Integrating Support Vector Machine Research Articles

EnsemPseU: Identifying Pseudouridine Sites With an Ensemble Approach

Pseudouridine (Ψ) is the most prevalent RNA modification, which is formed from uridine through an isomerization reaction. With the increasing availability of genomic and proteomic samples, computer-aided pseudouridine-synthase-specific Ψ site recognition is becoming possible. In this paper, we propose an ensemble approach to identify pseudouridine sites, named EnsemPseU. First, five sequence-encoding strategies, namely, kmer, binary encoding, enhanced nucleic acid composition (ENAC), nucleotide chemical property (NCP), and nucleotide density (ND), were applied to extract sequence information. Then, chi-square feature selection was used to reduce the feature dimensionality and remove redundant information. Finally, an ensemble algorithm integrating support vector machine (SVM), extreme gradient boosting (XGBoost), naive Bayes (NB), k-nearest neighbor (KNN), and random forest (RF) was used to build our prediction model. Upon testing, the results showed that the accuracy improved 5.3% for H. sapiens, 6.09% for S. cerevisiae, and 5.55% for M. musculus after chi-square feature selection. Moreover, upon evaluation via 10-fold cross-validation and an independent test, our proposed model EnsemPseU outperformed the other best existing model. The source code and data sets are available at https://github.com/biyue1026/EnsemPseU.

A Depression Recognition Method for College Students Using Deep Integrated Support Vector Algorithm

The infinite increase in population, the pressure of survival, and the pressure of learning make the competition between people more and more fierce. Some college students have also been in a state of anxiety and panic for a long time, and mental health diseases have shown an explosive growth trend. The development of social networks such as Weibo, QQ, and WeChat not only provides more convenient communication methods for college students, but also provides a new emotional vent window for college students. They can record their living conditions in real time through social networks and interact with friends to express emotions and relieve stress. At the same time, the development of social networks has also provided a new way for the detection of depressed users. The current computer technology analyzes the user's social network data to detect the user's depression. This study uses text-level mining of Sina Weibo data from college students to detect depression among college students. First, collect text information of college student users in Sina Weibo, and construct the text information into input data that can be used for machine learning. Deep neural networks are used for feature extraction. An deep integrated support vector machine(DISVM) algorithm is introduced to classify the input data, and finally realize the recognition of depression. DISVM makes the recognition model more stable and improves the accuracy of depression diagnosis to a certain extent. Simulation experiments verify that the proposed depression recognition scheme can detect potential depression patients in the college student population through Sina Weibo data.

Predicting permeability changes with injecting CO2 in coal seams during CO2 geological sequestration: A comparative study among six SVM-based hybrid models

CO2 geological sequestration and enhanced coal bed methane extraction is a significant CO2 utilization approach with dual-meaning of energy and environment, and coal permeability is considered as one of the critical parameters for evaluating this method. To better predict permeability changes with injecting CO2 in coal seams, six SVM-based hybrid models integrating support vector machine (SVM) with intelligent optimization algorithms are proposed and compared, SVM is used for the relationship modelling between CO2 permeability and its influencing variables, and six intelligent optimization algorithms, including artificial bee colony (ABC), cuckoo search (CS), particle swarm optimization (PSO), differential evolution (DE), gray wolf optimizer (GWO), DE-GWO, are used for the hyper-parameters tuning. A total of 125 data samples for CO2 permeability are retrieved from the reported studies to train and verify the proposed models. The input variables for the predictive models include CO2 injection pressure, effective stress, temperature, buried depth and coal rank, and the corresponding output variable is CO2 permeability. The predictive model performance is evaluated and compared by correlation coefficient (R), root mean square error (RMSE) and mean absolute error (MAE). The predictive results denote that the prediction performance of the six hybrid models from high to low is DEGWO-SVM, GWO-SVM, PSO-SVM, CS-SVM, DE-SVM, ABC-SVM, and the DEGWO-SVM hybrid model is recommended to predict permeability changes with injecting CO2 in coal seams. At the same time, the mean impact value (MIV) is used to investigate the relative importance of each input variable. The relative importance scores of CO2 injection pressure, effective stress, temperature, buried depth and coal rank are 0.0248, 0.4617, 0.0211, 0.1102, and 0.3822, respectively. The research results have important guiding significance for CO2 permeability prediction and CO2 sequestration in coal seams.

Tool Wear Estimation in End Milling of Titanium Alloy Using NPE and a Novel WOA-SVM Model

Accurate identification of the tool wear states in machining of titanium alloys continues to be a thorny problem. Rapid construction of accurate and effective tool wear predictive models with regard to various titanium alloys is indispensable for promoting the development of intelligent manufacturing. This article presents a novel WOA-SVM model that integrates support vector machine (SVM) and whale optimization algorithm (WOA), which is utilized for accurate estimation of the tool wear in end milling of titanium alloy Ti-6Al-4V under variable cutting conditions. The signal features in three domains are extracted from the original cutting forces and constitute the “complete features” that are adopted as the monitoring features. Besides, neighborhood preserving embedding (NPE) is utilized to fuse the monitoring features and realize dimension reduction, which aims at improving the modeling efficiency of the novel WOA-SVM model. The experimental results show that under the premise of guaranteeing prediction accuracy, the utilization of NPE reduces the modeling time-consumption of WOA-SVM by more than 90%. In comparison with the commonly used methods (such as PSO-SVM and GSA-SVM), WOA-SVM takes on comparable prediction accuracy and can reduce the modeling time-consumption by more than 30% in most cases. Moreover, the WOA-SVM model has better prediction accuracy and stability than other classical methods, such as k-nearest neighbor (k-NN), feedforward neural network (FFNN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and classification and regression tree (CART). Therefore, the combination of NPE and WOA-SVM provides a guarantee for rapidly constructing accurate tool wear predictive models in machining of titanium alloys.

Prediction of S-nitrosylation sites by integrating support vector machines and random forest.

Cysteine S-nitrosylation is a type of reversible post-translational modification of proteins, which controls diverse biological processes. It is associated with redox-based cellular signaling to protect against oxidative stress. The identification of S-nitrosylation sites is an important step to reveal the function of proteins; however, experimental identification of S-nitrosylation is expensive and time-consuming work. Hence, sequence-based computational prediction of potential S-nitrosylation sites is highly sought before experimentation. Herein, a novel predictor PreSNO has been developed that integrates multiple encoding schemes by the support vector machine and random forest algorithms. The PreSNO achieved an accuracy and Matthews correlation coefficient value of 0.752 and 0.252 respectively in classifying between SNO and non-SNO sites when evaluated on the independent dataset, outperforming the existing methods. The web application of the PreSNO and its associated datasets are freely available at http://kurata14.bio.kyutech.ac.jp/PreSNO/.

Performance-based fault diagnosis of a gas turbine engine using an integrated support vector machine and artificial neural network method

An effective and reliable gas path diagnostic method that could be used to detect, isolate, and identify gas turbine degradations is crucial in a gas turbine condition-based maintenance. In this paper, we proposed a new combined technique of artificial neural network and support vector machine for a two-shaft industrial gas turbine engine gas path diagnostics. To this end, an autoassociative neural network is used as a preprocessor to minimize noise and generate necessary features, a nested support vector machine to classify gas path faults, and a multilayer perceptron to assess the magnitude of the faults. The necessary data to train and test the method are obtained from a performance model of the case engine under steady-state operating conditions. The test results indicate that the proposed method can diagnose both single- and multiple-component faults successfully and shows a clear advantage over some other methods in terms of multiple fault diagnosis. Moreover, 5-8 sets of measurements have been used to assess the prediction accuracy, and only a 2.3% difference was observed. This result indicates that the proposed method can be used for multiple fault diagnosis of gas turbines with limited measurements.

Audible Sound-Based Intelligent Evaluation for Aluminum Alloy in Robotic Pulsed GTAW: Mechanism, Feature Selection, and Defect Detection

Aluminum alloy is the main structure material in aerospace industry. Online defect detection for aluminum alloy in pulsed gas tungsten arc welding (GTAW) is still challenging, especially for increasing application of robotics. This paper presents an intelligent methodology for real-time evaluation of weld penetration defects based on arc audible sound sensing for aluminum alloy in robotic-pulsed GTAW. The generation mechanism of arc sound was investigated using correlation analysis, high-speed camera observing and frequency spectrum analysis before denoising of arc sound. Two feature selection approaches based on Fisher distance and principal component analysis (PCA) were developed to select the frequency components related to seam defects, and then, their performance were qualitatively and quantitatively analyzed. Finally, a new classification model integrating support vector machine with grid search optimization and cross-validation (SVM-GSCV) was established to identify underpenetration, normal penetration, and burning through. The proposed methodologies were verified to be effective with high accuracy and robustness. This paper can provide some guidance for condition monitoring of additive manufacturing (AM) or process industry.

Integrating support vector machine and graph cuts for medical image segmentation

Medical image segmentation remains a challenged problem because of intensity inhomogeneity and surrounding complex background. In this paper, we propose a novel method for medical image segmentation by integrating support vector machine and graph cuts. Particularly, a novel localized training scheme is proposed to train a classifier for each pixel based on the target image information, and then a novel graph cuts-based segmentation method that combines the constraint information of machine learning result, the edge information, the local information, and the remote-local information is proposed for post-processing. Instead of delineating an initialized curve around the object boundary, we directly draw a narrowband mask for the initialization in the paper. Experiments on synthetic and medical images demonstrate that the proposed method can achieve better performance than the state-of-the-art.

Credit card fraud forecasting model based on clustering analysis and integrated support vector machine

At present, with the popularization of credit cards, credit card fraud increases gradually. Based on this, this paper designs a credit card fraud prediction model based on cluster analysis and integrated support vector machine using computer technology. First of all, adjust and reduce the Unbalanced state based on K-means clustering analysis combined with more than half of the random samples. Secondly, the use of the idea of integrated learning to further deal with the Unbalanced state of the data and increase classifier’s awareness of minorities. Finally, we tested the algorithm, and the result showed that the proposed algorithm effectively reduced the cost of accidental injury, which provides a great possibility for the card issuer to effectively reduce the economic losses caused by credit card fraud, which has laid a good theoretical basis and foundation for practical application.

An Analysis of Quality Monitoring and Control System using Real- Time, Integrated Cost Effectiveness and Support Vector Machine

The quality monitoring and control (QMC) has been an essential process in the manufacturing industries to ensure product reliability. With the advancements in big-data analytics, machine-learning based QMC has become more and more popular in various manufacturing industries, such as automotive and electronic companies. At the same time, the cost effectiveness (CE) of the QMC is perceived as a main decision criterion that explicitly accounts for inspection efforts and has a direct relationship with the QMC capability. In this paper, the integrated support vector machine (SVM)-based automated QMC system with the adjusted CE, CEadj, model is proposed. Unlike existing models, the proposed model explicitly incorporates inspection-related expenses (i.e., warranty cost, rework cost, inspection cost) and error types (i.e., type-I and -II errors) in the CEadj framework to guide the SVM algorithm. The proposed automated QMC system is verified and validated using a door-trim manufacturing process of an automotive industry. Next, a designed experiment is performed to assess the sensitivity analysis of the proposed framework. The proposed model is found to be effective and could be exploited as an alternative or complementary tool for the traditional quality inspection system.

Financial distress prediction: The case of French small and medium-sized firms

Financial distress prediction is a central issue in empirical finance that has drawn a lot of research interests in the literature. This paper aims to predict the financial distress of French small and medium firms using Logit model, Artificial Neural Networks, Support Vector Machine techniques, Partial Least Squares, and a hybrid model integrating Support Vector Machine with Partial Least Squares. Empirical results indicate that for one year prior to financial distress, Support Vector Machine is the best classifier with an overall accuracy of 88.57%. Meanwhile, in the case of two years prior to financial distress, the hybrid model outperforms Support Vector Machine, Logit model, Partial Least Squares, andArtificial Neural Networks with an overall accuracy of 94.28%. Distressed firms are found to be smaller, more leveraged and with lower repayment capacity. Moreover, they have lower liquidity, profitability, and solvency ratios. Besides the academic research contribution, our findings can be useful for managers, investors, and creditors. With respect to managers, our findings provide them with early warnings signals of performance deterioration in order to take corrective actions and reduce the financial distress risk. For investors, understanding the main factors leading to financial distress allows them to avoid investing in risky firms. Creditors should correctly evaluate the firm financial situation and be vigilant to signs of impending financial distress to avoid capital loss and costs related to counterpart risk.

Prediction of the product formation with the AdaBoost algorithm in bioprocesses

The state variables such as product formation could provide important information for the optimisation of fermentation processes. Since the kinetic modelling is difficult for bioprocesses, the product formation is predicted by integrating support vector machine (SVM) with the AdaBoost algorithm. The AdaBoost algorithm is used for adaptively boosting the performance of SVM weak learners. The prediction approach is tested by using 2-keto-L-gulonic acid (2-KGA) cultivation as an example. The validation results using the data from industrial 2-KGA cultivation demonstrate that the prediction approach has good generalisation performance and noise tolerance.

Prediction of the product formation with the AdaBoost algorithm in bioprocesses

The state variables such as product formation could provide important information for the optimisation of fermentation processes. Since the kinetic modelling is difficult for bioprocesses, the product formation is predicted by integrating support vector machine (SVM) with the AdaBoost algorithm. The AdaBoost algorithm is used for adaptively boosting the performance of SVM weak learners. The prediction approach is tested by using 2-keto-L-gulonic acid (2-KGA) cultivation as an example. The validation results using the data from industrial 2-KGA cultivation demonstrate that the prediction approach has good generalisation performance and noise tolerance.

Dengue Spread Modeling in the Absence of Sufficient Epidemiological Parameters: Comparison of SARIMA and SVM Time Series Models

Dengue remains to be a major public health concern in the Philippines, claiming hundreds of lives every year ( Jaymalin 2017). Given limited data for deriving necessary epidemiological parameters in developing deterministic disease models, forecasting as a means in controlling and anticipating outbreaks remains a challenge. In this study, two-time series models, namely Seasonal Autoregressive Integrated Moving Average (SARIMA) and Support Vector Machine (SVM), were developed without the requirement for prior epidemiological parameters. Performances of the models in predicting dengue incidences in the Western Visayas Region of the Philippines were compared by measuring the Root Mean Square Error (RMSE) and Mean Average Error (MAE). Results showed that the models were both effective in forecasting Dengue incidences for epidemiological surveillance as validated by historical data. SARIMA model yielded average RMSE and MAE scores of 16.8187 and 11.4640, respectively. Meanwhile, SVM model achieved scores of 11.8723 and 7.7369, respectively. With the data and setup used, this study showed that SVM outperformed SARIMA in forecasting Dengue incidences. Furthermore, preliminary investigation of one-month lagged climate variables using Random Forest Regressor’s feature ranking yielded rain intensity and value as top possible dengue incidence climate predictors.

Improving the Long Lead-Time Inundation Forecasts Using Effective Typhoon Characteristics

In this paper, a new type of inundation forecasting model with the effective typhoon characteristics is proposed by integrating support vector machine (SVM) with multi-objective genetic algorithm (MOGA). Firstly, a comparison of the proposed model and an existing model based on back-propagation network (BPN) is made to highlight the improvement in forecasting performance. Next, the proposed model is compared with the SVM-based model without typhoon characteristics to investigate the influence of typhoon characteristics on inundation forecasting. Effective typhoon characteristics for improving forecasting performance are identified as well. An application to Chiayi City, Taiwan, is conducted to demonstrate the superiority of the proposed model. The results confirm that the proposed model with the effective typhoon characteristics does improve the forecasting performance and the improvement increases with increasing lead-time, especially for long lead-time forecasting. The proposed model is capable of optimizing the input to decrease the negative impact when increasing forecast lead time. In conclusion, effective typhoon characteristics are recommended as key inputs for inundation forecasting during typhoons.

Predicting protein–protein interaction sites using modified support vector machine

Protein–protein interaction plays a fundamental role in many biological processes and diseases. Characterizing protein interaction sites is crucial for the understanding of the mechanism of protein–protein interaction and their cellular functions. In this paper, we proposed a method based on integrated support vector machine (SVM) with a hybrid kernel to predict protein interaction sites. First, a number of features of the protein interaction sites were extracted. Secondly, the technique of sliding window was used to construct a protein feature space based on the influence of the adjacent residues. Thirdly, to avoid the impact of imbalance of the data set on prediction accuracy, we employed boost-strap to re-sample the data. Finally, we built a SVM classifier, whose hybrid kernel comprised a Gaussian kernel and a Polynomial kernel. In addition, an improved particle swarm optimization (PSO) algorithm was applied to optimize the SVM parameters. Experimental results show that the PSO-optimized SVM classifier outperforms existing methods.

RETRACTED ARTICLE: Horizontal global solar radiation estimation using hybrid SVM-firefly and SVM-wavelet algorithms: a case study

Nowadays, investigations are being performed to identify proper approaches for prediction of solar radiation in the absence of measured data. In this context, application of hybrid techniques has attracted many attentions since it offers some advantages by utilizing the specific nature of each technique to achieve more preciseness. This study aims at assessing the suitability of two hybrid approaches to predict monthly mean daily horizontal global solar radiation. For this aim, two hybrid approaches by integrating support vector machine (SVM) with firefly algorithm (FFA) and wavelet transform (WT) algorithm named, respectively, SVM-FFA and SVM-WT are developed. Then two different SVM-FFA and SVM-WT models are established by considering different meteorological parameters: (1) relative sunshine duration and (2) relative sunshine duration, air temperature difference, average air temperature and relative humidity. The results indicate that hybridizing the SVM with FFA and WT algorithms would be promising as both SVM-FFA and SVM-WT approaches show higher performance than single SVM. Also, model (2) of each hybrid approach provides more accuracy. Furthermore, the statistical results reveal the superiority of SVM-WT over SVM-FFA in terms of predictions accuracy.

A Parallel Digital VLSI Architecture for Integrated Support Vector Machine Training and Classification

This paper presents a parallel digital VLSI architecture for combined support vector machine (SVM) training and classification. For the first time, cascade SVM, a powerful training algorithm, is leveraged to significantly improve the scalability of hardware-based SVM training and develop an efficient parallel VLSI architecture. The presented architecture achieves excellent scalability by spreading the training workload of a given data set over multiple SVM processing units with minimal communication overhead. Hardware-friendly implementation of the cascade algorithm is employed to achieve low hardware overhead and allow for training over data sets of variable size. In the proposed parallel cascade architecture, a multilayer system bus and multiple distributed memories are used to fully exploit parallelism. In addition, the proposed architecture is rather flexible and can be tailored to realize hybrid use of hardware parallel processing and temporal reuse of processing resources, leading to good tradeoffs between throughput, silicon overhead and power dissipation. Several parallel cascade SVM processors have been designed with a commercial 90-nm CMOS technology, which provide up to a $561\times $ training time speedup and a significant estimated $21\,859\times $ energy reduction compared with the software SVM algorithm running on a 45-nm commercial general-purpose CPU.

Flood susceptibility analysis and its verification using a novel ensemble support vector machine and frequency ratio method

Flood is one of the most commonly occurred natural hazards worldwide. Severe flood occurrences in Kelantan, Malaysia cause damage to both life and property every year. Due to the huge losses in this area, development of appropriate flood modeling is required for the government. Remote sensing and geographic information system techniques can support overall flood management as they can produce rapid data collection and analysis for hydrological studies. The existing models for flood mapping have some weak points that may improve through more sophisticated and ensemble methods. The current research aimed to propose a novel ensemble method by integrating support vector machine (SVM) and frequency ratio (FR) to produce spatial modeling in flood susceptibility assessment. In the literature, mostly statistical and machine learning methods are used individually; however, their integration can enhance the final output. The FR model can perform bivariate statistical analysis and evaluate the correlation between the flooding and classes of each conditioning factors. The weights achieved by FR can be assigned to each conditioning factor and the resulted factors can be used in SVM analysis. In order to examine the efficiency of the proposed ensemble method and to show the proficiency of SVM, another machine learning algorithm such as decision tree (DT) was applied and the results were compared. To perform the methods, the upper catchment of the Kelantan basin in Malaysia was chosen. First, a flood inventory map with a total of 155 flood locations were extracted from various sources over the study area. The flood inventory map was randomly divided into two dataset; 70 % (115 flood locations) for the purpose of training and the remaining 30 % (40 flood locations) was used for validation. The spatial database included digital elevation model, curvature, geology, river, stream power index, rainfall, land use/cover, soil type, topographic wetness index and slope. For model validation, area under curve method was used and both success and prediction rate curves were calculated. The validation results for ensemble method showed 88.71 and 85.21 % for success rate and prediction rate respectively. The DT model showed 87.00 and 82.00 % for the success rate and prediction rate respectively. It is evident that the accuracies were increased using the ensemble method. The acquired results proved the efficiency of the proposed ensemble method as rapid, accurate and reasonable in flood susceptibility assessment.

An EPC Forecasting Method for Stock Index Based on Integrating Empirical Mode Decomposition, SVM and Cuckoo Search Algorithm

AbstractIn order to improve the forecasting accuracy, a hybrid error-correction approach by integrating support vector machine (SVM), empirical mode decomposition (EMD) and the improved cuckoo search algorithm (ICS) was introduced in this study. By using two indexes as examples, the empirical study shows our proposed approach by means of synchronously predict the prediction error which used to correct the preliminary predicted values has better prediction precision than other five competing approaches, furthermore, the improved strategies for cuckoo search algorithm has better performance than other three evolutionary algorithms in parameters selection.