The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
A novel feature selection method was proposed for electromyography (EMG)-based affective recognition. First of all, correlation analysis was used to reduce the dimension of original feature subset; then adaptive Tabu search algorithm combined with intensification and diversification strategies was adopted for feature selection, and mutation operator of genetic algorithm (GA) was implemented as the...
In this paper, several clustering methodologies are investigated in order to group together wind parks with close statistical behaviour. The proposed approach is practically founded on a fast incremental algorithm. The latter requires the definition of an objective function which is based in the present case on the definition of a Pearson correlation coefficient level. The advantage of such a clustering...
A fundamental problem in machine learning is to discriminate a representative set of features on which to construct a classification model for a particular task. This paper presents a feature selection algorithm RF-MI for multiple classes based on ReliefF algorithm and Mutual Information (MI) measure. RF-MI algorithm gets a feature subset by excluding irrelevant and redundant features from original...
In this paper two different methods for non-technical losses (NTL) detection are analyzed and new approach is proposed, based on the noticed drawbacks. It is shown that NTL can be successfully detected by a neural network trained by “artificial”, i.e., generated samples. This approach eliminates the need for many hard-to-obtain real life samples and the network can easily be trained to detect some...
Mining of association rules has become an important area in the research on data mining. However the traditional approaches based on support-confidence framework maybe generate a great number of redundant and wrong association rules. In order to solve the problems, a correlation measure is defined and added to the mining algorithm for association rules. According to the value of correlation measure,...
H.264/AVC is a powerful and high performance video compression standard. To achieve high coding efficiency, it employs many new techniques, such as spatial prediction in intra coding, adaptive block size inter coding and motion compensation, multiple reference pictures, and so on. Especially the complex intra and inter encoding modes, obtain notable coding gains. The high computation burden of mode...
Aimed to improve the detection efficiency of information-hiding blind detection system, the present study proposes an SADRID-I image steganalysis algorithm that is based on improved differential matrix, according to the high dimensions and correlation of image features. Using attribute significance and differential matrix that stems from rough sets theory, the algorithm can implement D attributes...
According to the definition of autocorrelation matrix (ACM), this paper was proposed a new autocorrelation matrix construction method. In comparison to traditional least square algorithm based on eigenvalue decomposition (LS-EVD), the presented method improved the estimation accuracy of eigenvalue. In addition, dividing signal subspace and noise subspace in a new method, which reduced the estimation...
The basic algorithms for both lossless and lossy compression of images are discussed in the paper. The criteria of the contrastive analysis of compression algorithms are chosen. The classes of images are discussed. The results of algorithms realization and testing in MATLAB environment are shown in the diagrams and discussed. The recommendations on future improvement of image compression algorithms...
In text categorization, feature selection is an effective feature dimension-reduction methods. To solve the problems of unadaptable high original feature space dimension, too much irrelevance, data redundancy and difficulties in selecting a threshold, we propose an improved LAM feature selection algorithm (ILAMFS). Firstly, combining the gold segmentation and the LAM algorithm based on the characteristics...
Naive Bayes algorithm is a simple and efficient classification algorithm, but its conditional independence assumption is not always true in real life which is affected to some extent. Weighted Naive Bayesian classifier relax the conditional independence assumption to increase accuracy. Based on Identifiably matrix of Rough Set, a new weighted naive Bayes method based on attribute frequency is proposed...
Module partition is the key part of the modular design technology which has been widely used in increasing fields. To divide product objective and reasonably will build the good foundation for each stage of the module design realized. Firstly, the principles of module partition are proposed in this paper. Then the hybrid fuzzy clustering algorithm based on PSO is used to make module division of the...
Due to the explosive issue of attributes combination, the minimum reduction of decision-making tables is the NP-hard issue. But it is convenient and fast to decide the reduction scope by fractal dimension. The paper discusses the relationship between inherent dimension and the fractal dimension, explains the fractal dimension of one data set reflects the inherent characteristics of the data, and proposes...
In the field of imbalance learning and cost sensitive learning, minimization of the classification error rate is not an appropriate approach due to class skew and cost distributions. Thus the area under the ROC Curve (AUC) has been widely utilized to assess the performance of the classifiers in such cases. The Maximum AUC Linear Classifier (MALC), aiming at maximizing AUC directly, is a nonparametric...
E-commerce commodities contain a large number of associated information, an algorithm on how to mine association rules based on rough set is proposed in this paper. The different feature vectors extracted from different types of commodities can be looked as a prerequisite for getting association rules. By using the knowledge reduction theory the associated commodities can be drown as a minimum set...
In the process of iris classification, a new classification distance with adjustable weight which takes advantage of whole phase information to encode is proposed. The method is to use feature extraction function to do the extraction toward all iris image, which could obtain real and imaginary part iris information. Then, tangent function is used to transform extracted real and imaginary part characteristics...
Correlation algorithm in the frequency domain is now in widespread use due to Fast Fourier Transform Algorithm. However, its calculating speed and accuracy need to be further studied for the limited length sequence. A kind of discrete correlation algorithm in the frequency domain was derived based on analyzing the correlation theorem and vectors in the frequency domain. As is shown from the simulation...
Cluster analysis, a set of tools for building groups from multivariate data objects is extensively applied in many fields. One of the most widely used classical approaches of clustering is K-means algorithm. Kohonen's Self Organizing map is a neural network clustering methodology that maps an n-dimensional input data to a lower dimensional output map. In this study, we have compared K-means algorithm...
One major challenge in image compression is to efficiently represent and encode high-frequency structure components in images, such as edges, contours, and texture regions. To address this issue for lossy image compression, in our previous work, we proposed a scheme to learn local image structures and efficiently predict image data based on this structure information. In this work, we applied this...
In recent years, Internet traffic classification using machine learning has become a new direction in network measurement. Because supervised clustering algorithm need accuracy of training sets and it can not classify unknown application, we introduced complex network's community detecting algorithm, a new unsupervised classify algorithm, which have previously not been used for network traffic classification...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.