Abstract
Classification is one of the most common tasks in statistics and machine learning. Constructing a classifier is highly nontrivial, as the data analyst must optimize predicting performance, often while preserving interpretability, model parsimony, and scientific plausibility. Building a classifier is an iterative process, wherein the data analyst alternates between data exploration, variable selection, and model assessment. The Visual Classification Toolbox was designed to assist an analyst in performing exploratory visualizations to facilitate classifier construction and to visually compare performance of various classifiers. Allowing the analyst to visually inspect the behavior of the classifier with respect to different aspects of the data is central to this task.
Authors
Amit Meir - amitmeir (a) uw (dt) edu
Jonathan Fintzi - fintzij (a) uw (dt) edu
Summary Image
Running Instructions
- Download the repository.
- Open as an R shiny project in R-Studio.
- Click on the 'Run App' button.