Decision Tree Induction Based on Efficient Tree Restructuring

By Paul E. Utgoff, Neil C. Berkman, J. Clouse. 1997

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

Learning Optimal Decision Trees with SAT (2018, methodology): “One concrete example used in the paper is ITI [Utgoff et al. , 1997a] 2 .”
Top-down induction of decision trees classifiers - a survey (2005, methodology): “Utgoff [68], [69] presents several methods for updating decision trees incrementally.”
Dynamic Weighted Majority: An Ensemble Method for Drifting Concepts (2007, methodology): “DWM-ITI DWM-ITI uses as a base learner an incremental algorithm for inducing decision trees, called Incremental Tree Inducer, or ITI (Utgoff et al., 1997).”
Automatic Construction of Decision Trees from Data: A Multi-Disciplinary Survey (1998, methodology): “proposed incremental tree induction methods in the context of univariate decision trees [361, 363, 364] as well as multivariate trees [365].”
DDD: A New Ensemble Approach for Dealing with Concept Drift (2012, methodology): “[35] as the base learners for each ensemble.”
The Impact of Diversity on Online Ensemble Learning in the Presence of Concept Drift (2010, methodology): “The ensembles were composed of 25 ITI on-line decision trees [46].”
Decision trees for mining data streams (2006, unknown): “The incorporation of new information involves re-structuring the actual tree [17,29,31].”
Hybrid decision tree (2002, methodology): “Utgoff et al. (1997) designed an incremental tree induction algorithm ITI based upon a tree revision mechanism that updated the splits in the branch where the new example passed down.”
Machine learning: a review of classification and combining techniques (2006, unknown): “Neural networks are usually more able to easily provide incremental learning than decision trees (Saad 1998), even though there are some algorithms for incremental learning of decision trees such as (Utgoff et al. 1997) and (McSherry 1999).”
Minimising Decision Tree Size as Combinatorial Optimisation (2009, methodology): “In the rightmost columns of Table 2, we report the size of the pruned decision trees computed by ITI and W EKA .”
Learning the structure of Factored Markov Decision Processes in reinforcement learning problems (2006, methodology): “In spiti, we use the decision tree induction algorithm named iti (Utgoff et al., 1997) and noted ITI(Tree(F ),A, ς).”
Dynamic weighted majority: a new ensemble method for tracking concept drift (2003, methodology): “Using an incremental version of naive Bayes and Incremental Tree Inducer [25] as base algorithms, we evaluated the method using two synthetic problems involving concept drift: the STAGGER Concepts [23] and the " SEA Concepts " , a problem recently proposed in the data mining community [24].”
On Explaining Decision Trees (2020, methodology): “Concretely, we use the well-known DT training tools ITI (Incremental Tree Induction) [65,31] and IAI (Interpretable AI) [10,28].”
Integrating Guidance into Relational Reinforcement Learning (2004, background): “One consequence is that it is not feasible to store all the statistics necessary for tree restructuring as done by Utgoff, Berkman, and Clouse (1997).”
Learn$^{++}$ .NC: Combining Ensemble of Classifiers With Dynamically Weighted Consult-and-Vote for Efficient Incremental Learning of New Classes (2009, methodology): “tion of classifier weights [27], [28], or architecture update by growing/pruning of decision trees, nodes or clusters [29]–[31], or both as in STAGGER [32] and hybrid decision trees [5].”
On-line Random Forests (2009, methodology): “There exist incremental methods for single decision trees but they are either memory intensive, because every node sees and stores all the data [22], or have to discard important information if parent nodes change.”