Search

In the new "gig" economy, a user plays the role of a consumer as well as a service provider. As a service provider, drivers travelling from a source to a destination may opportunistically pickup and drop-off packages along the way if that does not add significantly to their trip distance or time. This gives rise to a new business offering called Package Delivery as a Service (PDaaS) that brokers package pickups and deliveries at one end and connects them to drivers on the other end, thus creating an ecosystem of supply and demand. The dramatic cost savings of such a service model come from the fact that the driver is already en-route to their destination and the package delivery adds a small overhead to an already pre-planned trip. From a technical perspective, this problem introduces new technical challenges that are uncommon in the literature. The driver may want to optimise for distance or time. Furthermore, new packages arrive for delivery all the time and are assigned to various drivers continuously. This means that the algorithm has to work in an environment that is dynamic, thereby precluding most standard road network precomputation efforts. Furthermore, the number of packages that are available for delivery could be in the hundreds of thousands, which has to be quickly pruned down for the algorithm to scale. The paper proposes a variation called dual Dijkstra's that combines a forward and a backward scan in order to find delivery options that satisfy the constraints specified by the driver. The new dual heuristic improves the standard single Dijkstra's approach by narrowing down the search space, thus resulting in significant speed-ups over the standard algorithms. Furthermore, a combination of dual Dijkstra's with a heuristic landmark approach results in a dramatic speed-up compared to the baseline algorithms. Experimental results show that the proposed approach can offer drivers a choice of packages to deliver under specified constraints of time or distance, and providing sub-second response time despite the complexity of the problem involved. As the number of packages in the system increases, the matchmaking process becomes easier resulting in faster response times. The scalability of the PDaaS infrastructure is demonstrated using extensive experimental results. [ABSTRACT FROM AUTHOR]

The paper aimed to examine the price development of gold, silver, and platinum from 1 January 2019 to 27 November 2023. The aim was also to predict the price development of pairs of metals - gold and silver, gold and platinum, and silver and platinum until the end of 2024. The Nearest Neighbours method in Wolfram Mathematica software was used to achieve the set goal. In terms of the development of metal pairs until the end of 2024, it was found that the most profitable pairs for investments are gold-platinum and silver-platinum. If an investor is to invest in a single metal, platinum is recommended. In the case of a pair of metals, it would be appropriate to choose gold-platinum or silver-platinum. The research is limited in terms of the metals selected, a narrower set of historical data to make a prediction, or the selected method. [ABSTRACT FROM AUTHOR]

We study the |$k$| -nearest neighbour classifier (⁠|$k$| -NN) of probability measures under the Wasserstein distance. We show that the |$k$| -NN classifier is not universally consistent on the space of measures supported in |$(0,1)$|⁠. As any Euclidean ball contains a copy of |$(0,1)$|⁠ , one should not expect to obtain universal consistency without some restriction on the base metric space, or the Wasserstein space itself. To this end, via the notion of |$\sigma $| -finite metric dimension, we show that the |$k$| -NN classifier is universally consistent on spaces of discrete measures (and more generally, |$\sigma $| -finite uniformly discrete measures) with rational mass. In addition, by studying the geodesic structures of the Wasserstein spaces for |$p=1$| and |$p=2$|⁠ , we show that the |$k$| -NN classifier is universally consistent on spaces of measures supported on a finite set, the space of Gaussian measures and spaces of measures with finite wavelet series densities. [ABSTRACT FROM AUTHOR]

Entropy of multivariate distributions may be estimated based on the distances of nearest neighbours from each sample from a statistical ensemble. This technique has been applied on biomolecular systems for estimating both conformational and translational/rotational entropy. The degrees of freedom which mostly define conformational entropy are torsion angles with their periodicity. In this work, tree structures and algorithms to quickly generate lists of nearest neighbours for periodic and non-periodic data are reviewed and applied to biomolecular conformations as described by torsion angles. The effect of dimensionality, number of samples, and number of neighbours on the computational time is assessed. The main conclusion is that using proper data structures and algorithms can greatly reduce the complexity of nearest neighbours lists generation, which is the bottleneck step in nearest neighbours entropy estimation.

Graphs are versatile structures for the representation of many real-world data. Deep Learning on graphs is currently able to solve a wide range of problems with excellent results. However, both the generation of graphs and the handling of large graphs still remain open challenges. This work aims to introduce techniques for generating large graphs and test the approach on a complex problem such as the calculation of dose distribution in oncological radiotherapy applications. To this end, we introduced a pooling technique (ReNN-Pool) capable of sampling nodes that are spatially uniform without computational requirements in both model training and inference. By construction, the ReNN-Pool also allows the definition of a symmetric un-pooling operation to recover the original dimensionality of the graphs. We also present a Variational AutoEncoder (VAE) for generating graphs, based on the defined pooling and un-pooling operations, which employs convolutional graph layers in both encoding and decoding phases. The performance of the model was tested on both the realistic use case of a cylindrical graph dataset for a radiotherapy application and the standard benchmark dataset sprite. Compared to other graph pooling techniques, ReNN-Pool proved to improve both performance and computational requirements. [ABSTRACT FROM AUTHOR]

Entropy of multivariate distributions may be estimated based on the distances of nearest neighbours from each sample from a statistical ensemble. This technique has been applied on biomolecular systems for estimating both conformational and translational/rotational entropy. The degrees of freedom which mostly define conformational entropy are torsion angles with their periodicity. In this work, tree structures and algorithms to quickly generate lists of nearest neighbours for periodic and non-periodic data are reviewed and applied to biomolecular conformations as described by torsion angles. The effect of dimensionality, number of samples, and number of neighbours on the computational time is assessed. The main conclusion is that using proper data structures and algorithms can greatly reduce the complexity of nearest neighbours lists generation, which is the bottleneck step in nearest neighbours entropy estimation. [ABSTRACT FROM AUTHOR]

Clustering on the data with multiple aspects, such as multi-view or multi-type relational data, has become popular in recent years due to their wide applicability. The approach using manifold learning with the Non-negative Matrix Factorization (NMF) framework, that learns the accurate low-rank representation of the multi-dimensional data, has shown effectiveness. We propose to include the inter-manifold in the NMF framework, utilizing the distance information of data points of different data types (or views) to learn the diverse manifold for data clustering. Empirical analysis reveals that the proposed method can find partial representations of various interrelated types and select useful features during clustering. Results on several datasets demonstrate that the proposed method outperforms the state-of-the-art multi-aspect data clustering methods in both accuracy and efficiency. [ABSTRACT FROM AUTHOR]

Graphs are versatile structures for the representation of many real-world data. Deep Learning on graphs is currently able to solve a wide range of problems with excellent results. However, both the generation of graphs and the handling of large graphs still remain open challenges. This work aims to introduce techniques for generating large graphs and test the approach on a complex problem such as the calculation of dose distribution in oncological radiotherapy applications. To this end, we introduced a pooling technique (ReNN-Pool) capable of sampling nodes that are spatially uniform without computational requirements in both model training and inference. By construction, the ReNN-Pool also allows the definition of a symmetric un-pooling operation to recover the original dimensionality of the graphs. We also present a Variational AutoEncoder (VAE) for generating graphs, based on the defined pooling and un-pooling operations, which employs convolutional graph layers in both encoding and decoding phases. The performance of the model was tested on both the realistic use case of a cylindrical graph dataset for a radiotherapy application and the standard benchmark dataset sprite. Compared to other graph pooling techniques, ReNN-Pool proved to improve both performance and computational requirements.

The structure of B-DNA, the physiological form of the DNA molecule, has been a central topic in biology, chemistry and physics. Far from uniform and rigid, the double helix was revealed as a flexible and structurally polymorphic molecule. Conformational changes that lead to local and global changes in the helix geometry are mediated by a complex choreography of base and backbone rearrangements affecting the ability of the B-DNA to recognize ligands and consequently on its functionality. In this sense, the knowledge obtained from the sequence-dependent structural properties of B-DNA has always been thought crucial to rationalize how ligands and, most notably, proteins recognize B-DNA and modulate its activity, i.e. the structural basis of gene regulation. Honouring the anniversary of the first high-resolution X-ray structure of a B-DNA molecule, in this contribution, we present the most important discoveries of the last 40 years on the sequence-dependent structural and dynamical properties of B-DNA, from the early beginnings to the current frontiers in the field. [ABSTRACT FROM AUTHOR]

In this paper the problem of measuring factor importance on patient length of stay in an emergency department is discussed. Historical dataset contains average patient length of stay per day. Factors are agreed with domain expert. The task is to provide factors' impact measure on specific day that does not belong to the historical dataset (new observation) and average length of stay for that day is higher than specified threshold. Observations are represented as multidimensional numeric vectors. Each dimension represents factor. The basic idea consists of identifying appropriate neighbourhood and measure distances between the new observation and its neighbourhood in the historical dataset with respect to each factor. Impact measure of a factor is derived from the Error Sum of Squares. Factor impact is proportional to distance between the observation and its neighbourhood with respect to the dimension representing that factor. Nearest neighbour and clustering methods for neighbourhood determination are considered. [ABSTRACT FROM AUTHOR]

Advances in data acquisition have generated an enormous amount of data that captures business, commercial, technological and scientific information. However, some occurrences are rare or unusual, irrespective of a large amount of data available. These rare occurrences in data mining are usually referred to as outliers or anomalies. All these rare occurrences are infrequent. Sometimes it varies from 0.01% to 10% depending on the type of application. In recent years, outlier detection has become important in many applications and has attracted considerable attention among the increasing number of data mining techniques. Focusing on this has resulted in several outlier detection algorithms, mostly based on distance or density. However, each method has its inherent weaknesses. Methods based on distance have problems with local density, and methods based on density have problems with low-density patterns. In this paper, we present a new outlier detection algorithm based on the relevant attribute analysis (ODRA) for local outlier detection in a high-dimensional dataset. There are two phases of the proposed algorithm. During the preliminary stage, we present a data reduction method that reduces the data set by pruning irrelevant attributes and data points. In the second phase, we propose an outlier detection method based on k-NN kernel density estimation. The experimental results on 15 UCI machine learning repository datasets show the supremacy and effectiveness of our proposed approach over state-of-the-art outlier detection methods. [ABSTRACT FROM AUTHOR]

This paper addresses reliable and efficient calculation of the mode of a multivariate sample, which is a classical fusion function. In particular, we focus on the inputs given on the unit simplex, when aggregating elements of Atanassov intuitionistic fuzzy sets, interval‐valued fuzzy sets and their extensions, as well as compositional data. We outline the use of a specially designed 2‐additive fuzzy measures and the Choquet integral for the purposes of reducing computational complexity in higher dimensions. We present computational analysis and benchmark four different methods of density‐based mode estimation. [ABSTRACT FROM AUTHOR]

This paper aims to develop a metric indexing method that uses users' queries for reducing the search cost of similarity search systems and for avoiding the insertion cost in dynamic data sets. We have proposed an indexing method which is able to improve its structure based on users' queries. The proposed method, called I-Clusters, is a metric clustering based method, extended from the List of Clusters method. This method decreases the construction costs, and it improves the search cost after the execution of queries. The I-Clusters method allows solving the trade-off between the construction cost and the searching cost, and it also allows indexing dynamic datasets without additional cost of objects insertion. The experiment results show that the I-Clusters method significantly reduces the search cost based on queries execution, and the search performance of the proposed method can reach that of List of Clusters. [ABSTRACT FROM AUTHOR]

Associated with the establishment of removal of water sources in savanna ecosystems is the issue of the effects of such management actions on animal movement and habitat selection, longer term implications on population levels, and impacts of such change on habitat degradation and soil erosion. Extant metrics used to describe the spatial distribution of water sources on the landscape often fall short of providing source-specific information, making them hard to apply in small-scale management settings. Using the Klaserie Private Nature Reserve (KPNR) as a case study, we compare a buffer framework, describing distances to water, a nearest neighbour framework, a spatial location-allocation framework (SLAF) created in a geographic information system (GIS). These three frameworks can be combined into one GIS to demonstrate site-specific information on water source distribution, in addition to system-wide descriptions. The visually accessible quality of a GIS allows qualitative input from managers and property owners to achieve quantifiable management goals. The duality of database and visual representation provides a useful tool to assess the role of individual water sources and can easily be updated to reflect changes in their distribution.

Associated with the establishment or removal of water sources in savanna ecosystems is the issue of the effects of such management actions on animal movement and habitat selection, longer term implications on population levels, and impacts of such change on habitat degradation and soil erosion. Extant metrics used to describe the spatial distribution of water sources on the landscape often fall short of providing source-specific information, making them hard to apply in small-scale management settings. Using the Klaserie Private Nature Reserve (KPNR) as a case study, we compare a buffer framework, describing distances to water, a nearest neighbour framework, and a spatial location-allocation framework (SLAF) created in a geographic information system (GIS). These three frameworks can be combined into one GIS to demonstrate site-specific information on water source distribution, in addition to system-wide descriptions. The visually accessible quality of a GIS allows qualitative input from managers and property owners to achieve quantifiable management goals. The duality of database and visual representation provides a useful tool to assess the role of individual water sources and can easily be updated to reflect changes in their distribution.

Fault detection and location are important and front-end tasks in assuring the reliability of power electronic circuits. In essence, both tasks can be considered as the classification problem. This paper presents a fast fault classification method for power electronic circuits by using the support vector machine (SVM) as a classifier and the wavelet transform as a feature extraction technique. Using one-against-rest SVM and one-against-one SVM are two general approaches to fault classification in power electronic circuits. However, these methods have a high computational complexity, therefore in this design we employ a directed acyclic graph (DAG) SVM to implement the fault classification. The DAG SVM is close to the one-against-one SVM regarding its classification performance, but it is much faster. Moreover, in the presented approach, the DAG SVM is improved by introducing the method of Knearest neighbours to reduce some computations, so that the classification time can be further reduced. A rectifier and an inverter are demonstrated to prove effectiveness of the presented design.

This paper examines the prediction of disturbances based on their past measurements using k -nearest neighbours. The aim is to provide a prediction of a measured disturbance to a controller, in order to improve the feed-forward action. This prediction method works in an unsupervised way, it is robust against changes of the characteristics of the disturbance, and its functioning is simple and transparent. The method is tested on data from industrial process plants and compared with predictions from an autoregressive model. A qualitative as well as a quantitative method for analysing the predictability of the time series is provided. As an example, the method is implemented in an MPC framework to control a simple benchmark model. [ABSTRACT FROM AUTHOR]

Aims: The aim of this study is to estimate monthly precipitation by support vector regression and the nearest neighbourhood methods using meteorological variables data of Chabahar station. Methods and Results: Monthly precipitation was modelled by using two support vector regression and the nearest neighbourhood methods based on the two proposed input combinations. Conclusions: The results showed that the support vector regression method using normalized polynomial kernel function has higher accuracy and it has lower estimation error than the nearest neighbour method. Significance and Impact of the Study: Precipitation is one of the most important parts of the water cycle and plays an important role in assessing the climatic characteristics of each region. Modelling of monthly precipitation values for a variety of purposes, such as flood and sediment control, runoff, sediment, irrigation planning, and river basin management, is very important. The modelling of precipitation in each region requires the existence of accurately measured historical data such as humidity, temperature, wind speed, etc. Limitations such as insufficient knowledge of precipitation on spatial and temporal scales as well as the complexity of the relationship between precipitation-related climatic parameters make it impossible to estimate precipitation using conventional inaccurate and unreliable methods. [ABSTRACT FROM AUTHOR]

Millions of smart meters that are able to collect individual load curves, that is, electricity consumption time series, of residential and business customers at fine scale time grids are now deployed by electricity companies all around the world. It may be complex and costly to transmit and exploit such a large quantity of information, therefore it can be relevant to use survey sampling techniques to estimate mean load curves of specific groups of customers. Data collection, like every mass process, may undergo technical problems at every point of the metering and collection chain resulting in missing values. We consider imputation approaches (linear interpolation, kernel smoothing, nearest neighbours, principal analysis by conditional estimation) that take advantage of the specificities of the data, that is to say the strong relation between the consumption at different instants of time. The performances of these techniques are compared on a real example of Irish electricity load curves under various scenarios of missing data. A general variance approximation of total estimators is also given which encompasses nearest neighbours, kernel smoothers imputation and linear imputation methods. The Canadian Journal of Statistics 47: 65–89; 2019 © 2018 Statistical Society of Canada [ABSTRACT FROM AUTHOR]

Abstract A novel data driven approach to batch process monitoring is presented, which combines the k-Nearest Neighbour rule with the dynamic time warping (DTW) distance. This online method (DTW-NN) calculates the DTW distance between an ongoing batch, and each batch in a reference database of batches produced under normal operating conditions (NOC). The sum of the k smallest DTW distances is monitored. If a fault occurs in the ongoing batch, then this distance increases and an alarm is generated. The monitoring statistic is easy to interpret, being a direct measure of similarity of the ongoing batch to its nearest NOC predecessors and the method makes no distributional assumptions regarding normal operating conditions. DTW-NN is applied to four extensive datasets from simulated batch production of penicillin, and tested on a wide variety of fault types, magnitudes and onset times. Performance of DTW-NN is contrasted with a benchmark multiway PCA approach, and DTW-NN is shown to perform particularly well when there is clustering of batches under NOC. [ABSTRACT FROM AUTHOR]

A biannual survey of physico-chemical quality indices of 104 irrigation-water wells located in a cultivated plain of a Mediterranean island catchment was conducted using a multi-parameter probe. The campaign was planned so as to differentiate between the dry and wet seasons. The acquired data constituted the test bed for evaluating the results and the features of four spatial interpolation methods, i.e. ordinary kriging, universal kriging, inverse distance weighted and nearest neighbours, against those of the recently introduced bilinear surface smoothing (BSS). In several cases, BSS outperformed the other interpolation methods, especially during the two-fold cross-validation procedure. The study emphasizes the fact that both in situ measurements and good mathematical techniques for studying the spatial distribution of water quality indices are pivotal to agricultural practice management. In the specific case studied, the spatio-temporal variability of water quality parameters and the need for monitoring were evident, as low irrigation water quality was encountered throughout the study area. [ABSTRACT FROM AUTHOR]

Abstract: Von Willebrand disease (VWD) is one of the main inherited coagulation disorders. It is caused by a deficiency and/or a dysfunction of the von Willebrand factor (VWF), a fundamental multimeric glycoprotein involved in the hemostasis process. Correct detection of the disease is not an easy task because the disease manifests itself in many variants and a high intra‐subject variability is observed. For these reasons, the diagnostic clinical trials typically rely on a 24‐h sampling protocol, which makes the overall test long, stressful, and costly. Using a new pharmacokinetic model derived from Galvanin et al.'s 2014 study, this study aims at i) assessing the theoretical possibility to perform a shorter clinical test and ii) proposing a set of model‐based diagnostic methods as a support for the clinical team. A preliminary information analysis is performed in order to understand which sampling instants are more informative for model identification. This allowed us to propose a novel, 8‐h diagnostic protocol, which is still able to ensure model identifiability. Three alternative diagnostic methods are then proposed based on this short‐length clinical protocol. One of them directly uses the pharmacokinetic model, whereas the other two are based on the use of three indices (two pharmacokinetic indices, namely clearance, total VWF released, and as third index the basal multimer ratio) to formulate the diagnosis problem as a classification one. The classification problem is then solved using K‐nearest neighbours and linear discriminant analysis. Results show the theoretical feasibility of a VWD diagnosis based on a shorter protocol. [ABSTRACT FROM AUTHOR]

Fault detection and location are important and front-end tasks in assuring the reliability of power electronic circuits. In essence, both tasks can be considered as the classification problem. This paper presents a fast fault classification method for power electronic circuits by using the support vector machine (SVM) as a classifier and the wavelet transform as a feature extraction technique. Using one-against-rest SVM and one-against-one SVM are two general approaches to fault classification in power electronic circuits. However, these methods have a high computational complexity, therefore in this design we employ a directed acyclic graph (DAG) SVM to implement the fault classification. The DAG SVM is close to the one-against-one SVM regarding its classification performance, but it is much faster. Moreover, in the presented approach, the DAG SVM is improved by introducing the method of K-nearest neighbours to reduce some computations, so that the classification time can be further reduced. A rectifier and an inverter are demonstrated to prove effectiveness of the presented design. [ABSTRACT FROM AUTHOR]

Accurate property valuation is important for property purchasers, investors and for mortgage-providers to assess credit risk in the mortgage market. Automated valuation models (AVM) are being developed to provide cheap, objective valuations that allow dynamic updating of property values over the term of a mortgage. A useful feature of automated valuations is to provide a region of plausible price estimates for each individual property, rather than just a single point estimate. This would allow buyers and sellers to understand uncertainty on pricing individual properties and mortgage providers to include conservatism in their credit risk assessment. In this study, Conformal Predictors (CP) are used to provide such region predictions, whilst strictly controlling for predictive accuracy. We show how an AVM can be constructed using a CP, based on an underlying k-nearest neighbours approach. Time trend in property prices is dealt with by assuming a systematic effect over time and adjusting prices in the training data accordingly. The AVM is tested on a large data set of London property prices. Region predictions are shown to be reliable and the efficiency, ie region width, of property price predictions is investigated. In particular, a regression model is constructed to model the uncertainty in price prediction linked to property characteristics. [ABSTRACT FROM AUTHOR]

Copyright of Forestry Chronicle is the property of Canadian Institute of Forestry and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The k-nearest neighbour ( kNN) problem appears in many different fields of computer science, such as computer animation and robotics. In crowd simulation, kNN queries are typically used by a collision-avoidance method to prevent unnecessary computations. Many different methods for finding these neighbours exist, but it is unclear which will work best in crowd simulations, an application which is characterised by low dimensionality and frequent change of the data points. We therefore compare several data structures for performing kNN queries. We find that the nanoflann implementation of a k-d tree offers the best performance by far on many different scenarios, processing 100,000 agents in about 35 ms on a fast consumer PC. [ABSTRACT FROM AUTHOR]

• A new approach, GLOBAL-LOCAL, was developed to predict SOC at the local scale from vis–NIR data. • GLOBAL-LOCAL requires a representative set from the samples to be predicted. • GLOBAL-LOCAL improved SOC prediction compared to the entire spectral library. • GLOBAL-LOCAL was more accurate than using site-specific reference samples. • GLOBAL-LOCAL outperformed three other published approaches. There is much interest in the use of large visible near infrared (vis–NIR) soil spectral libraries for rapid soil analysis and to further soils research. However, the use of such large and diverse spectral libraries often provides biased predictions at local scales. Here, we present a new approach, GLOBAL-LOCAL, which aims to use large libraries to develop spectroscopic models that can fit locally. This is achieved by selecting the most common spectral neighbours of a small representative set (Lab) from the target site using a combination of principal component analysis (PCA) and k -nearest neighbours, followed by selection of the calibration with the lowest root mean square error (RMSE) in predicting the Lab , which is then added to the calibration to derive the GLOBAL-LOCAL. We evaluated this approach by predicting the soil organic carbon (SOC) content in two different target sites, one in Canada (Lévis, QC; N = 111) and the other in Finland (Maaninka; N = 101). Partial least squares regression (PLSR) was used for model development and validation. The soil spectral library used was a combination of that developed by the World Agroforestry Center and International Soil Reference Information Centre (N = 3875), and that developed at the Quebec Research and Development Centre (N = 1051). Based on a preliminary analysis, we selected 25 Lab samples from each target site using conditioned Latin hypercube sampling (cLHS). Prediction accuracy by GLOBAL-LOCAL for Lévis and Maaninka (RMSE = 1.41 and 3.08 g kg−1, RPIQ = 5.9 and 4.1, respectively) were better than using the global calibration (RMSE = 20.3 and 23.6 g kg−1, RPIQ = 0.41 and 0.53, respectively). The GLOBAL-LOCAL was better or similar to calibrations with site-specific samples (RMSE = 2.38 and 4.68 g kg−1, RPIQ = 3.5 and 2.7, respectively) and Lab samples only (RMSE = 1.41 and 2.54 g kg−1, RPIQ = 5.9 and 4.9, for Lévis and Maaninka, respectively). In addition, GLOBAL-LOCAL outperformed three other approaches, RS-LOCAL, the spectrum-based learner and LOCAL. While 25 Lab samples was regarded as optimum by cLHS, further examination revealed that as little as 10 Lab samples may be sufficient for the GLOBAL-LOCAL. The method uses the information in the large spectral library so that it can substantially reduce analytical cost and greatly improve predictions on local scales. [ABSTRACT FROM AUTHOR]

A popular and easy to implement classifier is the k-Nearest Neighbour (k-NN). However, sequentially searching for nearest neighbours in large datasets leads to inefficient classification because of the high computational cost involved. This article presents an adaptive hybrid and cluster-based method for speeding up the k-NN classifier. The proposed method reduces the computational cost as much as possible while maintaining classification accuracy at high levels. The method is based on the well-known k-means clustering algorithm and consists of two main parts: (i) a pre-processing algorithm that builds a two-level, cluster-based data structure, and (ii) a hybrid classifier that classifies new items by accessing either the first or the second level of the data structure. The proposed approach was tested on seven real life datasets and the experiential measurements were statistically validated by the Wilcoxon signed ranks test. The results show that the proposed classification method can be used either to achieve high accuracy with slightly higher cost or to reduce the cost at a minimum level with slightly lower accuracy. [ABSTRACT FROM AUTHOR]

Abstract: Human Epithelial (HEp-2) cells are commonly used in the Indirect Immunofluorescence (IIF) tests to detect autoimmune diseases. The diagnosis consists of searching and classification to specific patterns created by Anti-Nuclear Antibodies (ANAs) in the patient serum. Evaluation of the IIF test is mostly done by humans, which means that it is highly dependent on the experience and expertise of the physician. Therefore, a significant amount of research has been focused on the development of computer aided diagnostic systems which could help with the analysis of images from microscopes. This work deals with the design and development of HEp-2 cells classifier. The classifier is able to categorize pre-segmented images of HEp-2 cells into 6 classes. The core of this engine consists of the following image descriptors: Haralick features, Local Binary Patterns, SIFT, surface description and a granulometry-based descriptor. These descriptors produce vectors that form metric spaces. k-NN classification is based on aggregated distance function which combines several features together. An extensive set of evaluations was performed on the publicly available MIVIA HEp-2 images dataset which allows a direct comparison of our approach with other solutions. The results show that our approach is one of the leading classifiers when comparing with other participants in the HEp-2 Cells Classification Contest[1]. [Copyright &y& Elsevier]

Abstract: Nearest neighbour queries (or kNN) have been used in many disciplines, including spatial databases. kNN queries have gone beyond simply finding nearest objects. It is, therefore, important to understand the full spectrum of kNN queries, even before starting to work on how to process and optimise such queries. The aim of this paper is to give a complete picture of what kNN queries are capable of. In this study, we present a survey of kNN queries, in which we propose a taxonomy of kNN queries, comprising four perspectives: (i) a Space perspective, (ii) a Result perspective, (iii) a Query-Point perspective, and (iv) a Relationship perspective. These give a comprehensive overview of kNN queries. As kNN is a large area of research, in this paper, we confine the discussion to kNN queries on stationary objects. [Copyright &y& Elsevier]