Support Vector Machine Prediction Research Articles

Rice Origin Tracing Technology Based on Fluorescence Spectroscopy and Stoichiometry.

The origin of agricultural products is crucial to their quality and safety. This study explored the differences in chemical composition and structure of rice from different origins using fluorescence detection technology. These differences are mainly affected by climate, environment, geology and other factors. By identifying the fluorescence characteristic absorption peaks of the same rice seed varieties from different origins, and comparing them with known or standard samples, this study aims to authenticate rice, protect brands, and achieve traceability. The study selected the same variety of rice seed planted in different regions of Jilin Province in the same year as samples. Fluorescence spectroscopy was used to collect spectral data, which was preprocessed by normalization, smoothing, and wavelet transformation to remove noise, scattering, and burrs. The processed spectral data was used as input for the long short-term memory (LSTM) model. The study focused on the processing and analysis of rice spectra based on NZ-WT-processed data. To simplify the model, uninformative variable elimination (UVE) and successive projections algorithm (SPA) were used to screen the best wavelengths. These wavelengths were used as input for the support vector machine (SVM) prediction model to achieve efficient and accurate predictions. Within the fluorescence spectral range of 475-525 nm and 665-690 nm, absorption peaks of nicotinamide adenine dinucleotide (NADPH), riboflavin (B2), starch, and protein were observed. The origin tracing prediction model established using SVM exhibited stable performance with a classification accuracy of up to 99.5%.The experiment demonstrated that fluorescence spectroscopy technology has high discrimination accuracy in tracing the origin of rice, providing a new method for rapid identification of rice origin.

Ensemble machine learning prediction of anaerobic co-digestion of manure and thermally pretreated harvest residues

This study aimed to clarify the statistical accuracy assessment approaches used in recent biogas prediction studies using state-of-the-art ensemble machine learning approach according to 10-fold cross-validation in 100 repetitions. Three thermally pretreated harvest residue types (maize stover, sunflower stalk and soybean straw) and manure were anaerobically co-digested, measuring biogas and methane yield alongside eight thermal preprocessing and biomass covariates. These were the inputs to an ensemble machine learning approach for biogas and methane yield prediction, employing three feature selection approaches. The Support Vector Machine prediction with the Recursive Feature Elimination resulted in the highest prediction accuracy, achieving the coefficient of determination of 0.820 and 0.823 for biogas and methane yield prediction, respectively. This study demonstrated an extreme dependency of prediction accuracy to input dataset properties, which could only be mitigated with ensemble machine learning and strongly suggested that the split-sample approach, often used in previous studies, should be avoided.

Studying Stock Prediction in the Context of Machine Learning Exemplified by Analyzing SVM and LSTM Models

The impact of stocks on economic development is significant, with most companies employing quantitative finance methods to analyze and trade stocks. In the present day, there is a growing plethora of stock prediction methods, requiring individuals to carefully select the appropriate means of prediction. This study aims to present the highly regarded machine learning algorithms of recent years. Machine learning enables the extraction of trading strategies by learning patterns from historical stock data. Deep learning, a subset of machine learning, eliminates the need for researchers to manually extract features, instead utilizing neural network models that have proven effective for stock analysis. This paper also provides an extensive analysis of Support Vector Machine (SVM) and Long Short-Term Memory (LSTM), elucidating their algorithmic principles and assessing their strengths and weaknesses based on the findings of other researchers. It is observed that SVM is effective for predicting small quantities, but lacks suitability for non-stationary data prediction due to the non-informative nature of financial time series data, which significantly diminishes the accuracy of SVM predictions. In contrast, LSTM emerges as the most comprehensive single algorithm, exhibiting high accuracy and efficiency across all aspects. The paper also explores advanced hybrid models, which demonstrate that combining the advantages of multiple models can lead to higher accuracy and efficiency. The future of stock prediction algorithms holds promise for development, with the potential to construct an intelligent system that automatically selects data for analysis.

Support vector machine prediction of individual Autism Diagnostic Observation Schedule (ADOS) scores based on neural responses during live eye-to-eye contact

Social difficulties during interactions with others are central to autism spectrum disorder (ASD). Understanding the links between these social difficulties and their underlying neural processes is a primary aim focused on improved diagnosis and treatment. In keeping with this goal, we have developed a multivariate classification method based on neural data acquired by functional near infrared spectroscopy, fNIRS, during live eye-to-eye contact with adults who were either typically developed (TD) or individuals with ASD. The ASD diagnosis was based on the gold-standard Autism Diagnostic Observation Schedule (ADOS) which also provides an index of symptom severity. Using a nested cross-validation method, a support vector machine (SVM) was trained to discriminate between ASD and TD groups based on the neural responses during eye-to-eye contact. ADOS scores were not applied in the classification training. To test the hypothesis that SVM identifies neural activity patterns related to one of the neural mechanisms underlying the behavioral symptoms of ASD, we determined the correlation coefficient between the SVM scores and the individual ADOS scores. Consistent with the hypothesis, the correlation between observed and predicted ADOS scores was 0.72 (p < 0.002). Findings suggest that multivariate classification methods combined with the live interaction paradigm of eye-to-eye contact provide a promising approach to link neural processes and social difficulties in individuals with ASD.

An integrated data mining algorithms and meta-heuristic technique to predict the readmission risk of diabetic patients

Reducing hospital readmission rate is a significant challenge in the healthcare industry for managers and policymakers seeking to improve healthcare and lower costs. This study integrates data mining and meta-heuristic techniques to predict the early readmission probability of diabetic patients within 30 days of discharge. The research dataset was obtained from the UC Irvine Machine Learning Repository, including 101765 instances with 50 features representing patient and hospital outcomes, collected from 130 US hospitals. After data preprocessing, including cleansing, sampling, and normalization, a Chi-square analysis is done to confirm and rank the 20 identified factors affecting the readmission risk. As the algorithms' performance could vary based on the features’ characteristics, several classification algorithms, including a Random Forest (RF), Neural Network (NN), and Support Vector Machine (SVM), are employed. Moreover, the Genetic Algorithm (GA) is integrated into the SVM algorithm, called GA-SVM, for hyper-parameter tuning and increasing the prediction accuracy. The performance of the models was evaluated using accuracy, recall, precision, and f-measure metrics. The results indicate that the accuracy of RF, GA-SVM, SVM, and NN are calculated respectively as 74.04 %, 73.52 %, 72.40 %, and 70.44 %. Using GA to adjust c and gamma hyper-parameters led to a 1.12 % increase in SVM prediction accuracy. In response to increasing demand and considering poor hospital conditions, particularly during epidemics, these findings point out the potential benefits of a more tailored methodology in managing diabetic patients.

An implementation of inertia control strategy for grid-connected solar system using moth-flame optimization algorithm

As the solar power system power system grows rapidly, inertia control strategy (ICS) becomes crucial to enable stable grid integration. However, the existing ICS lacks of dynamic weather analysis with maximum power point tracking (MPPT) and fault-ride through (FRT) capabilities such as low voltage ride-through (LVRT) and high voltage ride-through (HVRT). In this work, an inertia weighting strategy and the Cauchy mutation operator are introduced to improve the moth-flame optimization (MFO) algorithm to support vector machine prediction of photovoltaic power generation. In this paper, the proposed adaptive VICS with variable moment of inertia (J) and damping factor ( D P ) demonstrates its effectiveness with faster frequency recovery, less overshooting and continuous stable operation under grid fault and dynamic weather. The MFO algorithm is used to implement inertia control strategies for grid-connected solar systems. Accurate simulation results confirm the inertia control of the emulsion and the control of the solar system. The results of the simulation show a significant improvement in frequency with the designed MFO and compared to Horse Herd Optimization (HHO). The proposed method contributes to improve photovoltaic energy prediction, reduces the impact of photovoltaic power penetration into the grid and maintains the system reliability.

Determination of the Acid and Peroxide Values of Vegetable Oils by Raman Spectroscopy with Competitive Adaptive Reweighted Sampling (CARS) and Back Propagation Neural Network (BPNN)

Acid and peroxide values are important indicators to evaluate the hydrolysis or oxidation level of vegetable oil. In order to achieve rapid determination of these parameters by Raman spectroscopy, rapeseed and soybean oil were analyzed, and a competitive adaptive reweighted sampling-back propagation neural network (CARS-BPNN) method was developed. First, the Raman spectrum was preprocessed to reduce noise using moving window smoothing. Next, competitive adaptive reweighted sampling (CARS), successive projections algorithm (SPA), and uninformative variable elimination (UVE) was applied to extract the Raman characteristic information that indicated the acid and peroxide values. Based on the extracted characteristic information, back propagation neural network (BPNN), partial least squares regression (PLSR), and support vector machine (SVM) prediction models were established and compared. The results showed that the best accuracy was obtained using CARS-BPNN. For rapeseed oil, the coefficients of determination (R2) of acid value and peroxide value were 0.9207 and 0.9772. The root mean square error (RMSE) values were 0.0909 and 0.0022, respectively. For soybean oil, the coefficients of determination (R 2) of acid and peroxide values were 0.9381 and 0.9716. The RMSE values were 0.1250 and 0.0027, respectively. A good correlation coefficient and a small RMSE value were obtained. The results indicate that Raman spectroscopy combined with CARS-BPNN accurately evaluated the acid and peroxide values, which provides a reference for the optimization of the Raman spectrum and creates a new approach for the evaluation of vegetable oil quality.

Debris Flow Scale Prediction Based on Correlation Analysis and Improved Support Vector Machine

The occurrence of debris flows are a significant threat to human lives and property. Estimating the debris flow scale is a crucial parameter for assessing disaster losses in such events. Currently, the commonly used method for estimating debris flow runoff relies on fitting techniques, which often yield low prediction accuracy and limited data representation capabilities. Addressing these challenges, this study proposes an improved grey wolf algorithm optimized support vector machine prediction model. The model’s effectiveness is validated using data from 72 debris flow events in Beichuan County. The results demonstrate a prediction accuracy of 95.9% using this approach, indicating its strong predictive capabilities for debris flow scale. Additionally, it is observed that the basin area, the basin relative, and the main channel length are the key factors influencing debris flow scale in Beichuan County.

Calculation of exact Shapley values for explaining support vector machine models using the radial basis function kernel

Machine learning (ML) algorithms are extensively used in pharmaceutical research. Most ML models have black-box character, thus preventing the interpretation of predictions. However, rationalizing model decisions is of critical importance if predictions should aid in experimental design. Accordingly, in interdisciplinary research, there is growing interest in explaining ML models. Methods devised for this purpose are a part of the explainable artificial intelligence (XAI) spectrum of approaches. In XAI, the Shapley value concept originating from cooperative game theory has become popular for identifying features determining predictions. The Shapley value concept has been adapted as a model-agnostic approach for explaining predictions. Since the computational time required for Shapley value calculations scales exponentially with the number of features used, local approximations such as Shapley additive explanations (SHAP) are usually required in ML. The support vector machine (SVM) algorithm is one of the most popular ML methods in pharmaceutical research and beyond. SVM models are often explained using SHAP. However, there is only limited correlation between SHAP and exact Shapley values, as previously demonstrated for SVM calculations using the Tanimoto kernel, which limits SVM model explanation. Since the Tanimoto kernel is a special kernel function mostly applied for assessing chemical similarity, we have developed the Shapley value-expressed radial basis function (SVERAD), a computationally efficient approach for the calculation of exact Shapley values for SVM models based upon radial basis function kernels that are widely applied in different areas. SVERAD is shown to produce meaningful explanations of SVM predictions.

Experimental and support vector machine predictions of entropy generations and exergy efficiency of Fe3O4–SiO2/Water hybrid nanofluid in a plate heat exchanger

Several experiments of Fe3O4–SiO2/water hybrid nanofluids with volumetric concentrations ranging from 0.2 % to 1.0 % circulating in the cold-side of a plate heat exchanger at flow rates ranging from 0.05 kg/s to 0.1166 kg/s are performed. Under these ranges of flow rates and volumetric concentrations, the flow of Fe3O4–SiO2/water hybrid nanofluids remains laminar. The results of these experiments are predicted with support vector machine (SVM) algorithm to determine hybrid nanofluid entropy generation thermal, entropy generation frictional, and efficiency of exergy. Fe3O4–SiO2 nanomaterials was synthesized with reduction of chemicals and insitu development techniques, with XRD, FTIR and VSM instruments, characterizations were done. The SVM model gives large precision predictions of the measured data with correlations coefficients of 0.9944, 0.99798, and 0.99428 for frictional entropy generation, thermal entropy generation and exergy efficiency. At a flow rate of 0.1166 kg/s in the cold-side of PHE, the exergy efficiency is found to be 77.96 % for water (Reynolds number of 935.4) and with 1.0 vol% of Fe3O4–SiO2/water hybrid nanofluid in the cold-side of PHE, the efficiency is increased to 82.97 %, respectively. Under similar conditions of 0.1166 kg/s of flow circulation and 1.0 % vol. concentration of hybrid nanofluid, the thermal entropy generation is dropped off to 18.37 %, but the frictional entropy generation is increased by 20.97 %, compared to water, with the results that the total entropy generation drops off by 15.91 %, compared to water data. Preliminary curve-fitting correlations have been developed for the frictional entropy generation, thermal entropy generation, and exergy efficiency.

Decomposition-integration-based prediction study on the development trend of film industry

Movies have the unique ability to both generate income and spread culture, thus predicting the direction of the film industry's growth has garnered a lot of interest. Given the volatility of the movie industry's entire box office revenue dataset and the peculiarities of tiny samples, this article incorporates the decomposition-integration notion to build the EEMD-PSO-LSSVM model movie box office prediction model. The historical box office data are first broken down into many components using the ensemble empirical modal decomposition technique, and then, distinct sequences are predicted using the least squares support vector machine prediction method with particle swarm optimization, and ultimately, the predictions for each sequence are combined. The experimental results demonstrate the effectiveness of the decomposition-integration technique in illustrating the data fluctuation characteristics of quarterly movie box office revenues. When compared to other models, the model proposed in this study has clear advantages in the problem of predicting the time series data of box office revenues that are non-linear, non-smooth, and non-large samples.

Prediction of Surface Settlement in the Cross-passage by FIG-SVM Approach

Introduction: An optimized support vector machine prediction model (SVM) with fuzzy information granulation (FIG) is established for surface prediction in this paper. Based on the measured data, cubic spline interpolation was processed and FIG approach was applied to granulate the surface settlement parameters into fuzzy particles Low, R and Up, and the particles are used to represent the range of the measured data variation. Method: For each fuzzy particle, particle swarm optimization (PSO) was used to select the best penalty and kernel function parameters to minimize the K-fold cross-validation (K-CV) error. With the optimized parameters, the prediction model was trained for the nonlinear prediction of fuzzy particles. Finally, the surface settlement trend in terms of Low, R, Up for the next construction stage can be predicted effectively based on the previous test data. Taking the surface settlement monitoring point D5-5 as an example, the error of the predicted range of surface settlement at D5-5 is 5.82%, 5.42%, and 8.0%, respectively Result: The error of other predicted points is also less than 10%, indicating the effectiveness of the prediction model. Compared with the numerical simulation results, the accuracy of the prediction model was further verified. Conclusion: Combined with the simulation method, the "simulation - prediction" patent scheme for monitoring the surface settlement of the Cross Passage is proposed in this paper. The research results indicate that the model proposed in this paper can easily and effectively predict the range and trend of changes in surface settlement, and is suitable for practical engineering applications.

Goaf risk prediction based on IAOA–SVM and numerical simulation: A case study

In regard to goaf risk prediction, due to the low accuracy and single prediction method, this study proposes a method that combines the improved arithmetic optimization algorithm (IAOA) – support vector machines (SVM) with GoCAD–FLAC3D numerical simulation. Thus, goaf risk is comprehensively predicted. From the perspectives of geological and engineering conditions, eight factors that affect goaf stability and 176 sets of sample data were determined. We utilized eight influencing factors such as rock mass structure, geological structure, and goaf burial depth as inputs, and the goaf risk level as the output. Moreover, an IAOA–SVM goaf risk prediction model was established. The 30 goaf areas of Yangla Copper Mine in Yunnan Province were selected as the research subject. First, the rationality of mechanical parameter values in the numerical model was verified using the parameter inversion method. Second, based on the GoCAD–FLAC3D numerical simulation method, the goaf risk analysis in Yangla Copper Mine was performed. Subsequently, using numerical simulation verification, the goaf filling effect was analyzed. Finally, the prediction results of the IAOA–SVM model were compared with that of other intelligent algorithms. The results indicate that the numerical simulation results of the GoCAD–FLAC3D model are consistent with those of IAOA–SVM and the actual results, which further verifies the effectiveness and superiority of the IAOA–SVM prediction model. Therefore, an innovative approach for goaf risk prediction is developed.

Mechanical properties of Lop Nor salt rock fillers for subgrade and its forecast model construction

A deep exploration was conducted on the evolution law of the mechanical properties of salt rock fillers under the influence of different forming parameters in the Lop Nor area to promote the engineering application process of salt rock. The accuracy of primary regression, square regression, and support vector machine (SVM) regression algorithms in predicting the mechanical properties of salt rock fillers was compared, and the most accurate prediction model for the California bearing ratio (CBR) and rebound modulus of salt rock fillers were recommended. The results showed that the optimal brine content of SC-40, SC-20, and SC-10 salt rock fillers was between 8.2% and 9.3%, with a dry density of approximately 1.69–1.76 g/cm3. The CBR of salt rock samples gradually decreased with an increase in brine content, and the rebound modulus was higher than 90.6 MPa. As the compaction degree increased, the CBR value increased significantly, and the rebound modulus increased by approximately 28.1 MPa. As the immersion time increased, the mechanical properties of salt rock gradually decreased. Among the various regression models, the SVM prediction model had the highest accuracy index coefficient of determination (R2), whereas the mean absolute percentage error (MAPE), root mean square error (RMSE), and mean absolute error (MAE) were the smallest. Therefore, the SVM prediction model was recommended to accurately estimate the mechanical properties of salt rock roadbed fillers and provide a reference for the regulation of compaction parameters and the guarantee of bearing capacity characteristics of salt rock roadbeds.

CT-based radiomics model for the prediction of genomic alterations in renal cell carcinoma (RCC)

Abstract Background Radiogenomics is an emerging tool with applications in screening for molecular biomarkers in diagnostic and prognostic assessment through the extraction of quantitative data from medical images (Shui et al., Front Oncol 2021). In this study, we aim to develop machine learning (ML) models to predict the most common genomic alterations present in our subset of renal cell carcinoma patients (pts). Methods Retrospectively, pts with tissue genomic testing from CT-guided biopsy samples were identified. Genomic testing was done via GEM ExTra assay, a CAP-accredited, CLIA-certified test encompassing tumor whole exome sequencing and whole transcriptome sequencing (TGen; Phoenix, AZ). Biopsy sample collection sites were identified from pre-biopsy contrast CT images, and the lesions were segmented with ITK-SNAP software, from which 510 radiomic features were extracted with Pyradiomics. The Least Absolute Shrinkage and Selection Operator (LASSO) regression was used to select the most relevant features. Logistic regression (LR) and support vector machine (SVM) classifiers were built for the prediction of PBRM1, VHL, and SETD2 gene alterations. Multiple metrics were used to evaluate the predictive performance via leave-one-out cross-validation, including the area under the receiver operating characteristic (AUROC) and the area under the precision-recall curve (AUPRC). Feature importance was evaluated with Shapley additive explanations (SHAP) method. Results A total of 14 RCC pts (10:4 M:F) with genomic testing from CT-guided biopsies were identified. The majority of pts were White (85.7%) and had clear cell histology (71.4%). The most common locations for the CT-guided biopsy were lung (18.2%), soft tissue (13.6%), kidney (9.1%), and bone (9.1%). The most common alterations were seen in PBRM1 (50%), VHL (43.9%), and SETD2 (35.7%) genes. The PBRM1 gene was predicted with the highest AUROC (0.84) and AUPRC (0.88) with the SVM classifier, followed by the SETD2 gene (AUROC=0.78 and AUPRC=0.66) with LR classifier and VHL gene (AUROC=0.56 and AUPRC=0.65) with SVM classifier. Notably, all three models showed good sensitivity in classifying gene mutation status (PBRM1: 0.86; VHL: 0.88; SETD2: 0.89). Among all radiomic features, first-order features, Gray Level Size Zone Matrix features, and Gray Level Dependence Matrix features were found to be the most important features for predictions. Conclusions Using a CT-based radiomics analysis of the biopsy area, we showed that SVM and LR prediction models could predict PBRM1, VHL, and SETD2 mutations with high accuracy. These models may assist in identifying potentially actionable alterations and yield ease in treatment selection for RCC. Further extensive studies are warranted to validate our findings and improve our model. CDMRP DOD Funding: no

EEG-Based Emergency Braking Prediction Using Data Ablation and SVM Classification

Contemporary advanced driver assistance system (ADAS) features for semi-autonomous vehicles include braking assistance during collision avoidance. Although pre-collision detection typically relies on sensing systems to enable production vehicles to perceive oncoming road obstacles, the physiological state of the driver is not measured to predict emergency braking. On the other hand, previous driving simulation experiments have demonstrated the ability of regularized linear discriminant analysis (RLDA) to predict pre-collision braking using brain signals from multiple electroencephalogram (EEG) electrodes. In contrast, the current study used EEG data from these previous experiments to determine the quality of support vector machine (SVM) predictions as a first step towards realizing a brain-computer interface for emergency braking. Power spectral density (PSD) features were extracted from the EEG of one electrode to train and evaluate an SVM. Through a novel data ablation analysis, the optimal number of PSD components was determined to optimize model classification quality measured by the area under the curve (AUC). A comparison of the proposed model to the previous RLDA and other machine learning methods indicated that the SVM had a superior AU C. Thus, the proposed model is a candidate for assisting ADASs with pre-collision detection. Moreover, since the proposed model only utilized one electrode, our study potentially contributes to the facilitation of brain-computer interfaces for autonomous vehicles.

Effect of different spectral processing methods on the quantitative model of Astragalus polysaccharides detected by NIR spectroscopy

Quality control of Chinese herbal medicines requires reliable and rapid detection techniques, Many scholars have used NIR spectroscopy for research in Chinese herbal medicine industry. The aim of this study was to evaluate the feasibility of combining NIR spectroscopy and machine learning to predict the adulteration content of Astragalus polysaccharides. The astragalus polysaccharide and rice flour were mixed to form astragalus polysaccharide adulterant with different concentration. the Quantitative prediction models are built using the support vector machine (SVM). Different pre-processing methods of SG, SNV, WT, SG+SNV, SG+WT were used to process the spectra, and then the continuous projection algorithm (SPA) and uninformative variable elimination (UVE) were used to select the characteristic wavelengths. The SVM prediction model based on SG+SNV+SPA is optimal with an RPD of 2.653, indicating that the model has good predictive ability. The results showed that NIR technique could be used for quantitative analysis of Astragalus polysaccharides.

Construction and evaluation of in-house methylation-sensitive SNaPshot system and three classification prediction models for identifying the tissue origin of body fluid.

The identification of tissue origin of body fluid can provide clues and evidence for criminal case investigations. To establish an efficient method for identifying body fluid in forensic cases, eight novel body fluid-specific DNA methylation markers were selected in this study, and a multiplex singlebase extension reaction (SNaPshot) system for these markers was constructed for the identification of five common body fluids (venous blood, saliva, menstrual blood, vaginal fluid, and semen). The results indicated that the in-house system showed good species specificity, sensitivity, and ability to identify mixed biological samples. At the same time, an artificial body fluid prediction model and two machine learning prediction models based on the support vector machine (SVM) and random forest (RF) algorithms were constructed using previous research data, and these models were validated using the detection data obtained in this study (n=95). The accuracy of the prediction model based on experience was 95.79%; the prediction accuracy of the SVM prediction model was 100.00% for four kinds of body fluids except saliva (96.84%); and the prediction accuracy of the RF prediction model was 100.00% for all five kinds of body fluids. In conclusion, the in-house SNaPshot system and RF prediction model could achieve accurate tissue origin identification of body fluids.

Numerical investigation and prediction of the excavation face stability for river-crossing shield tunneling: An intelligent prediction model for limit support pressure

The instability of the shield tunnel's excavation face, particularly underwater tunnels, is one of the most hazardous factors in subsea tunnel excavation projects. Therefore, an accurate prediction of the limit support pressure to maintain the excavation face's stability is essential to minimize the possible risk of damage. This study proposed an intelligent prediction model to address these issues. First, the analytical formula of the limit support pressure of the excavation face was established through mechanical analysis to give the mechanics criterion for active instability of the excavation face. Subsequently, the numerical simulations were conducted to reveal the evolution mechanism of the excavation face instability. The result revealed that the buried depth of the tunnel, water depth, driving speed, internal friction angle, water content, cohesion, and support safety factor were the main factors affecting the limit support pressure. Furthermore, the support vector machine (SVM) model was established for predicting the limit support pressure. The prediction results agreed well with the actual data, thereby indicating the feasibility and convenient implementation of the SVM predictor. The proposed model was validated as an effective method for predicting the limit support pressure for further engineering applications.

Combining GA-SVM and NSGA-Ⅲ multi-objective optimization to reduce the emission and fuel consumption of high-pressure common-rail diesel engine

This research proposed a multi-objective optimization approach that combines Non-dominated Sorting Genetic Algorithms (NSGA) Ⅲ and support vector machine (SVM) to reduce diesel engine emissions while enhancing economic performance and calibration efficiency. In order to obtain accurate experimental data on diesel engines, a space-filling design method was proposed based on the prediction modeling of diesel engine performance. The SVM prediction model for diesel engine performance was established. A genetic algorithm (GA) was introduced to optimize the SVM model's penalty factor and radial basis parameters, thereby improving its prediction accuracy. The multi-objective optimization approach optimized the braking specific fuel consumption (BSFC), NOx, and CO. The results show that: the GA-SVM diesel engine performance prediction model has excellent prediction performance and generalization ability for BSFC, NOx, and CO, with R2 values of 0.981, 0.979, and 0.968, respectively. GA-SVM was used to evaluate the fitness of the NSGA-III optimal set. This not only ensures optimization accuracy but also improves working efficiency. After optimization, the BSFC of the diesel engine was reduced by 1.67%, NOx emission was reduced by 27.01%, CO emission was reduced by 19.15%, and noticeable optimization results were obtained. This work has important reference value for the automatic calibration of diesel engine control parameters, improving the economy and emission of diesel engines.