Least Squares Support Vector Machine Classifier Research Articles

Pipeline Leak Detection: A Comprehensive Deep Learning Model Using CWT Image Analysis and an Optimized DBN-GA-LSSVM Framework.

Detecting pipeline leaks is an essential factor in maintaining the integrity of fluid transport systems. This paper introduces an advanced deep learning framework that uses continuous wavelet transform (CWT) images for precise detection of such leaks. Transforming acoustic signals from pipelines under various conditions into CWT scalograms, followed by signal processing by non-local means and adaptive histogram equalization, results in new enhanced leak-induced scalograms (ELIS) that capture detailed energy fluctuations across time-frequency scales. The fundamental approach takes advantage of a deep belief network (DBN) fine-tuned with a genetic algorithm (GA) and unified with a least squares support vector machine (LSSVM) to improve feature extraction and classification accuracy. The DBN-GA framework precisely extracts informative features, while the LSSVM classifier precisely distinguishes between leaky and non-leak conditions. By concentrating solely on the advanced capabilities of ELIS processed through an optimized DBN-GA-LSSVM model, this research achieves high detection accuracy and reliability, making a significant contribution to pipeline monitoring and maintenance. This innovative approach to capturing complex signal patterns can be applied to real-time leak detection and critical infrastructure safety in several industrial applications.

A NEW ROLLER BEARING FAULT DIAGNOSIS METHOD BASED ON VMD ENERGY ENTROPY AND BSOA-LSSVM

This paper presents a new method for roller bearing fault diagnosis based on least square support vector machine (LSSVM) with parameters optimized by Backtracking Search Optimization Algorithm (BSOA), namely BSOA-LSSVM. First, roller bearing acceleration vibration signals are decomposed into functions by using Variational Mode Decomposition (VMD) method. Second, initial feature matrices are extracted from those functions by energy entropy to obtain feature matrix. Third, these values serve as input vector for BSOA - LSSVM classifier. Experimental results showed that the proposed method gave the higher classification accuracy (100%) and shorter computational time than other method.

Photoplethysmography signal-based automated diagnosis of type-2 diabetes using tunable-Q wavelet transform and least-square support vector machine classifier

Photoplethysmography signal-based automated diagnosis of type-2 diabetes using tunable-Q wavelet transform and least-square support vector machine classifier

A proposed model for diabetes mellitus classification using coyote optimization algorithm and least squares support vector machine

One of the most dangerous health diseases affecting the world's population is diabetes mellitus (DM), and its diagnosis is the key to its treatment. Several methods have been implemented to diagnose diabetes patients. In this work, a hybrid model which combines of coyote optimization algorithm (COA) and least squares support vector machine (LS-SVM) is proposed to classify of Type-II-DM patients. LS-SVM classifier is applied for classification process but it's very sensitive when its parameter values are changed. To overcome this problem, COA algorithm is implemented to optimize parameters of the LS-SVM classifier. This is the goal of the proposed model called the COA-LS-SVM. The proposed model is implemented and evaluated using pima Indians diabetes dataset (PIDD). Also, it's compared with several classification algorithms that were implemented on the same PIDD. The experimental results demonstrated the effectiveness of the proposed model and its superiority over other algorithms, as it could accomplish an average classification accuracy of 98.811%.

Wireless IoT and Cyber-Physical System for Health Monitoring Using Honey Badger Optimized Least-Squares Support-Vector Machine.

Health monitoring is a prominent factor in a person's daily life. Healthcare for the elderly is becoming increasingly important as the population ages and grows. The health of an Elderly patient needs frequent examination because the health deteriorates with an increasing age profile. IoT is utilized everywhere in the health industry to identify and communicate with the patients by the professional. A cyber-physical system (CPS) is used to combine physical processes with communication and computation. CPS and IoT are both wirelessly connected via information and communication technologies. The novelty of the research lies in the Honey Badger (HB) algorithm optimized Least-squares Support-Vector Machine (LS-SVM) architecture proposed in this paper for monitoring multi parameters to categorize and determine the abnormal patient details present in the dataset. Since the performance of the LS-SVM is highly dependent on the width coefficient and regularization factor, the HB algorithm is employed in this study to optimize both parameters. The HB algorithm is capable of solving the medical problem that has a complex search space and it also improves the convergence performance of the LS-SVM classifier by achieving a tradeoff between the exploration and exploitation phases. The HB optimized LS-SVM classifier predicts the patients with deteriorating health conditions and evaluates the accuracy of the results obtained. In the end, the statistical data is provided to the caretaker via a smartphone application as a monthly statistical report. The proposed model offers a Positive Predictive Value (PPV), Negative Predictive Value (NPV), and an Area Under the Curve (AUC) score of 0.9478, 0.9587, and 0.9617 respectively which is relatively higher than the conventional techniques such as Decision tree, Random Forest, and Support Vector Machine (SVM) classifier. The simulation results demonstrate that the proposed model efficiently models the sensor parameters and offers timely support to elderly patients.

Extracting epileptic features in EEGs using a dual-tree complex wavelet transform coupled with a classification algorithm

The detection of epileptic seizures from electroencephalogram (EEG) signals is traditionally performed by clinical experts through visual inspection. It is a long process, is error prone, and requires a highly trained expert. In this research, a new method is presented for seizure classification for EEG signals using a dual-tree complex wavelet transform (DT-CWT) and fast Fourier transform (FFT) coupled with a least square support vector machine (LS-SVM) classifier. In this method, each EEG signal is divided into four segments. Each segment is further split into smaller sub-segments. The DT-CWT is applied to decompose each sub-segment into detailed and approximation coefficients (real and imaginary parts). The obtained coefficients by the DT-CWT at each decomposition level are passed through an FFT to identify the relevant frequency bands. Finally, a set of effective features are extracted from the sub-segments, and are then forwarded to the LS-SVM classifier to classify epileptic EEGs. In this paper, two epileptic EEG databases from Bonn and Bern Universities are used to evaluate the extracted features using the proposed method. The experimental results demonstrate that the method obtained an average accuracy of 97.7% and 96.8% for the Bonn and Bern databases, respectively. The results prove that the proposed DT-CWT and FFT based features extraction is an effective way to extract discriminative information from brain signals. The obtained results are also compared to those by k-means and Naïve Bayes classifiers as well as with the results from the previous methods reported for classifying epileptic seizures and identifying the focal and non-focal EEG signals. The obtained results show that the proposed method outperforms the others and it is effective in detecting epileptic seziures in EEG signals. The technique can be adopted to aid neurologists to better diagnose neurological disorders and for an early seizure warning system.

Surrogate-Assisted Differential Evolution Using Knowledge-Transfer-Based Sampling for Expensive Optimization Problems

To sufficiently reuse the knowledge from previous optimization efforts, a surrogate-assisted differential evolution using knowledge-transfer-based sampling (denoted as SADE-KTS) method is proposed for solving expensive black-box optimization problems. In SADE-KTS, a novel knowledge-transfer-based sampling method is integrated with the differential evolution framework to generate promising initial sample points. In this way, a least-squares support vector machine classifier is constructed based on the prior optimization knowledge database to calibrate the initial sample points adaptively, which improves the exploration performance via transferring the existed optimization efforts to the current optimization task. Moreover, the radial basis function and kriging surrogates are employed to replace the expensive simulation models for evolutionary operations, where the tailored differential evolution operators are cooperated with the sequential quadratic programming optimizer to lead the search to the global optimum efficiently. A number of numerical benchmarks are tested to illustrate the optimization capacity of SADE-KTS compared with several competitive optimization algorithms. Finally, SADE-KTS is applied to an airfoil aerodynamic knowledge-based optimization problem considering the existed optimization knowledge, which demonstrates the practicality and effectiveness of the proposed SADE-KTS in engineering practices.

Incipient Chiller Fault Diagnosis Using an Optimized Least Squares Support Vector Machine With Gravitational Search Algorithm

Operational faults in centrifugal chillers will lead to high energy consumption, poor indoor thermal comfort, and low operational safety, and thus it is of significance to detect and diagnose the anomalies timely and effectively, especially for those at their incipient stages. The least squares support vector machine (LSSVM) has been regarded as an effective algorithm for multiclass classification. One of the most difficult issues in LSSVM is parameter tuning. Therefore, this paper reports a development of a gravitational search algorithm (GSA) optimized LSSVM method for incipient fault diagnosis in centrifugal chillers. Considering the inadequacies of conventional principle component analysis (PCA) algorithm for nonlinear data transformation, kernel principle component analysis (KPCA) was firstly employed to reduce the dimensionality of the original input data. Secondly, an optimized “one against one” multi-class LSSVM classifier was developed and its penalty constant and kernel bandwidth were tuned by GSA. Based on the fault samples of seven typical faults at their incipient stages in chillers from ASHRAE RP 1043, the proposed GSA optimized LSSVM fault diagnostic model was trained and validated. For the purpose of demonstrating the priority of the proposed fault diagnosis method, the obtained results were compared to that of using the LSSVM classifier optimized by another two algorithms, namely, the conventional cross-validation method and particle swarm optimizer. Results showed that the best fault diagnosis performance could be achieved using the proposed GSA-LSSVM classifier. The overall average fault diagnosis accuracy for the least severity faults was reported over 95%.

Inspection method of combine assembly quality based on optimized VMD

Aiming at the problems of low assembly accuracy and difficult to detect assembly quality of combine, a method of combine assembly quality detection based on sparrow search algorithm (SSA) optimized variational mode decomposition (VMD) and particle swarm optimization (PSO) optimized least squares support vector machine (LSSVM) was proposed, Firstly, the sparrow search algorithm is used to obtain the optimal VMD decomposition modal parameter K and penalty factor α, then the combined vibration signal of combine harvester is decomposed into intrinsic modal components of different center frequencies by using the best parameter combination [K, α]. Finally, the feature vector is used as the input of LSSVM classifier to classify different fault features. The analysis results show that the classification accuracy of SSA-VMD joint feature extraction method is 99.5%, which is 17.5% and 9.5% higher than ensemble empirical mode decomposition (EEMD) and fixed parameter VMD, which verifies the superiority of this method in the detection of combine assembly quality.

Fault diagnosis of the constant current remote power supply system in CUINs based on the improved water cycle algorithm

Cabled Underwater Information Networks (CUINs) is an important platform for ocean observation where the Constant Current Remote Power Supply System (CCRPSS) guarantees the safe and normal operation of CUINs. The CCRPSS is mainly composed of the main node and underwater cables which require frequent fault diagnosis but how to improve the fault diagnosis rate is a difficult problem. This paper proposed a fault diagnosis method for the CCRPSS based on the Improved Water Cycle Algorithm (IWCA) and the multi-classifier group based on the Least Squares Support Vector Machine (LSSVM). Firstly, the multi-feature extraction method is used to obtain the characteristic information in the time and frequency domain; secondly, IWCA is established by combining the traditional Water Cycle Algorithm (WCA) with the chaotic mutation strategy, the elite memory strategy and the population reconstruction strategy. By applying 13 typical test functions to performance test, it can be found that the IWCA can effectively improve the global search ability and balance of the WCA algorithm. At last, IWCA is used to optimize the parameters of the LSSVM classifier and improve the classification efficiency. The comparisons of simulated results with traditional methods show that the proposed diagnostic model can not only obtain complete fault feature information, but also obtain the optimal classification parameters of LSSVM faster. Therefore, the proposed diagnostic method is verified to be suitable for the fault diagnosis of the constant current remote power supply system in CUINs.

Recognition of Emotion in Indian Classical Dance Using EMG Signal

Indian classical dance forms like Kathak are an enrichment of Indian culture and tradition. These dance forms glorify its beauty by expressing nine emotions (Navras) such as Adbhut (amazed), Bhayanaka (fearful), Hasya (humorous), Karuna (tragic), Raudra (fierce), Shringar (loving smile), Veer (heroic), Bibhatsa (disgusted), and Shant (peaceful). Identifying correct emotions is an important task. The objective of this research work is to recognize Navras in the Kathak dance. Proposed research work can assist dance teachers in an accurate and unbiased evaluation process of dance examination. This research work analyzed the Electromyogram (EMG) signals acquired from eleven subjects. The EMG signals collected from the various locations on the face and neck represent the emotions and head movement. The EMG signals are processed to extract integrated EMG (IEMG) features. This research introduced a new feature named 'difference in IEMG feature' for improving the accuracy of emotion recognition. For the classification of nine emotions, the Least Square Support Vector Machine (LSSVM), Nonlinear Autoregressive Exogenous Network (NARX), and Long- and Short-Term Memory (LSTM) classifiers were used. The classifiers' performance is judged with head motion and without head motion. The classification accuracies are calculated using a maximum, variance, and mean of the feature. LSSVM, NARX, and LSTM classifiers achieved 60.80%, 81.67%, and 92.28% classification accuracies, respectively, using the IEMG feature and head motion. Using the new feature, LSSVM, NARX, and LSTM classifiers achieved 64.29%, 81.27%, and 93.63% classification accuracies, respectively. The overall classification accuracy improved by 1.46% by using the new feature.

Metaheuristic Optimized Multi-Level Classification Learning System for Engineering Management

Multi-class classification is one of the major challenges in machine learning and an ongoing research issue. Classification algorithms are generally binary, but they must be extended to multi-class problems for real-world application. Multi-class classification is more complex than binary classification. In binary classification, only the decision boundaries of one class are to be known, whereas in multiclass classification, several boundaries are involved. The objective of this investigation is to propose a metaheuristic, optimized, multi-level classification learning system for forecasting in civil and construction engineering. The proposed system integrates the firefly algorithm (FA), metaheuristic intelligence, decomposition approaches, the one-against-one (OAO) method, and the least squares support vector machine (LSSVM). The enhanced FA automatically fine-tunes the hyperparameters of the LSSVM to construct an optimized LSSVM classification model. Ten benchmark functions are used to evaluate the performance of the enhanced optimization algorithm. Two binary-class datasets related to geotechnical engineering, concerning seismic bumps and soil liquefaction, are then used to clarify the application of the proposed system to binary problems. Further, this investigation uses multi-class cases in civil engineering and construction management to verify the effectiveness of the model in the diagnosis of faults in steel plates, quality of water in a reservoir, and determining urban land cover. The results reveal that the system predicts faults in steel plates with an accuracy of 91.085%, the quality of water in a reservoir with an accuracy of 93.650%, and urban land cover with an accuracy of 87.274%. To demonstrate the effectiveness of the proposed system, its predictive accuracy is compared with that of a non-optimized baseline model, single multi-class classification algorithms (sequential minimal optimization (SMO), the Multiclass Classifier, the Naïve Bayes, the library support vector machine (LibSVM) and logistic regression) and prior studies. The analytical results show that the proposed system is promising project analytics software to help decision makers solve multi-level classification problems in engineering applications.

Capacitively-coupled ECG and respiration for the unobtrusive detection of sleep apnea

Objective. The performance of a novel unobtrusive system based on capacitively-coupled electrocardiography (ccECG) combined with different respiratory measurements is evaluated for the detection of sleep apnea. Approach. A sleep apnea detection algorithm is proposed, which can be applied to electrocardiography (ECG) and ccECG, combined with different unobtrusive respiratory measurements, including ECG derived respiration (EDR), respiratory effort measured using the thoracic belt (TB) and capacitively-coupled bioimpedance (ccBioz). Several ECG, respiratory and cardiorespiratory features were defined, of which the most relevant ones were identified using a random forest based backwards wrapper. Using this relevant feature set, a least-squares support vector machine classifier was trained to decide if a one minute segment is apneic or not, based on the annotated polysomnography (PSG) data of 218 patients suspected of having sleep apnea. The obtained classifier was then tested on the PSG and capacitively-coupled data of 28 different patients. Main results. On the PSG data, an AUC of 76.3% was obtained when the ECG was combined with the EDR. Replacing the EDR with the TB led to an AUC of 80.0%. Using the ccECG and ccBioz or the ccECG and TB resulted in similar performances as on the PSG data, while using the ccECG and ccECG-based EDR resulted in a drop in AUC to 67.4%. Significance. This is the first study which tests an apnea detection algorithm on capacitively-coupled ECG and bioimpedance signals and shows promising results on the capacitively-coupled data set. However, it was shown that the EDR could not be accurately estimated from the ccECG signals. Further research into the effect that respiration has on the ccECG is needed to propose alternative EDR estimates.

Multi-Scale Sample Entropy-Based Energy Moment Features Applied to Fault Classification

Rolling element bearings are crucial, of high failure rate and easy damage parts of rotating machinery, and significantly affect safe and reliable production processes. Much more attention has been focused on fault diagnosis of rolling element bearings in recent years. This article presents a novel feature extraction scheme for the classification of multiple bearing faults. Multi-Scale Sample Entropy (M-SSampEn) is combined with Energy Moment (EM) to construct a time-domain Multi-Scale Sample Entropy-based Energy Moment (M-SSampEn-EM) feature extractor. The classifier model for the proposed fault classification system has been built using the Least Square Support Vector Machine (LS-SVM). The M-SSampEn-EM feature extractor is used to capture two-dimensional representative eigenvectors from multiple fault classes' bearing vibration data. Its separability performance is ensured by optimizing the feature-extraction parameters, including the Embedded Dimension and the Tolerance, etc. The LS-SVM classifier is also compared with two Neural Network (NN)-based classifiers, i.e., Radial Basis Function NN (RBFNN) and Probabilistic NN (PNN), to show it better generalization performance on bearing fault classification. The experimental study verifies the excellent capacity of the proposed approach in bearing fault classification.

Predictive diagnosis of chronic obstructive pulmonary disease using serum metabolic biomarkers and least-squares support vector machine.

ObjectiveDevelopment of biofluid‐based biomarkers is attractive for the diagnosis of chronic obstructive pulmonary disease (COPD) but still lacking. Thus, here we aimed to identify serum metabolic biomarkers for the diagnosis of COPD.MethodsIn this study, we investigated serum metabolic features between COPD patients (n = 54) and normal individuals (n = 74) using a 1H NMR‐based metabolomics approach and developed an integrated method of least‐squares support vector machine (LS‐SVM) and serum metabolic biomarkers to assist COPD diagnosis.ResultsWe observed a hypometabolic state in serum of COPD patients, as indicated by decreases in N‐acetyl‐glycoprotein (NAG), lipoprotein (LOP, mainly LDL/VLDL), polyunsaturated fatty acid (pUFA), glucose, alanine, leucine, histidine, valine, and lactate. Using an integrated method of multivariable and univariate analyses, NAG and LOP were identified as two important metabolites for distinguishing between COPD patients and controls. Subsequently, we developed a LS‐SVM classifier using these two markers and found that LS‐SVM classifiers with linear and polynomial kernels performed better than the classifier with RBF kernel. Linear and polynomial LS‐SVM classifiers can achieve the total accuracy rates of 80.77% and 84.62% and the AUC values of 0.87 and 0.90 for COPD diagnosis, respectively.ConclusionsThis study suggests that artificial intelligence integrated with serum metabolic biomarkers has a great potential for auxiliary diagnosis of COPD.

Unbalanced classification method using least squares support vector machine with sparse strategy for steel surface defects with label noise

Least squares support vector machine (LS-SVM) plays an important role in steel surface defects classification because of its high speed. However, the defect samples obtained from the real production line may be noise. LS-SVM suffers from the poor classification performance in the classification stage when there are noise samples. Thus, in the classification stage, it is necessary to design an effective algorithm to process the defects dataset obtained from the real production line. To this end, an adaptive weight function was employed to reduce the adverse effect of noise samples. Moreover, although LS-SVM offers fast speed, it still suffers from a high computational complexity if the number of training samples is large. The time for steel surface defects classification should be as short as possible. Therefore, a sparse strategy was adopted to prune the training samples. Finally, since the steel surface defects classification belongs to unbalanced data classification, LS-SVM algorithm is not applicable. Hence, the unbalanced data information was introduced to improve the classification performance. Comprehensively considering above-mentioned factors, an improved LS-SVM classification model was proposed, termed as ILS-SVM. Experimental results show that the new algorithm has the advantages of high speed and great anti-noise ability.

Automatic drowsiness detection using electroencephalogram signal

Drowsiness refers to the state of being sleepy or the state of minimal concentration. It is characterised by a decrease in a person's memory capacity and brain information processing speed. These conditions cause hazards in the real-time working environment such as driving, monitoring power generation and patient health etc. These hazards can be sidestepped by introducing the automatic drowsiness detection system. This Letter suggested the electroencephalogram (EEG)-based automatic drowsiness detection method. The clustering variational mode decomposition (CVMD) explores the non-stationary behaviour of EEG for drowsiness detection. In CVMD optimum allocation sampling analyses non-homogenous EEG signals and converts those into homogeneous EEG clusters. These clusters are then decomposed into band-limited modes. The oscillatory mode characteristics are extracted in terms of several features. These features are fed as input into the least-squares support vector machine classifier. The proposed method provides better drowsiness detection performance in comparison with the different methods using the same data set.

Fault Diagnosis of Nonlinear Analog Circuit Based on Generalized Frequency Response Function and LSSVM Classifier Fusion

For fault diagnosis of nonlinear analog circuit, a novel method based on generalized frequency response function (GFRF) and least square support vector machine (LSSVM) classifier fusion is presented. The sinusoidal signal is used as the input of analog circuit, and then, the generalized frequency response functions are estimated directly by the time-domain formulations. The discrete Fourier transform of measurement data is avoided. After obtaining the generalized frequency response functions, the amplitudes of the GFRFs are chosen as the fault feature parameters. A classifier fusion algorithm based on least square support vector machine (LSSVM) is used for fault identification. Two LSSVM multifault classifiers with different kernel functions are constructed as subclassifiers. Fault diagnosis experiments of resistor-capacitance (RC) circuit and Sallen Key filter are carried out, respectively. The results show that the estimated GFRFs of the circuit are accurate, and the fault diagnosis method can get high recognition rate.

New imbalanced fault diagnosis framework based on Cluster-MWMOTE and MFO-optimized LS-SVM using limited and complex bearing data

Due to the complexity of their working conditions, historical rolling bearing datasets are mostly limited and imbalanced. The fault data may be composed of multiple subclusters; that is, the historical rolling bearing data have both between-class and within-class imbalances. While support vector machines (e.g., least squares support vector machines (LS-SVMs)) offer advantages when dealing with limited data, traditional fault diagnosis using an LS-SVM has the disadvantages of easy failure of complex imbalanced data and large dependence on the classifier hyperparameters. Therefore, this paper presents a new imbalanced fault diagnosis framework based on a cluster-majority weighted minority oversampling technique (Cluster-MWMOTE) and a moth-flame optimization (MFO)-based LS-SVM classifier. As an extension of MWMOTE, our proposed Cluster-MWMOTE combines the clustering algorithm represented by agglomerative hierarchical clustering (AHC) with MWMOTE. Unlike MWMOTE, Cluster-MWMOTE can avoid the ignoring of small subclusters of faulty (minority) instances far from normal (majority) instances. That is, Cluster-MWMOTE further improves the adaptation to within-class imbalances. As a novel heuristic intelligent algorithm, MFO exhibits faster convergence and higher precision than the traditional optimization algorithms (e.g., particle swarm optimization (PSO) and genetic algorithm (GA)). Therefore, we utilize MFO to optimize the hyperparameters (Sigma &γ) of the LS-SVM classifier for the first time. The fault diagnosis results represented by CWRU and IMS bearing data suggest that the proposed framework provides higher fault diagnosis recognition rates and algorithm robustness than 16 existing algorithms.

A Comparative Analysis of Wavelet-Based FEMG Signal Denoising with Threshold Functions and Facial Expression Classification Using SVM and LSSVM

This work presents a technique for the analysis of facial electromyogram signal activities to classify five different facial expressions for computer-muscle interfacing applications. Facial electromyogram (FEMG) is a technique for recording the asynchronous activation of neuronal inside the face muscles with non-invasive electrodes. FEMG pattern recognition is a difficult task for the researcher, where classification accuracy is key concerns. Artifacts, such as eyeblink activity and electroencephalogram (EEG) signals interference, can corrupt these FEMG signals and directly affected the classification results. In this work, a robust wavelet-based thresholding technique, which comprised of a wavelet transform (WT) method and the statistical threshold, is proposed to remove the different artifacts from FEMG datasets and improve recognition accuracy rate. A set of five different raw FEMG data set was analyzed. Four wavelet basis functions, namely, haar, coif3, sym3, and bior4.4, were considered. The performance parameters (signal-to-artifact ratio (SAR) and normalized mean square error (NMSE) were utilized to understand the effect of the proposed signal denoising protocol. After denoising, the effectiveness of different statically features has been extracted. Two pattern recognition algorithms support vector machine (SVM) and the least square support vector machine (LSSVM) are implemented to classify extracted features. The performance accuracy of SVM and LSSVM classifier was evaluated and compared to know which classifier is the best for facial expression classification. The results showed that: (i) the proposed technique for denoising, improves the performance parameter results; (ii) The proposed work gives the best 95.24% classification accuracy.