جستجوی مقالات مرتبط با کلیدواژه "support vector machine" در نشریات گروه "ریاضی"

‎The utilization of model checking has been suggested as a formal verification technique for analyzing critical systems‎. ‎However‎, ‎the primary challenge in applying to complex systems is the state space explosion problem‎. ‎To address this issue‎, ‎bisimulation minimization has emerged as a prominent method for reducing the number of states in a system‎, ‎aiming to overcome the difficulties associated with the state space explosion problem‎. ‎For systems with stochastic behaviors‎, ‎probabilistic bisimulation is employed to minimize a given model‎, ‎obtaining its equivalent form with fewer states‎. ‎In this paper‎, ‎we propose a novel technique to partition the state space of a given probabilistic model to its bisimulation classes‎. ‎This technique uses the PRISM program of a given model and constructs some small versions of the model to train a classifier‎. ‎It then applies supervised machine learning techniques to approximately classify the related partition‎. ‎The resulting partition is then used to accelerate the standard bisimulation technique‎, ‎significantly reducing the running time of the method‎. ‎The experimental results show that the approach can decrease significantly the running time compared to state-of-the-art tools‎.

The paper discusses the limitations of emotion recognition in Persian speech due to inefficient feature extraction and classification tools‎. ‎To address this‎, ‎we propose a new method for detecting hidden emotions in Persian speech with higher recognition accuracy‎. ‎The method involves four steps‎: ‎preprocessing‎, ‎feature description‎, ‎feature extraction‎, ‎and classification‎. ‎The input signal is normalized in the preprocessing step using single-channel vector conversion and signal resampling‎. ‎Feature descriptions are performed using Mel-Frequency Cepstral Coefficients and Spectro-Temporal Modulation techniques‎, ‎which produce separate feature matrices‎. ‎These matrices are then merged and used for feature extraction through a Convolutional Neural Network‎. ‎Finally‎, ‎a Support Vector Machine with a linear kernel function is used for emotion classification‎. ‎The proposed method is evaluated using the Sharif Emotional Speech dataset and achieves an average accuracy of 80.9% in classifying emotions in Persian speech‎.

In this study, a support vector machine (SVM) based technique for timing irrigation projects is presented, and one of the most accurate predictive models in calculating the final project duration within the contract documents, where the research problem is projects are not completed within the contract period because most of the total project duration is determined In an unthoughtful manner by the employer. Linear regression models were applied to data and information for several projects, and a significant improvement in forecast accuracy was obtained.

Investors, creditors, government and other users of financial statements rely on financial information given by the managers of firms to make logical and reasonable decisions. In many cases, the purposes of providers are contradictory to the users’ ones. Therefore, auditing is a tool to enhance the reliability of financial statements presented by firms. In the current research, the selection of an optimal method to predict the report type of independent auditor has been addressed and two approaches of vector machine and neural network have been compared. It was conducted during 2008-2017. 84 firms were reviewed. To train and test the research variables, Voka software has been implemented. The dependent variable is the report type of auditor. Results indicated that the accuracy of the support vector machine algorithm was computed as 66.13% and 56.74% for the training and testing sections, respectively. As well, the accuracy of the neural network model was 61.24% and 55.02% in the training and testing sections, respectively. The support vector machine model was more effective than the neural network.

Parkinson's Disease (PD) is a neurodegenerative disorder that affects predominantly neurons in the brain. The main purpose of this paper is to define a way in detecting the PD in its early stages. This has been achieved through the use of recorded speech, a biomarker in the natural environment in its original state.  In this paper, the Mel-Frequency Cepstral Coefficients (MFCC)  method is utilized to extract features from the recorded speech. The principal component analysis (PCA) and Genetic algorithm (GA) are then applied for feature extraction/selection. Once the features are selected, multiple classifiers are then applied for classification. Performance metrics such as accuracy, specificity, and sensitivity are measured. The result shows that Support Vector Machine (SVM) along with the GA has shown optimal performance.

The fuzzy support vector machine is one of the most exceptional methods to deal with uncertainty in the classification problem. The membership function is a proper way to model uncertainty. The goal of the membership function is to distinguish the different points in terms of their importance. The ordinary design of the membership function relies on the distance of the observations to the class center. However, the class center is affected by the presence of outliers.  To prevent this effect, we utilized an unsupervised learning method called the Gaussian mixture model in the structure of the membership function. The proposed membership function is presented in two different categories distance-based and Bayes-based. Unlike the classical membership function, the contribution of outliers in the training phase decreased by diminishing their degree of importance.  Hybridizing the classic fuzzy support vector machine classifier with the Gaussian mixture model will enhance the classification accuracy and also will prevent overfitting problems. The superiority of the proposed methods assessed by the synthetic and benchmarking dataset. The statistical significance is assessed by using the non-parametric Friedman and post-hoc Nemenyi tests.

Ensuring the production of non-defect high-quality tires is an essential part of the tire industry. X-ray inspection is one of the best methods to detect tire defects. In this paper, a new approach has been presented for detecting tire defects in X-ray images based on an entropy filter, the extraction of texture properties of patches by Local Binary Pattern, and, finally, the classification of defects using the Support Vector Machine method. In the proposed method, an entropy filter was first applied to the input. The parts of the image with different patterns were then selected as candidate regions and these regions were classified by the patch classifier. All the defects were detected and classified and, finally, the efficiency of the algorithm was evaluated. By applying this algorithm to the dataset the best performance was obtained by the LBP descriptor and the linear SVM classifier with 98\% defect location accuracy and 97\% defect detection accuracy were achieved. In order to analyze the performance, used the deep model as a classifier, thus demonstrating that the deep model has a high capability for learning complex patterns. This proposed method is sensitive to local texture and could well describe texture information, which is appropriate for most kinds of tire defects.

Local and global based methods are two main trends for face recognition. Local approaches extract salient features by processing different parts of the image whereas global approaches find a general template for face of each person. Unfortunately, most global approaches work under controlled environments and they are sensitive to changes in the illumination. On the other hand, local approaches are more robust but finding their optimal parameters is a challenging task. This work proposes a new local-based approach that automatically tunes its parameters. The proposed method incorporates different techniques. In the first step, convolutional neural network (CNN) is employed as a trainable feature extraction procedure. In the second step, different metaheuristic methods are merged with CNN so that its best structure is found automatically. Finally, in the last step the decision is made by employing proper multi-class support vector machine (SVM). In this fashion a fully automated system is developed that is self-tuned and do not need manual adjustments. Simulation results demonstrate efficacy of the proposed method.

In this paper, a novel method for video content summarization has been proposed by calculating the fractal dimension of frames. Summarization of the video is the first step in automatic video analysis. In this paper, we use the support vector machine (SVM) and the decision - tree to identify the shot boundary and classify them. In order to compute the fractal dimension, the numerical method has been expressed. The results of this implementation were also reported on TRECVID 2006 data collection. The results show that the relative advantage of the method presented in this paper is compared to other articles.

Fuzzification of support vector machine has been utilized to deal with outlier and noise problem. This importance is achieved, by the means of fuzzy membership function, which is generally built based on the distance of the points to the class centroid. The focus of this research is twofold. Firstly, by taking the advantage of robust statistics in the fuzzy SVM, more emphasis on reducing the impact of outliers on the generalizability of SVM has been placed. Moreover, the variety of membership function for the elliptical data has been designated, based on the classic and robust Mahalanobis distance. Minimum covariance determinant and orthogonalised Gnanadesikan Kettenring estimators are employed in the structure of the robust--fuzzy SVM.By implementing the new membership function, the disadvantages of the traditional fuzzy membership function has been rectified. Simulated and real benchmarking data set confirm the effectiveness of the proposed methods. Compared with the traditional SVM and fuzzy SVM, these methods give a better performance on reducing the effects of outliers and significantly improves the classification accuracy and generalization.

Support Vector Regression (SVR) solves regression problems based on the concept of Support Vector Machine (SVM). In this paper, a new model of SVR with probabilistic constraints is proposed that any of output data and bias are considered the random variables with uniform probability functions. Using the new proposed method, the optimal hyperplane regression can be obtained by solving a quadratic optimization problem. The proposed method is illustrated by several simulated data and real data sets for both models (linearand nonlinear) with probabilistic constraints.