Nearest-neighbor Type Research Articles

Analisa Predictive Berbasis Supervised Machine Learning Terhadap Kerusakan Peralatan Pembangkit

Predictive maintenance is a treatment for the actual operation of the equipment to optimize the company's operations. The output of predictive program maintenance is data, this treatment includes the type of "condition based maintenance" where changes in the condition of the machine or equipment are detected so that proactive actions are taken before the occurrence of machine damage. The K-nearest Neighbor (K-NN) algorithm is a simple supervised machine learning algorithm that is used to solve problems based on classification and regression. K-NN works by finding the query distance and all database examples, selecting a certain number of examples (K) adjacent to the query, then selecting the frequent label (in classification) or the average label (in regression). The purpose of this algorithm is to classify new object conditions based on attributes and samples from the training database. So that a predictive analysis is carried out on the damage to generating equipment using the machine learning application method of the Nearest Neighbor type or the classification of conditions used to predict the age or condition of an equipment by modeling according to the standard Operation & Maintenance of equipment. By doing predictive analysis, maintenance will lead to condition based maintenance so that the KPI (Key Performance Indicator) of operating performance in the form of increasing values, such as Capacity Factor (CF), Equivalent Availbility Factor (EAF) becomes optimal and prevents the generator from tripping suddenly. which is called sudden outage frequency (SdOF), as well as more efficient maintenance costs.

A nearest-neighbor-based ensemble classifier and its large-sample optimality

ABSTRACT A nonparametric approach is proposed to combine several individual classifiers in order to construct an asymptotically more accurate classification rule in the sense that its misclassification error rate is, asymptotically, at least as low as that of the best individual classifier. The proposed method uses a nearest neighbour type approach to estimate the conditional expectation of the class associated with a new observation (conditional on the vector of individual predictions). Both mechanics and the theoretical validity of the proposed approach are discussed. As an interesting byproduct of our results, it is shown that the proposed method can also be applied to any single classifier in which case the resulting new classifier will be at least as good as the original one. Several numerical examples, involving both real and simulated data, are also given. These numerical studies further confirm the superiority of the proposed classifier.

Phase transitions for a model with uncountable spin space on the Cayley tree: the general case

In this paper we complete the analysis of a statistical mechanics model on Cayley trees of any degree, started in Botirov (Positivity 21(3):955–961, 2017), Eshkabilov et al. (J Stat Phys 147(4):779–794, 2012), Eshkabilov and Rozikov (Math Phys Anal Geom 13:275–286, 2010), Botirov et al. (Lobachevskii J Math 34(3):256–263 2013) and Jahnel et al. (Math Phys Anal Geom 17:323–331 2014). The potential is of nearest-neighbor type and the local state space is compact but uncountable. Based on the system parameters we prove existence of a critical value $$\theta _{\mathrm{c}}$$ such that for $$\theta \le \theta _{\mathrm{c}}$$ there is a unique translation-invariant splitting Gibbs measure. For $$\theta _{\mathrm{c}}<\theta $$ there is a phase transition with exactly three translation-invariant splitting Gibbs measures. The proof rests on an analysis of fixed points of an associated non-linear Hammerstein integral operator for the boundary laws.

The Role of Intrinsically Disordered Linker and Salt Type on Cooperativity of E. Coli SSB Binding to ssDNA

E. coli single strand DNA binding protein (SSB) is a key protein in DNA replication, recombination and repair. SSB functions as a homotetramer and binds ssDNA in different modes that differ in their excluded site size and cooperative binding to ssDNA. Two types of cooperativity have been identified previously: the high “unlimited”, nearest-neighbor type of cooperativity associated with the 35 site size ((SSB)35) binding mode at low [NaCl], and a lower, “limited” cooperativity associated with formation of dimers of tetramers in the 65 site size ((SSB)65) binding mode at [NaCl] > 200 mM. Using sedimentation velocity and equilibrium titration methods we have examined how salt concentration and type affects these cooperativities and binding mode transitions on long ssDNA, focusing on the effects of KGlu, since glutamate is the major monovalent anion in E. coli. Unexpectedly, we find that in contrast to NaCl, KGlu promotes highly cooperative binding to ssDNA even at high concentrations of 0.5 M, conditions that promote the ((SSB)65) mode, suggesting that highly cooperative binding is likely occurring in vivo. We also show that all cooperative binding behavior absolutely requires the unstructured C-terminal region (intrinsically disordered linker (IDL)) connecting the SSB DNA binding core and the 9 amino acid C-terminal end (tip) needed for SSB binding to other proteins involved in genome maintenance (supported by NIH GM030498 to TML).

Construction of three-dimensional models of bimetallic nanoparticles based on X-ray absorption spectroscopy data

A new method for constructing three-dimensional models of bimetallic nanoparticles is proposed. This method, which is based on X-ray absorption spectroscopy data on the number and type of nearest neighbors, provides information on the distribution of types of atoms over the nanoparticle volume. The application of the method to the study of the structures of platinum–copper and platinum–silver nanoparticles of metal–carbon electrocatalysts has allowed to distinguish the nanoparticles with a core–shell structure from the nanoparticles with structure of disordered alloy or clusterized solid solution.

Reflection couplings and contraction rates for diffusions

We consider contractivity for diffusion semigroups w.r.t. Kantorovich ($L^1$ Wasserstein) distances based on appropriately chosen concave functions. These distances are inbetween total variation and usual Wasserstein distances. It is shown that by appropriate explicit choices of the underlying distance, contractivity with rates of close to optimal order can be obtained in several fundamental classes of examples where contractivity w.r.t. standard Wasserstein distances fails. Applications include overdamped Langevin diffusions with locally non-convex potentials, products of these processes, and systems of weakly interacting diffusions, both of mean-field and nearest neighbour type.

On high dimensional two-sample tests based on nearest neighbors

In this article, we propose new multivariate two-sample tests based on nearest neighbor type coincidences. While several existing tests for the multivariate two-sample problem perform poorly for high dimensional data, and many of them are not applicable when the dimension exceeds the sample size, these proposed tests can be conveniently used in the high dimension low sample size (HDLSS) situations. Unlike Schilling (1986) [26] and Henze’s (1988) test based on nearest neighbors, under fairly general conditions, these new tests are found to be consistent in HDLSS asymptotic regime, where the sample size remains fixed and the dimension grows to infinity. Several high dimensional simulated and real data sets are analyzed to compare their empirical performance with some popular two-sample tests available in the literature. We further investigate the behavior of these proposed tests in classical asymptotic regime, where the dimension of the data remains fixed and the sample size tends to infinity. In such cases, they turn out to be asymptotically distribution-free and consistent under general alternatives.

Isolated centres versus defect associates in Sm3+-doped CeO2: a spectroscopic investigation

The interactions between Sm3+ and oxygen vacancies in CeO2 are probed by the use of tuneable laser excited time-resolved photoluminescence and Raman spectroscopies. It is found that Sm3+ (with doping concentrations of 0.1, 0.3, 1 and 5 wt%) substitutes largely for Ce4+ in sites with cubic symmetry and the corresponding emission is sensitized via the Ce4+–O2− charge-transfer band of CeO2. It is established from the photoluminescence spectra measured at long delay after the laser pulse that the local environment around cubic Sm3+ centres is not changed with concentration and ceria size. In addition to cubic symmetry Sm3+ centres, low-symmetry Sm3+ centres tentatively assigned to the Sm3+–oxygen vacancy associates of nearest-neighbour type are also observed. Their emission is preferentially excited via the weak f–f absorption transitions of Sm3+. A relatively strong concentration-induced quenching of Sm3+ emission was inferred from the decrease in the average emission lifetimes from 2.1 ms (0.1 wt%) to 0.87 ms (5 wt%). The local environments of Sm3+ and Eu3+ in CeO2 are also compared on the basis of their emission spectra and decays.

Wetting in electrolyte solutions

Wetting of a charged substrate by an electrolyte solution is investigated by means of classical density functional theory applied to a lattice model. Within the present model the pure, i.e., salt-free solvent, for which all interactions are of the nearest-neighbor type only, exhibits a second-order wetting transition for all strengths of the substrate-particle and the particle-particle interactions for which the wetting transition temperature is nonzero. The influences of the substrate charge density and of the ionic strength on the wetting transition temperature and on the order of the wetting transition are studied. If the substrate is neutral, the addition of salt to the solvent changes neither the order nor the transition temperature of the wetting transition of the system. If the surface charge is nonzero, upon adding salt this continuous wetting transition changes to first-order within the wide range of substrate surface charge densities and ionic strengths studied here. As the substrate surface charge density is increased, at fixed ionic strength, the wetting transition temperature decreases and the prewetting line associated with the first-order wetting transition becomes longer. This decrease of the wetting transition temperature upon increasing the surface charge density becomes more pronounced by decreasing the ionic strength.

Development of spiral wave in a regular network of excitatory neurons due to stochastic poisoning of ion channels

Some experimental evidences show that spiral wave could be observed in the cortex of brain, and the propagation of this spiral wave plays an important role in signal communication as a pacemaker. The profile of spiral wave generated in a numerical way is often perfect while the observed profile in experiments is not perfect and smooth. In this paper, formation and development of spiral wave in a regular network of Morris–Lecar neurons, which neurons are placed on nodes uniformly in a two-dimensional array and each node is coupled with nearest-neighbor type, are investigated by considering the effect of stochastic ion channels poisoning and channel noise. The formation and selection of spiral wave could be detected as follows. (1) External forcing currents with diversity are imposed on neurons in the network of excitatory neurons with nearest-neighbor connection, a target-like wave emerges and its potential mechanism is discussed; (2) artificial defects and local poisoned area are selected in the network to induce new wave to interact with the target wave; (3) spiral wave can be induced to occupy the network when the target wave is blocked by the artificial defects or poisoned area with regular border lines; (4) the stochastic poisoning effect is introduced by randomly modifying the border lines (areas) of specific regions in the network. It is found that spiral wave can be also developed to occupy the network under appropriate poisoning ratio. The process of growth for the poisoned area of ion channels poisoning is measured, the effect of channels noise is also investigated. It is confirmed that perfect spiral wave emerges in the network under gradient poisoning even if the channel noise is considered.

Computing sparse derivatives and consecutive zeros problem

We describe a substitution based sparse Jacobian matrix determination method using algorithmic differentiation. Utilizing the a priori known sparsity pattern, a compression scheme is determined using graph coloring. The "compressed pattern" of the Jacobian matrix is then reordered into a form suitable for computation by substitution. We show that the column reordering of the compressed pattern matrix (so as to align the zero entries into consecutive locations in each row) can be viewed as a variant of traveling salesman problem. Preliminary computational results show that on the test problems the performance of nearest-neighbor type heuristic algorithms is highly encouraging.

Strain Modulation of Defect Structure in Gadolinia-Doped Ceria

The relationship between the strain and defect structure in gadolinia-doped ceria has been studied with atomistic simulations. It was found that the next nearest neighbor type associations are preferred under zero or tensile applied strains, but the nearest neighbor type defect associations became favorable under compressive strains. Molecular dynamics simulations suggest that the defect structure can be modulated by strain through oxygen diffusion into or out of the nearest neighbor sites of dopants. The activation energy of such a process was estimated to be 0.82 eV. The decrease or increase in the amount of nearest neighbor associations also causes expansion or contraction, respectively, of the material volume. On the basis of such a mechanism, an explanation of the “chemical strain/stress” phenomenon observed in gadolinia-doped ceria, as well as the benefit of zero or moderate tensile strain for electrolyte applications, is proposed.

Woody dominance in a semi-arid savanna rangeland – Evidence for competitive self-thinning

Self-thinning has been considered an important mechanism facilitating species coexistence in nature. The purpose of this study was to examine the role of self-thinning in regulating the density of woody plants in semi-arid savannas. Previous models (e.g. the ‘honeycomb rippling model’) have postulated that, as trees grow older (inferred by size), the interplant distance, taken to reflect competition-induced mortality, becomes greater and more even. Therefore, the opening of canopies in the long-run promotes species coexistence. To study the role of self-thinning in a semi-arid savanna we selected four savanna patches and sampled the plant density, size, and distance between the tallest woody plant and its nearest conspecific and heterospecific neighbors, along with other environmental variables. Interplant distance was positively influenced by species, tree size, woody species diversity, and nearest neighbor type (conspecific vs. heterospecific). Distances did not increase evenly with increasing size of the tallest tree or shrub, however. We conclude that self-thinning is evident in woody encroached semi-arid savanna patches, as suggested by the honeycomb rippling model. The mechanism may play a significant role for the long-term stability of the savanna by opening dense canopies and facilitating tree–grass coexistence.

Expander Families as Information Patterns for Distributed Control of Vehicle Platoons

A state space formulation of the distributed control of a 1-D vehicle platoon is considered. The objective is to understand the effects of the underlying information exchange pattern between the vehicles on the control performance of the platoon. The symmetric control case is considered where each vehicle gives equal weight to all the information available to it in determining its control law. It is shown that expander families of graphs when used as information patterns result in stability margins decaying to zero at rate at most O(1/N); an improvement over the previously known O(1/N2) decay with nearest neighbor type information patterns.

Effects of Long-Range Correlations on Nonmagnetic Mott Transitions in Hubbard Model on Square Lattice

The mechanism of Mott transition in the Hubbard model on the square lattice is studied without explicit introduction of magnetic and superconducting correlations, using a variational Monte Carlo method. In the trial wave functions, we consider various types of binding factors between a doubly-occupied site (doublon, D) and an empty site (holon, H), like a long-range type as well as a conventional nearest-neighbor type, and add independent long-range D-D (H-H) factors. It is found that a wide choice of D-H binding factor leads to Mott transitions at critical values near the band width. We renew the D-H binding picture of Mott transitions by introducing two characteristic length scales, the D-H binding length l_{DH} and the minimum D-D distance l_{DD}, which we appropriately estimate. A Mott transition takes place at l_{DH}=l_{DD}. In the metallic regime (l_{DH}>l_{DD}), the domains of D-H pairs overlap with one another, thereby doublons and holons can move independently by exchanging the partners one after another. In contrast, the D-D factors give only a minor contribution to the Mott transition.

Improving the Identification of RBF Predictive Models to Forecast the Portuguese Electricity Consumption

The Portuguese power grid company wants to improve the accuracy of the electricity load demand forecast within an horizon of 48 hours, in order to identify the need of reserves to be allocated in the Iberian Market. In this work we present updated results on the identification of radial basis function neural network load demand predictive models. The methodology follows the principles already employed by the authors in different applications: the NN models are trained by the Levenberg-Marquardt algorithm using a modified training criterion, and the model structure (number of neurons and input terms) is evolved using a multi-objective genetic algorithm. The set of goals and objectives used in the model optimisation reflect different requirements in the design: obtaining good generalisation ability, good balance between one-step-ahead prediction accuracy and model complexity, and good multi-step prediction accuracy. In this work the prediction horizon was increased, the model fitness assessment was altered, and the model structure search space was enlarged. Results are also presented for a predictive nearest neighbour type approach, which establishes a baseline for predictive methods comparison.

Tests of Bivariate Exchangeability

Summary We discuss the hypothesis of bivariate exchangeability and show that testing bivariate exchangeability is related to the two‐sample testing of equality of distribution functions. We consider three test statistics based on the ordering of the Euclidean interpoint distances. The runs test of exchangeability counts the runs among the observations and their mirror images on the minimal spanning tree. The nearest neighbour test of exchangeability is based on the number of nearest neighbour type coincidences among the observations and their folded images on the plane. The rank test of exchangeability compares the within and between ranks of the interpoint distances. We also consider the sign test of exchangeability, which uses the signs of the observations in specific regions, and a bootstrap test of exchangeability based on the maximum distance between the mirror images. We compare the power of these methods in a Monte Carlo study which shows different power orderings of the methods, depending on the alternative hypothesis.

A k-nearest neighbor approach for functional regression

Let ( X , Y ) be a random pair taking values in H × R , where H is an infinite dimensional separable Hilbert space. We establish weak consistency of a nearest neighbor type estimator of the regression function of Y on X based on independent observations of the pair ( X , Y ) . As a general strategy, we propose to reduce the infinite dimension of H by considering only the first d coefficients of an expansion of X in an orthonormal system of H , and then to perform k -nearest neighbor regression in R d . Both the dimension and the number of neighbors are automatically selected from the observations using a simple data-dependent splitting device.

A comparison of two zone-visitation sequencing strategies in a distribution centre

The environment of this study is modelled after a distribution centre in Taiwan. One unique characteristic of this environment is that the picking area has been divided into different zones. A picker often has to visit more than one zone to pick all the items in an order. This environment has one important restriction - no more than one picker can simultaneously be in the same zone. This paper focuses on one problem of this environment, i.e. finding the best zone-visitation sequence a picker can follow. Two zone-visitation sequencing strategies - the fixed zone-visitation sequencing strategy and the dynamic zone-visitation sequencing strategy - are proposed and studied for the problem. Both strategies incorporate a two-phase optimization process. An initial starting solution is found using a nearest-neighbour type generative algorithm NPZF, then a simulated annealing improvement algorithm SA brings this solution towards optimality. This ensures that each strategy is represented by good exemplars of their class. Simulation experiments are conducted to evaluate the performance of the proposed strategies in three performance measures - total travel distance of pickers (TTD), total required system Time to complete the picking operations (TRST), and total picking time of pickers (TPT). The simulation results show that the fixed strategy is about 20.9% better than the dynamic strategy in TTD, but about 17% less effective than the dynamic strategy in TRST and TPT. This indicates that the dynamic strategy will be a better strategy if one can prevent the possible physical fatigue experienced by pickers due to the greater travel distance.

Long-range integrable spin chains and plane-wave matrix theory

Quantum spin chains arise naturally from perturbativelarge-N field theories and matrix models. The Hamiltonian of such a model is a long-rangedeformation of nearest-neighbour type interactions. Here, we study the most generallong-range integrable spin chain with spins transforming in the fundamental representationof . We derive the Hamiltonian and the corresponding asymptotic Bethe ansatz at theleading four perturbative orders with several free parameters. Furthermore, we proposeBethe equations for all orders and identify the moduli of the integrable system. We finallyapply our results to plane-wave matrix theory and show that the Hamiltonian in aclosed sector is not of this form and therefore not integrable beyond the firstperturbative order. This also implies that the complete model is not integrable.