface and provides a Rules Mining window and a Graphing window. The rules mining and graphing functionality is based on the methodologies outlined in Hahsler et al. (2006) and Hahsler et al. (2016).

Installation Methods

Standard Shiny Installation

To run the application using Shiny, execute the following code in your R environment:

if (!require("pacman")) install.packages("pacman")
pacman::p_load("Rgraphviz", "shiny", "tidyverse", "arules", "arulesViz", "colorspace", "visNetwork", "plotly", "igraph", "shinythemes", "readxl", "DT")
runGitHub(
  'Assocs',
  username = 'statisticsguru1',
  ref = "main"
)

Local Installation via RInno

A local installation package is available through RInno. Follow these steps for the local installation:

1. Download the installer:

   • Download the app.rar file from this link and save it to a specific location.
   • Extract the folder to your preferred location.

2. Launch the installer:

   • Navigate to the folder RInno_installer and launch the installer.
   • Continue the installation with default settings.
   • Specify the location where you want to install the app.
   • Ignore any errors that appear by clicking "OK."

3. Finish the installation:

   • The app will install and launch its first session.

Using Assocs

Launching the Application

After installation, you can launch the app by clicking the app icon found among your Windows applications.

Application Tabs

The application consists of two main tabs:

• Tab A: Rules
  • This tab is for loading data, mining rules, and downloading the mined rules.
• Tab B: Rules
  • This tab is for graphing the mined rules and downloading the graphs.

Mining Rules

To mine rules, follow these steps:

1. Data Upload:

   • Click "Browse" to select a data file (CSV or Excel). The default file type is CSV.
   • Specify if the first row contains variable names (default is TRUE).
   • Specify the file type (default is CSV).

2. Rules Input:

   • Set the minimum support value (default is 0).
   • Set the minimum confidence value (default is 0).
   • Set the minimum length of item sets (default is 2).
   • Set the maximum length of item sets (default is 2).
   • Optionally exclude zero counts (default is selected).
   • Click "Mine rules" to start mining the rules.

3. Download Rules:

   • The mined rules will be displayed, and you can download them by clicking the "Download" button.

Graphing Rules

After extracting the rules, you can graph them using the Graphing window:

1. Static Graphs:

   • Select the graph type from the dropdown.
   • Specify the measure of interestingness (default is lift).
   • Specify the shading measure (default is support).
   • Click "Go" to generate the graph.
   • Download the graph by clicking the "Download graph" button.

2. Interactive Graphs:

   • The inputs for interactive graphs are similar to static graphs.
   • The graph can be downloaded directly from the graph interface.

References

• Hahsler M. & Karpienko R. (2016). Visualizing association rules in hierarchical groups. Journal of Business Economics 87(3) 317–335. doi: 10.1007/s11573-016-0822-8
• Hahsler M. (2017). arulesViz: Interactive Visualization of Association Rules with R. The R Journal 9(2) 163. doi: 10.32614/rj-2017-047

Appendix

How to Save CSV from Excel

1. Click "Save As".
2. Select "CSV (Comma delimited)" as the file type.
3. Click "Save".