107,281 Hits in 3.8 sec

Instance classification with prototype selection

Josip Krapac, Florent Perronnin, Teddy Furon, Hervé Jégou
2014 Proceedings of International Conference on Multimedia Retrieval - ICMR '14  
We propose a novel algorithm for the selection of class-specific prototypes which are used in a voting-based classification scheme.  ...  We address the problem of instance classification: our goal is to annotate images with tags corresponding to objects classes which exhibit small intra-class variations such as logos, products or landmarks  ...  Section 2 reviews the state-of-the-art in object recognition and instance-level image retrieval. Section 3 describes the proposed approach for instance classification based on prototype selection.  ... 
doi:10.1145/2578726.2578786 dblp:conf/mir/KrapacPFJ14 fatcat:b44ht67c5bhntfnskfyeqctayq

A Visual Mining Approach to Improved Multiple-Instance Learning

Sonia Castelo, Moacir Ponti, Rosane Minghim
2021 Algorithms  
With support of the MILTree layout, the initial classification model was updated by changing the training set, which is composed of the prototype instances.  ...  This problem is often addressed by selecting an instance to represent each bag, transforming an MIL problem into standard supervised learning.  ...  After selecting the training set with MILTree and using the proposed prototype selection methods, all training bags were correctly classified, which means that all instances chosen as instance prototypes  ... 
doi:10.3390/a14120344 fatcat:mfyuh7w6evbftbf7orwntcy5um

Prototype and Feature Selection by Sampling and Random Mutation Hill Climbing Algorithms [chapter]

David B. Skalak
1994 Machine Learning Proceedings 1994  
With the goal of reducing computational costs without sacrificing accuracy, we describe two algorithms to find sets of prototypes for nearest neighbor classification.  ...  Here, the term "prototypes" refers to the reference instances used in a nearest neighbor computation -the instances with respect to which similarity is assessed in order to assign a class to a new data  ...  selecting only a small handful of instances as prototypes.  ... 
doi:10.1016/b978-1-55860-335-6.50043-x dblp:conf/icml/Skalak94 fatcat:qo6wkvp3lfaihcct2ho3werqau

MIS-Boost: Multiple Instance Selection Boosting [article]

Emre Akbas, Bernard Ghanem, Narendra Ahuja
2011 arXiv   pre-print
We also apply MIS-Boost to large-scale image classification, where we show that the automatically selected prototypes map to visually meaningful image regions.  ...  In this paper, we present a new multiple instance learning (MIL) method, called MIS-Boost, which learns discriminative instance prototypes by explicit instance selection in a boosting framework.  ...  We propose the following additive model as our bag classification function: F (B) = sign M m=1 f m (B) , (1) where each f m (B), called a base classifier, is associated with a prototype instance p m ∈  ... 
arXiv:1109.2388v1 fatcat:w36wm4ealrbo3lgaidtrjsatki

Simultaneous generation of prototypes and features through genetic programming

Mauricio Garcia-Limon, Hugo Jair Escalante, Eduardo Morales, Alicia Morales-Reyes
2014 Proceedings of the 2014 conference on Genetic and evolutionary computation - GECCO '14  
This paper introduces a genetic programming approach to tackle the simultaneous generation of prototypes and features to be used for classification with a NN classifier.  ...  The proposed method learns to combine instances and attributes to produce a set of prototypes and a new feature space for each class of the classification problem via genetic programming.  ...  Mauricio García-Limón was supported by CONACyT with scholarship No. 345683.  ... 
doi:10.1145/2576768.2598356 dblp:conf/gecco/Garcia-LimonEMM14 fatcat:rzjwhq6edzd2rj2lovupuahj2q

Correcting the hub occurrence prediction bias in many dimensions

Nenad Tomasev, Krisztian Buza, Dunja Mladenic
2016 Computer Science and Information Systems  
The existing prototype selection methods implement different criteria for selecting relevant points to use in classification, which constitutes a selection bias.  ...  This study examines the nature of the instance selection bias in intrinsically high-dimensional data.  ...  We have shown that this coupling of instance selection with classification requires the selection methods to output the unbiased prototype hubness estimates recalculated on the training data after the  ... 
doi:10.2298/csis140929039t fatcat:c5veelea4bb2ho7lgw6btmzu4q

Evolutionary Adaptive Self-Generating Prototypes for imbalanced datasets

Dayvid V. R. Oliveira, George D. C. Cavalcanti, Tsang Ing Ren, Ricardo M. A. Silva
2015 2015 International Joint Conference on Neural Networks (IJCNN)  
Many PG techniques that have a high classification accuracy in regular datasets, have a poor performance when dealing with imbalanced datasets.  ...  The Adaptive Self-Generating Prototypes was proposed to tackle the SGP problem with imbalanced datasets, but, in doing so, the reduction rate is compromised.  ...  Prototype selection (PS) is a process of instance reduction that removes instances that are redundant or irrelevant to the classification, and selects the representative ones.  ... 
doi:10.1109/ijcnn.2015.7280702 dblp:conf/ijcnn/OliveiraCRS15 fatcat:i3ny3cvfpjbp5pkokxyjeb3xba

Support Vector Based Prototype Selection Method for Nearest Neighbor Rules [chapter]

Yuangui Li, Zhonghui Hu, Yunze Cai, Weidong Zhang
2005 Lecture Notes in Computer Science  
We proposed a new prototype selection method based on support vectors for nearest neighbor rules. It selects prototypes only from support vectors.  ...  During classification, for unknown example, it can be classified into the same class as the nearest neighbor in feature space among all the prototypes.  ...  They only indicated that it is feasible to select prototypes for NN with SVM and didn't compare the performance with common instance reduction method.  ... 
doi:10.1007/11539087_68 fatcat:xtvfohm6tjcetjtxesn77lldqu

Quad division prototype selection-based k-nearest neighbor classifier for click fraud detection from highly skewed user click dataset

Deepti Sisodia, Dilip Singh Sisodia
2021 Engineering Science and Technology, an International Journal  
The results show improved classification performance with QDPSKNN in terms of precision, recall, f-measure, g-mean, reduction rate and execution time, compared to existing prototype selection methods in  ...  The performance is also compared with one baseline model (k-NN) and four other prototype selection methods such as NearMiss-1, NearMiss-2, NearMiss-3, and Condensed Nearest-Neighbor.  ...  Nearest-Neighbor based prototype selection methods Existing k-Nearest-Neighbor model In this section, we will discuss the modeling of k-NN [33] , and the problems associated with imbalanced classification  ... 
doi:10.1016/j.jestch.2021.05.015 fatcat:vtrs3fijhbagtdiowlhdmqjx5q

Cluster-based instance selection for machine classification

Ireneusz Czarnowski
2011 Knowledge and Information Systems  
The paper proposes a cluster-based instance selection approach with the learning process executed by the team of agents and discusses its four variants.  ...  The basic assumption is that instance selection is carried out after the training data have been grouped into clusters.  ...  Acknowledgments This research has been supported by the Polish Ministry of Science and Higher Education with grant for years 2008-2010.  ... 
doi:10.1007/s10115-010-0375-z fatcat:y5tkp5sjqrgsnhtynmxdtf7nrq

Survey of Nearest Neighbor Condensing Techniques

2011 International Journal of Advanced Computer Science and Applications  
This drawback was dealt by the researchers' community as the problem of prototype selection. Trying to solve this problem several techniques presented as condensing techniques were proposed.  ...  For this, one possibility is to hybridize them with other algorithms, called modern heuristics or metaheuristics, which, themselves, can improve the solution.  ...  PROTOTYPE SELECTION Prototype selection is the process of finding representative patterns from the data, which can help in reducing these data.  ... 
doi:10.14569/ijacsa.2011.021110 fatcat:wsfyw6dbgffl3o7t6fxoafvfwe

Multiple-instance learning with pairwise instance similarity

Liming Yuan, Jiafeng Liu, Xianglong Tang
2014 International Journal of Applied Mathematics and Computer Science  
instance prototypes.  ...  The basic idea is selecting from every training bag a pair of the most similar instances as instance prototypes and then mapping training bags into the embedding space that is constructed from all the  ...  MILES regards all training instances as the initial instance prototypes and performs the instance selection implicitly by learning a 1-norm SVM with a linear kernel.  ... 
doi:10.2478/amcs-2014-0041 fatcat:ngb5h6a47bgqtleajhxlwloa5u

Bayesian instance selection for the nearest neighbor rule

Sylvain Ferrandiz, Marc Boullé
2010 Machine Learning  
In this paper, we focus on the second, challenging issue: instance selection.  ...  The performance of these methods relies on three crucial choices: a distance metric, a set of prototypes and a classification scheme.  ...  Instance selection algorithms usually apply a K-NN rule with a winner-takes-all voting procedure. This will be referred to as the K-NN classification scheme.  ... 
doi:10.1007/s10994-010-5170-2 fatcat:ga5qytsttnb3pfpkvozse36k2e

A Genetic Prototype Learner

Sandip Sen, Leslie Knight
1995 International Joint Conference on Artificial Intelligence  
Given a set of prototypes for each of the possible classes, the class of an input instance is determined by the prototype nearest to this instance.  ...  Supervised classification problems have received considerable attention from the machine learning community.  ...  An instance can be classified using prototypes by finding the prototype with which it shares most of its features, and then using the class of that prototype.  ... 
dblp:conf/ijcai/SenK95 fatcat:f33hqvzxpzdqxbth57glynkviq

Multiple Instance Cancer Detection by Boosting Regularised Trees [chapter]

Wenqi Li, Jianguo Zhang, Stephen J. McKenna
2015 Lecture Notes in Computer Science  
With images labelled at image-level, we first search a set of region-level prototypes by solving a submodular set cover problem.  ...  Regularised regression trees are then constructed and combined on the set of prototypes using a multiple instance boosting framework.  ...  Selecting instances as prototypes for bag classification was used previously with bags represented in terms of distances to prototypes [4, 7] .  ... 
doi:10.1007/978-3-319-24553-9_79 fatcat:ntpauqxrhjc77moyprtsf56gxm
« Previous Showing results 1 — 15 out of 107,281 results