An Empirical Comparison of Exact Nearest Neighbour Algorithms

Kibriya, Ashraf M.; Frank, Eibe

doi:10.1007/978-3-540-74976-9_16

Cited by 53 publications

(39 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For instance, when building a neighbourhood graph, the classical algorithm for kd-trees computes this graph so efficiently that quantizing the classical algorithm working only on the distances offers no significant advantage if the dimension d is small (Kibriya and Frank 2007). A fundamental open question is the study of the lower bounds for different clustering scenarios, be they classical or quantum mechanical.…”

Section: Fair Comparison Between Classical and Quantum Learning Algormentioning

confidence: 99%

Quantum speed-up for unsupervised learning

2012

View full text Add to dashboard Cite

We show how the quantum paradigm can be used to speed up unsupervised learning algorithms. More precisely, we explain how it is possible to accelerate learning algorithms by quantizing some of their subroutines. Quantization refers to the process that partially or totally converts a classical algorithm to its quantum counterpart in order to improve performance. In particular, we give quantized versions of clustering via minimum spanning tree, divisive clustering and k-medians that are faster than their classical analogues. We also describe a distributed version of k-medians that allows the participants to save on the global communication cost of the protocol compared to the classical version. Finally, we design quantum algorithms for the construction of a neighbourhood graph, outlier detection as well as smart initialization of the cluster centres.

show abstract

Section: Fair Comparison Between Classical and Quantum Learning Algormentioning

confidence: 99%

Quantum speed-up for unsupervised learning

2012

View full text Add to dashboard Cite

show abstract

“…Such a choice will yield low Bayes' error during classification. Furthermore, higher dimension (d) of the feature space (d ≥ 20) affects the optimal choice of k and hence the accuracy of the classifier [17]. It has been also observed that higher values of k make the kNN algorithm more robust to the outliers and also generate smoother class boundaries.…”

Section: Introductionmentioning

confidence: 93%

A probabilistic framework for dynamic k estimation in kNN classifiers with certainty factor

Bhattacharya

Ghosh

Chowdhury

2015

2015 Eighth International Conference on Advances in Pattern Recognition (ICAPR)

View full text Add to dashboard Cite

Accuracy of the well-known k-nearest neighbor (kNN) classifier heavily depends on the choice of k. The problem of estimating a suitable k for any test point becomes difficult due to several factors like the local distribution of training points around that test point, presence of outliers in the dataset, and, dimensionality of the feature space. In this paper, we propose a dynamic k estimation algorithm based on the neighbor density function of a test point and class variance as well as certainty factor information of the training points. Performance of the kNN algorithm with the proposed choice of k is evaluated on UCI benchmark datasets without employing any dimensionality reduction and/or explicit learning method(s). Experimental results clearly demonstrate the supremacy of the proposed method over the existing kNN algorithm with k = 1, 3, 5 and [√ ] (N denotes the number of training points) for different distance measures)

show abstract

“…We use Euclidean distance measuring as distance function as the following equation (10) [11]. n D(x, y) = L Jx; -Y;, (10) i=O Where D (x, y) is the distance function, x is the query sample, y is the sample from the training set, n is feature dimension.…”

Section: ) K-nearest Neighbor (Knn)mentioning

confidence: 99%

Hajj human event classification system using machine learning techniques

Zawbaa

Emary

Hassanien

et al. 2013

13th International Conference on Hybrid Intelligent Systems (HIS 2013)

View full text Add to dashboard Cite

In this paper, we proposed the new system for Hajj event classification in diverse and realistic Hajj videos and image scenes is investigated based on machine learning techniques. This challenging but important subject has mostly been ignored in the past due to several problems one of which is the lack of realistic and annotated video datasets. The main contribution of this work is to address the limitation and investigate the use of video for automatic annotation of human event classification. The proposed system consist of three main phases. Firstly, preprocessing phase which apply shot boundary detection algorithm for Hajj videos. After that feature extraction phase applying sparse coding based on Scale Invariant Feature Transform (SIFT) features. Finally, the event classification phase by applying several machine learning techniques including the K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Random Forests (RF) classifiers. Experiments with real data sets revealed the significant per formance advantage of the machine learning techniques over the scale invariant feature transform (SIFT) features selection method. Receiver operating characteristics (ROC) analysis is used to compare classifier performance.

show abstract

An Empirical Comparison of Exact Nearest Neighbour Algorithms

Cited by 53 publications

References 11 publications

Quantum speed-up for unsupervised learning

Quantum speed-up for unsupervised learning

A probabilistic framework for dynamic k estimation in kNN classifiers with certainty factor

Hajj human event classification system using machine learning techniques

Contact Info

Product

Resources

About