Narrow Results

In order to better realize the optimization of university education management, this paper puts forward the research on the optimization path of university education management under collaborative filtering algorithm. The optimization of higher education management is divided into several management directions. First, the teaching resource recommendation management, using collaborative filtering algorithm based on emotional tendency, analyzes the comments of users of teaching resources, and recommends the teaching resources to users next time according to the emotional tendency obtained from the analysis. Second, the library book recommendation management, constructing the characteristic model of library book borrowing users, uses collaborative filtering algorithm to generate nearest neighbors, calculating the similarity of users, and recommending books that meet their characteristics for users according to the similarity. Finally, the management of personalized course selection recommendation in colleges and universities, which uses the historical information of students to establish the evaluation matrix of course selection, searches the nearest neighbors according to the evaluation matrix, and produces the personalized course selection recommendation results of the students according to the nearest neighbors. According to the experiment, the recommended users can be satisfied with the recommended results in different directions of education management. After applying this method, the overall quality of college students has been significantly improved, the expenditure of education management in colleges and universities has been significantly reduced, and the optimization effect of education management in colleges and universities has been remarkable.

Changing the metric on the data may change the data distribution, hence a good distance metric can promote the performance of learning algorithm. In this paper, we address the semi-supervised distance metric learning (ML) problem to obtain the best nonlinear metric for the data. First, we describe the nonlinear metric by the multiple kernel representation. By this approach, we project the data into a high dimensional space, where the data can be well represented by linear ML. Then, we reformulate the linear ML by a minimization problem on the positive definite matrix group. Finally, we develop a two-step algorithm for solving this model and design an intrinsic steepest descent algorithm to learn the positive definite metric matrix. Experimental results validate that our proposed method is effective and outperforms several state-of-the-art ML methods.

Long range or multistep-ahead time series forecasting is an important issue in various fields of business, science and technology. In this paper, we have proposed a modified nearest neighbor based algorithm that can be used for long range time series forecasting. In the original time series, optimal selection of embedding dimension that can unfold the dynamics of the system is improved by using upsampling of the time series. Zeroth order cross-correlation and Euclidian distance criterion are used to select the nearest neighbor from up-sampled time series. Embedding dimension size and number of candidate vectors for nearest neighbor selection play an important role in forecasting. The size of embedding is optimized by using auto-correlation function (ACF) plot of the time series. It is observed that proposed algorithm outperforms the standard nearest neighbor algorithm. The cross-correlation based criteria shows better performance than Euclidean distance criteria.

Multistep ahead time series forecasting has become an important activity in various fields of science and technology due to its usefulness in future events management. Nearest neighbor search is a pattern matching algorithm for forecasting, and the accuracy of the method considerably depends on the similarity of the pattern found in the database with the reference pattern. Original time series is embedded into optimal dimension. The optimal dimension is determined by using autocorrelation function plot. The last vector in the embedded matrix is taken as the reference vector and all the previous vectors as candidate vectors. In nearest neighbor algorithm, the reference vector is matched with all the candidate vectors in terms of Euclidean distance and the best matched pattern is used for forecasting. In this paper, we have proposed a hybrid distance measure to improve the search of the nearest neighbor. The proposed method is based on cross-correlation and Euclidean distance. The candidate patterns are shortlisted by using cross-correlation and then Euclidean distance is used to select the best matched pattern. Moreover, in multistep ahead forecasting, standard nearest neighbor method introduces a bias in the search which results in higher forecasting errors. We have modified the search methodology to remove the bias by ignoring the latest forecasted value during the search of the nearest neighbor in the subsequent iteration. The proposed algorithm is evaluated on two benchmark time series as well as two real life time series.

The nearest neighbor method is one of the most widely used pattern classification methods. However its major drawback in practice is the curse of dimensionality. In this paper, we propose a new method to alleviate this problem significantly. In this method, we attempt to cover the training patterns of each class with a number of hyperspheres. The method attempts to design hyperspheres as compact as possible, and we pose this as a quadratic optimization problem. We performed several simulation experiments, and found that the proposed approach results in considerable speed-up over the k-nearest-neighbor method while maintaining the same level of accuray. It also significantly beats other prototype classification methods (Like LVQ, RCE and CCCD) in most performance aspects.

One-class extraction from remotely sensed imagery is researched with multi-class classifiers in this paper. With two supervised multi-class classifiers, Bayesian classifier and nearest neighbor classifier, we firstly analyzed the effect of the data distribution partitioning on one-class extraction from the remote sensing images. The data distribution partitioning refers to the way that the data set is partitioned before classification. As a parametric method, the Bayesian classifier achieved good classification performance when the data distribution was partitioned appropriately. While as a nonparametric method, the NN classifier did not require a detailed partitioning of the data distribution. For simplicity, the data set can be partitioned into two classes, the class of interest and the remainder, to extract the specific class. With appropriate partitioning of the data set, the specific class of interest was well extracted from remotely sensed imagery in the experiments. This study will be helpful for one-class extraction from remote sensing imagery with multi-class classifiers. It provides a way to improve the one-class classification from the aspect of data distribution partitioning.

Learning when limited to modification of some parameters has a limited scope; capability to modify the system structure is also needed to get a wider range of the learnable. In the case of artificial neural networks, learning by iterative adjustment of synaptic weights can only succeed if the network designer predefines an appropriate network structure, i.e. the number of hidden layers, units, and the size and shape of their receptive and projective fields. This paper advocates the view that the network structure should not, as is usually done, be determined by trial-and-error but should be computed by the learning algorithm. Incremental learning algorithms can modify the network structure by addition and/or removal of units and/or links. A survey of current connectionist literature is given on this line of thought. “Grow and Learn” (GAL) is a new algorithm that learns an association at one shot due to its being incremental and using a local representation. During the so-called “sleep” phase, units that were previously stored but which are no longer necessary due to recent modifications are removed to minimize network complexity. The incrementally constructed network can later be finetuned off-line to improve performance. Another method proposed that greatly increases recognition accuracy is to train a number of networks and vote over their responses. The algorithm and its variants were tested on recognition of handwritten numerals and seem promising especially in terms of learning speed. This makes the algorithm attractive for on-line learning tasks, e.g. in robotics. The biological plausibility of incremental learning is also discussed briefly.

The main aim of this paper is to introduce the single nearest neighbor approach for pattern recognition and the concept of incremental learning of a fuzzy classifier where decision making is based on data available up to time t rather than what may be available at the start of the trial, i.e. at t = 0. The single nearest neighbor method is explained in the context of solving the classic two-spiral benchmark. The proposed approach is further tested on the electronic nose coffee data to judge its performance on a real problem. This paper illustrates: (1) a novel fuzzy classifier system based on the single nearest neighbor method, (2) its application to the spiral benchmark taking the incremental pattern recognition approach, and (3) results obtained when solving the two-spiral problem with both nonincremental and incremental methods and coffee classification with the nonincremental method. The results show that incremental learning leads to improved recognition performance for spiral data and it is possible to study the behavioral characteristics of the classifier with possibility related parameters.

Synthesis and optimization of quantum circuits have received significant attention from researchers in recent years. Developments in the physical realization of qubits in quantum computing have led to new physical constraints to be addressed. One of the most important constraints that is considered by many researchers is the nearest neighbor constraint which limits the interaction distance between qubits for quantum gate operations. Various works have been reported in the literature that deal with nearest neighbor compliance in multi-dimensional (mostly 1D and 2D) qubit arrangements. This is normally achieved by inserting SWAP gates in the gate netlist to bring the interacting qubits closer together. The main objective function to minimize here is the number of SWAP gates. The present paper proposes an efficient qubit placement strategy in a three-dimensional (3D) grid that considers not only qubit interactions but also the relative positions of the gates in the circuit. Experimental evaluation on a number of benchmark circuits show that the proposed method reduces the number of SWAP gates by 16.2% to 47.0% on the average as compared to recently published works.

In the last couple of years, quantum computing has come out as emerging trends of computation not only due to its immense popularity but also for its commitment towards physical realization of quantum circuit in on-chip units. At the same time, the process of physical realization has faced several design constraints and one such problem is nearest neighbor (NN) enforcement which demands all the operating qubits to be placed adjacent in the implementable circuit. Though SWAP gate embedment can transform a design into NN architecture, it still creates overhead in the design. So, designing algorithms to restrict the use of SWAPs bears high importance.

Considering this fact, in this work, we are proposing a heuristic-based improved qubit placement strategy for efficient implementation of NN circuit. Two different design policies are being developed here. In the first scheme, a global reordering technique based on clustering approach is shown. In the second scheme, a local reordering technique based on look-ahead policy is developed. This look-ahead strategy considers the impact over the gates in the circuit and thereby estimates the effect using a cost metric to decide the suitable option for SWAP implementation. Furthermore, the joint use of both the ordering schemes also has been explored here. To ascertain the correctness of our design algorithms, we have tested them over a wide range of benchmarks and the obtained results are compared with some state-of-the-art design approaches. From this comparison, we have witnessed a considerable reduction on SWAP cost in our design scheme against the reported works’ results.

We present a new multi-dimensional data structure, which we call the skip quadtree (for point data in R²) or the skip octree (for point data in R^d, with constant d > 2). Our data structure combines the best features of two well-known data structures, in that it has the well-defined “box”-shaped regions of region quadtrees and the logarithmic-height search and update hierarchical structure of skip lists. Indeed, the bottom level of our structure is exactly a region quadtree (or octree for higher dimensional data). We describe efficient algorithms for inserting and deleting points in a skip quadtree, as well as fast methods for performing point location, approximate range, and approximate nearest neighbor queries.

We propose a Generalized Nearest Prototype Classifier (GNPC) as a common framework for a number of classification techniques. Specifically we consider clustering-and-relabeling; Parzen's classifier; radial basis functions (RBF) networks; learning vector quantization (LVQ) type classifiers; and nearest neighbor rules. To classify an unlabeled point x the GNPC combines the degrees of similarity of x to a set of prototypes. Five questions are addressed for these GNPC families: (1) How many prototypes do we need? (2) How are the prototypes found? (3) How are their class labels obtained? (4) How are the similarities defined? (5) How are the similarities and label information combined? The classification performance of a set of GNPCs is illustrated on two benchmark data sets: IRIS and the 2-spirals data. We study the resubstitution error of the GNPC as a function of the number of prototypes. Our conclusions are that: (a) unsupervised selection (or extraction) of prototypes followed by relabeling is inferior to the techniques that use labels to guide them towards prototypes; (b) the edited nearest neighbor rule is a viable option for GNPC design which has not received the attention it deserves.

Case-based reasoning (CBR) is a problem-solving paradigm that uses past experiences to solve new problems. Nearest neighbor is a common CBR algorithm for retrieving similar cases, whose similarity function is sensitive to irrelevant attributes. Taking the relevancy of the attributes into account can reduce this sensitivity, leading to a more effective retrieval of similar cases. In this paper, statistical evaluation is used for assigning relative importance of the attributes. This approach is applied to predict business failures in Australia using financial data. The results in this study indicate it is an effective and competitive alternative to predict business failures in a comprehensible manner. This study also investigates the usefulness of non-financial data derived from auditor's and directors' reports for business failure prediction. The results suggest that the particular non-financial attributes identified are not as effective as the financial attributes in explaining business failures.

This paper presents a novel method for computer vision-based static and dynamic hand gesture recognition. Haar-like feature-based cascaded classifier is used for hand area segmentation. Static hand gestures are recognized using linear discriminant analysis (LDA) and local binary pattern (LBP)-based feature extraction methods. Static hand gestures are classified using nearest neighbor (NN) algorithm. Dynamic hand gestures are recognized using the novel text-based principal directional features (PDFs), which are generated from the segmented image sequences. Longest common subsequence (LCS) algorithm is used to classify the dynamic gestures. For testing, the Chinese numeral gesture dataset containing static hand poses and directional gesture dataset containing complex dynamic gestures are prepared. The mean accuracy of LDA-based static hand gesture recognition on the Chinese numeral gesture dataset is 92.42%. The mean accuracy of LBP-based static hand gesture recognition on the Chinese numeral gesture dataset is 87.23%. The mean accuracy of the novel dynamic hand gesture recognition method using PDF on directional gesture dataset is 94%.