Probabilistic Instance-Based Learning

机译：基于概率实例的学习

获取原文

获取原文并翻译 | 示例

页面导航

摘要
著录项
相似文献
相关主题

摘要

Traditional instance-based learning methods base their predictions directly on (training) data that has been stored in the memory. The predictions are based on weighting the contributions of the individual stored instances by a distance function implementing a domain-dependent similarity metrics. This basic approach suffers from three drawbacks: computationally expensive prediction when the database grows large, overfitting in the presence of noisy data, and sensitivity to the selection of a proper distance function. We address all these issues by giving a probabilistic interpretation to instance-based learning, where the goal is to approximate predictive distributions of the attributes of interest. In this probabilistic view the instances are not individual data items but probability distributions, and we perform Bayesian inference with a mixture of such prototype distributions. We demonstrate the feasibility of the method empirically for a wide variety of public domain classification data sets.

机译：传统的基于实例的学习方法的预测直接基于（训练）已存储在内存中的数据。该预测基于通过实现依赖于域的相似性度量的距离函数对各个存储实例的贡献进行加权。这种基本方法具有三个缺点：数据库变大时计算量大的预测，在有噪声数据的情况下过拟合，以及对选择适当距离函数的敏感性。我们通过对基于实例的学习进行概率解释来解决所有这些问题，目标是近似估计感兴趣属性的预测分布。在这种概率视图中，实例不是单个数据项，而是概率分布，并且我们使用这些原型分布的混合来执行贝叶斯推断。我们通过经验证明了该方法对于各种公共领域分类数据集的可行性。

著录项

来源
《Machine learning》|1996年|507-515|共9页
会议地点 Bari(IT);Bari(IT)
作者
Henry Tirri; Petri Kontkanen; Petri Myllymaeki;
展开▼
作者单位

Complex Systems Computation Group (CoSCo) P.O.Box 26, Department of Computer Science FIN-00014 University of Helsinki, Finland;

Complex Systems Computation Group (CoSCo) P.O.Box 26, Department of Computer Science FIN-00014 University of Helsinki, Finland;

Complex Systems Computation Group (CoSCo) P.O.Box 26, Department of Computer Science FIN-00014 University of Helsinki, Finland;

展开▼
会议组织
原文格式 PDF
正文语种 eng
中图分类计算机的应用;
关键词

相似文献

外文文献
中文文献
专利

1. Making Instance-based Learning Theory usable and understandable: The Instance-based Learning Tool [J] . Varun Dutt, Cleotilde Gonzalez Computers in Human Behavior . 2012,第4期

机译：使基于实例的学习理论变得可用和易于理解：基于实例的学习工具
2. A probabilistic framework for the design of instance-based supervised ranking algorithms in an ordinal setting [J] . S. Lievens, B. De Baets, K. Cao-Van Annals of Operations Research . 2008,第octa期

机译：在序数环境中设计基于实例的监督排序算法的概率框架
3. A Relational Learning Algorithm Combining Relational Tri-training and Relational Instance-based Learning [J] . Yanjuan Li, Maozu Guo Journal of information and computational science . 2012,第2期

机译：结合关系三训练与关系实例学习的关系学习算法
4. Making Instance-based Learning Theory Usable, Transparent, and Understandable: Instance-based Learning Tool [C] . Varun Dutt, Cleotilde Gonzalez Human Factors and Ergonomics Society annual meeting . 2011

机译：使基于实例的学习理论变得可用，透明和可理解：基于实例的学习工具
5. Learning Latent Hierarchical Structures via Probabilistic Models and Deep Learning [D] . Arabshahi, Forough 2018

机译：通过概率模型和深度学习来学习潜在的层次结构
6. Nicotine improves probabilistic reward learning in wildtype but not alpha7 nAChR null mutants yet alpha7 nAChR agonists do not improve probabilistic learning. [O] . Morgane Milienne-Petiot, Kerin K Higa, Andrea Grim, -1

机译：尼古丁可改善野生型的概率奖励学习但不能改善alpha7 nAChR无效突变体但是alpha7 nAChR激动剂不能改善概率学习。
7. Probabilistic Wind Power Forecasting Approach via Instance-Based Transfer Learning Embedded Gradient Boosting Decision Trees [O] . Long Cai, Jie Gu, Jinghuan Ma, 2019

机译：基于实例的转移学习嵌入式梯度升压决策树的概率风力预测方法
8. General Instance-Based Learning Framework for Studying Intuitive Decision-Making in a Cognitive Architecture (Open Access, Publisher's Version). [R] . Thomson, R., Lebiere, C., Anderson, J. R., 2014

机译：基于通用实例的学习框架，用于研究认知架构中的直观决策（开放存取，出版商版）。

Probabilistic Instance-Based Learning

摘要

著录项

相似文献

相关主题

期刊订阅