Filters








236,629 Hits in 5.3 sec

A Reinforcement Learning Approach to Online Learning of Decision Trees [article]

Abhinav Garlapati, Aditi Raghunathan, Vaishnavh Nagarajan and Balaraman Ravindran
2015 arXiv   pre-print
Online decision tree learning algorithms typically examine all features of a new data point to update model parameters.  ...  We propose a novel alternative, Reinforcement Learning- based Decision Trees (RLDT), that uses Reinforcement Learning (RL) to actively examine a minimal number of features of a data point to classify it  ...  A challenge in searching over the space of decision trees is the combinatorial complexity of the problem.  ... 
arXiv:1507.06923v1 fatcat:sytaki5gjzazfaecjlm4nn6fzu

Solving Hybrid Markov Decision Processes [chapter]

Alberto Reyes, L. Enrique Sucar, Eduardo F. Morales, Pablo H. Ibargüengoytia
2006 Lecture Notes in Computer Science  
However, MDPs require an explicit representation of the state space and the probabilistic transition model which, in continuous or hybrid continuousdiscrete domains, are not always easy to define.  ...  This information is used to build a decision tree (C4.5) representing a small set of abstract states with equivalent rewards, and then is used to learn a probabilistic transition function using a Bayesian  ...  Acknowledgments This work was supported in part by IIE Project No. 12941 and CONACYT Project No. 47968.  ... 
doi:10.1007/11925231_22 fatcat:pfd65gbga5ayzcet7bgwd5xm3a

Optimizing Decision Trees Using Multi-objective Particle Swarm Optimization [chapter]

Jonathan E. Fieldsend
2009 Studies in Computational Intelligence  
in feature space to form their classification boundaries).  ...  Additionally feature selection is implicit within the decision tree structure.  ...  Decision trees As mentioned in the preceding section, the nodes in a decision tree act as rules, recursively partitioning the decision space.  ... 
doi:10.1007/978-3-642-03625-5_5 fatcat:kyqtlnyrzzeehot6jnlhbnrh6m

Simplest Streaming Trees [article]

Haoyin Xu, Jayanta Dey, Sambit Panda, Joshua T. Vogelstein
2022 arXiv   pre-print
We therefore developed the simplest possible extension of decision trees we could think of: given new data, simply update existing trees by continuing to grow them, and replace some old trees with new  ...  Decision forests, including random forests and gradient boosting trees, remain the leading machine learning methods for many real-world data problems, specifically on tabular data.  ...  This work is graciously funded by the Defense Advanced Research Projects Agency (DARPA) Lifelong Learning Machines program through contract FA8650-18-2-7834.  ... 
arXiv:2110.08483v4 fatcat:g6s643suuzap5dqubmx3xtp6yy

Adaptive Reinforcement Learning in Box-Pushing Robots

Jin-Ling Lin
2018 International Journal of Automation and Smart Technology  
First, a decision tree was applied to partition the state space according to temporary differences in reinforcement learning, so that a real valued action domain could be represented by a discrete space  ...  The box moves in the direction in which the robot asserts force. To push the box to the target point, the robot needs to learn how to adjust angles, avoid obstacles, and keep balance.  ...  First, the decision tree was applied to partition the continuous state space according to temporary differences of reinforcement learning, so that a real value domain can be represented by a discrete space  ... 
doi:10.5875/ausmt.v8i1.1551 fatcat:lnw2zwvaobhsfnz45dorhtd2wi

Comparative Analysis of Different Techniques for Novel Class Detection

Patel Jignasa N, Sheetal Mehta
2012 International Journal of Computer Applications Technology and Research  
Data stream mining is the process of extracting knowledge from continuous data. Data stream can be viewed as a sequence of relational touples arrives continuously at time varying.  ...  The decision tree continuously updates with new data points so that the most recent tree represents the most recent concept in data stream.  ...  IN [6] paper describe the decision tree learning algorithm The ID3 (Iterative Dichotomiser) technique builds decision tree using information theory.  ... 
doi:10.7753/ijcatr0103.1002 fatcat:bq2ndicybfalzbefo2o4qsoeo4

On the Application of a New Method of the Top-Down Decision Tree Incremental Pruning in Data Classification

Shao Hongbo, Zhou Jing, Wu Jianhui
2015 Open Automation and Control Systems Journal  
Decision tree, as an important branch of machine learning, has been successfully used in several areas.  ...  In order to overcome its defects, decision trees pruning is often adopted as a follow-up step of the decision trees learning algorithm to optimize decision trees.  ...  Seen from the perspective of hypothesis space search, a top-down decision space searching is used, which belong to the universal induction learning method.  ... 
doi:10.2174/1874444301507011922 fatcat:dhm4u5n3zjedrhiu53njd6vqjq

Towards Continuous Monitoring of Environment under Uncertainty: A Fuzzy Granular Decision Tree Approach

Preetham N. Reddy, Sahith N. Dambekodi, Tirtharaj Dash
2017 Annual India Software Engineering Conference  
In this paper, we propose to use a decision tree model that has the capability of handling uncertainty in the acquired data from the environment. e resulting model is called as Fuzzy Granular Decision  ...  Real-time decision making is another challenge in the eld on which the research community has been focusing on improving the performance of the underlying models. e underlying models are usually the learning  ...  Moreover, although a decision tree learner learns from the data by partitioning the space into multiple subspaces with conditions, it is already shown that the decision tree scales well with higher dimensional  ... 
dblp:conf/indiaSE/ReddyDD17 fatcat:ykw5hdloobfgzo62g3kniruyzq

Entity Embeddings of Categorical Variables [article]

Cheng Guo, Felix Berkhahn
2016 arXiv   pre-print
We map categorical variables in a function approximation problem into Euclidean spaces, which are the entity embeddings of the categorical variables.  ...  Entity embedding not only reduces memory usage and speeds up neural networks compared with one-hot encoding, but more importantly by mapping similar values close to each other in the embedding space it  ...  Single decision tree Decision trees partition the feature space X into M different sub-spaces R 1 , R 2 , . . . R M .  ... 
arXiv:1604.06737v1 fatcat:pvngy76mynbnljgfmilgbkjbty

Evaluating techniques for learning a feedback controller for low-cost manipulators

Oliver M. Cliff, T. Sildomar, Monteiro
2013 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems  
Learning algorithms to control robotic arms have introduced elegant solutions to the complexities faced in such systems.  ...  The agents were tested in a simulated domain for learning closed-loop policies of a simple task with no prior information.  ...  informative discussions over continuous MDP algorithms.  ... 
doi:10.1109/iros.2013.6696428 dblp:conf/iros/CliffSM13 fatcat:i3krcmnf5zfqnomketugis2rsm

Can hyperparameter tuning improve the performance of a super learner? A case study

Jenna Wong, Travis Manderson, Michal Abrahamowicz, David L Buckeridge, Robyn Tamblyn
2019 Epidemiology  
When implementing super learning, however, not tuning the hyperparameters of the algorithms in it may adversely affect the performance of the super learner.  ...  Super learning is an ensemble machine learning approach used increasingly as an alternative to classical prediction techniques.  ...  Decision tree Decision trees are non-parametric learning algorithms that apply a set of rules to partition the multi-dimensional space of covariates into hypercubes within which the outcome is more homogeneous  ... 
doi:10.1097/ede.0000000000001027 pmid:30985529 pmcid:PMC6553550 fatcat:3bjyf4ihnnam3brgnb6pkqf5ge

Analysis of Classification and Clustering based Novel Class Detection Techniques for Stream Data Mining

Kamini Tandel, Jignasa N. Patel
2015 International Journal of Engineering Research and  
Data stream is continuous and always change in nature. Data stream mining is the process of extracting knowledge form continuous data.  ...  Concept drift means data changes rapidly over time and novel class define as new class appear in continuous data stream.  ...  The decision tree classifier continuously updated so that it represents the most recent concept in the data stream.  ... 
doi:10.17577/ijertv4is100160 fatcat:cqtpjn4qxrb4pndtteg4uhxlfu

Abstraction and Refinement for Solving Continuous Markov Decision Processes

Alberto Reyes, Pablo H. Ibargüengoytia, Luis Enrique Sucar, Eduardo F. Morales
2006 European Workshop on Probabilistic Graphical Models  
In our approach, the reward function and transition model are learned from a random exploration of the environment, and can work with both, pure continuous spaces; or hybrid, with continuous and discrete  ...  We propose a novel approach for solving continuous and hybrid Markov Decision Processes (MDPs) based on two phases.  ...  Acknowledgments This work was supported in part by the Instituto de Investigaciones Eléctricas, México and CONACYT Project No. 47968.  ... 
dblp:conf/pgm/ReyesISM06 fatcat:exscliirinhc7eavwn3jvaeeli

A game strategy model in the digital curling system based on NFSP

Yuntao Han, Qibin Zhou, Fuqing Duan
2021 Complex & Intelligent Systems  
tree searching in continuous action space.  ...  AbstractThe digital curling game is a two-player zero-sum extensive game in a continuous action space.  ...  [18] proposed a Monte Carlo tree based action decision method and a Delaunay triangulation-based method respectively for the continuous action state space, both of which can select better strategy and  ... 
doi:10.1007/s40747-021-00345-6 fatcat:2i4rmut77jf37dfori5ke2m5ya

MODEL SELECTION VIA META-LEARNING: A COMPARATIVE STUDY

KALOUSIS ALEXANDROS, HILARIO MELANIE
2001 International journal on artificial intelligence tools  
The results show that decision trees and boosted decision trees models enhance the performance of the system.  ...  We compare the performance of meta-models produced by instancebased learners, decision trees and boosted decision trees.  ...  This work has been supported by Swiss OFES in the framework of ESPRIT IV LTR project METAL(26-357)  ... 
doi:10.1142/s0218213001000647 fatcat:xbjfa3csqvcu7jsxf4wccehxz4
« Previous Showing results 1 — 15 out of 236,629 results