Supervised and Unsupervised Discretization of Continuous Features

By James Dougherty, Ron Kohavi, Mehran Sahami. 1995

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model (2015, neutral): “For data with continuous features, a variety of procedures exist for antecedent mining [Fayyad and Irani (1993), Dougherty, Kohavi and Sahami (1995), Srikant and Agrawal (1996)].”
Mathematical optimization in classification and regression trees (2021, methodology): “We refer the reader to Liu et al. (2002) for a comprehensive review on enhancing classification and regression tree methods through the discretization of continuous predictor variables (Dougherty et al. 1995).”
Deep Neural Decision Trees (2018, methodology): “The core module we implement here is a soft binning function (Dougherty et al., 1995) that we will use to make the split decisions in DNDT.”
Clustering intrusion detection alarms to support root cause analysis (2003, background): “Dynamic generalization hierarchies, which are constructed at run-time to fit the actual data distribution, are a pragmatic way to mitigate this shortcoming [Dougherty et al. 1995; Han and Fu 1994; Lu 1997].”
Data Mining Algorithms to Classify Students (2008, methodology): “In addition, we have applied the equal-width method [13] to all the other attributes with three intervals and labels (LOW, MEDIUM and HIGH).”
Discretization: An Enabling Technique (2002, background): “As reported in a study (Dougherty et al., 1995), discretization makes learning more accurate and faster.”
A Survey of Discretization Techniques: Taxonomy and Empirical Analysis in Supervised Learning (2013, background): “Finally, we present the taxonomy.”
Web usage mining for predicting final marks of students that use Moodle courses (2013, methodology): “To all the other attributes, the equal-width method has been applied [83]; here three intervals have been used (each one depending on the range of each attribute) and the same three labels (LOW, MEDIUM, and HIGH).”
A network intrusion detection system based on a Hidden Naïve Bayes multiclass classifier (2012, background): “…methods in the last two decades to determine how continuous values should be grouped, how cut points should be positioned on the continuous scale and how many intervals should be used to generate datasets (Dougherty, Kohavi, & Sahami, 1995; Fayyad & Irani, 1993; Liu et al., 2002; Yang, 2002).”
On Embeddings for Numerical Features in Tabular Deep Learning (2022, unknown): “We point to the work by Dougherty et al. [11], which performs an overview of some classic approaches to binning and can serve as an entry point to the relevant literature on the topic.”
Activity Recognition Using Inertial Sensing for Healthcare, Wellbeing and Sports Applications: A Survey (2010, methodology): “Dougherty et al.[81] employ the Naïve Bayes for activity recognition and they investigate the effect of discretization on Naïve Bayes classiers.”
Information-Theoretic Feature Selection in Microarray Data Using Variable Complementarity (2008, methodology): “Each continuous variable is quantized according to two different methods: equal sized interval and equal frequency interval (see [15]).”
Bayesian Network Classifiers (1997, methodology): “Instead, we applied a pre-discretization step in the manner described by Dougherty et al. (1995). This pre-discretization is based on a variant of Fayyad and Irani’s (1993) discretization method.”