Support Vector Machine Method Research Articles

Perbandingan Metode Random Forest, Convolutional Neural Network, dan Support Vector Machine Untuk Klasifikasi Jenis Mangga

Mango is a fruit known as the "King of Fruit" due to its rich flavor, vast variability, and high nutritional value. Classifying mangoes based on their external appearance is the initial step in the process of identifying and categorizing mango types conventionally. The classification process can be performed by examining external features such as fruit color, shape, and size. Classifying different types of mango fruits accurately can assist researchers in developing superior varieties and also aid farmers for cultivation purposes, sales, distribution, and selecting the right varieties for local growth and weather conditions. This research conducts the classification of mango types based on color from mango images using machine learning. The study compares three methods, namely Random Forest, Support Vector Machine (SVM), and Convolutional Neural Network (CNN), to determine the best method for classifying mango types based on their images. The dataset underwent preprocessing, where image sizes were standardized to 300 x 300 pixels, and color was changed to grayscale. The dataset was then divided into training and testing data with a ratio of 70:30. Subsequently, the dataset was processed using three methods, and their accuracy results were compared. The findings indicate that the Random Forest method yielded the highest accuracy compared to the other methods, with an accuracy rate of 96%. The accuracy of the SVM method was 95%, and the accuracy of the CNN method was 33%. From these results, it can be concluded that the Random Forest method is highly effective for classifying mango types based on their image compared to SVM and CNN methods.

Identifying Fraud Financial Reports Based on Signs of Income Management Using Machine Learning Technology: The Case of Listed Companies in Vietnam

This study aims to use a measure of earnings management to predict companies whose financial statements have problems. This is an identification measure other than common measures to predict financial statement fraud such as measuring by the M-Score or the Z-score model that many previous studies have applied. In the income management measure, the author uses a measure of abnormal cash flow and abnormal expense flow to consider whether the corporate financial statements have problems or not. To do this, the author uses data from listed non-financial enterprises in the period from 2018 to 2022, with machine learning and deep learning algorithms, of which we focus on three main algorithms: ANN, SVM and RF. The results show that identifying problematic financial statements based on abnormal cash flows is quite effective with an accuracy of over 78% for the SVM method, while if using the RF method, the accuracy reaches over 82% but it is required to accept an increased processing time.

Machine learning for non-experts: A more accessible and simpler approach to automatic benthic habitat classification

Automating identification of benthic habitats from imagery, with Machine Learning (ML), is necessary to contribute efficiently and effectively to marine spatial planning. A promising method is to adapt pre-trained general convolutional neural networks (CNNs) to a new classification task (transfer learning). However, this is often inaccessible to a non-specialist, requiring large investments in computational resources and time (for user comprehension and model training). In this paper, we demonstrate a simpler transfer learning framework for classifying broad deep-sea benthic habitats. Specifically, we take an ‘off-the-shelf’ CNN (VGG16) and use it to extract features (pixel patterns) from benthic images (without further training). The default outputs of VGG16 are then fed in to a Support Vector Machine (SVM), a classical and simpler method than deep networks. For comparison, we also train the remaining classification layers of VGG16 using stochastic gradient descent. The discriminative power of these approaches is demonstrated on three benthic datasets (574–8353 images) from Norwegian waters; each using a unique imaging platform. Benthic habitats are broadly classified as Soft Substrate (sands, muds), Hard Substrate (gravels, cobbles and boulders) and Reef (Desmophyllum pertusum). We found that the relatively simplicity of the SVM classifier did not compromise performance. Results were competitive with the CNN classifier and consistently high, with test accuracy ranging from 0.87 to 0.95 (average = 0.9 (±0.04)) across datasets, somewhat increasing with dataset size. Impressively, these results were achieved 2.4–5× faster than CNN training and had significantly less dependency on high-specification hardware. Our suggested approach maximises conceptual and practical simplicity, representing a realistic baseline for novice users when approaching benthic habitat classification. This method has wide potential. It allows automated image grouping to aid annotation or further model selection, as well as screening of old-datasets. It is especially suited to offshore scenarios as it can provide quick, albeit crude, insights into habitat presence, allowing adaptation of sampling protocols in near real-time.

Risk adjusted EWMA control chart based on support vector machine with application to cardiac surgery data

In the current study, we demonstrate the use of a quality framework to review the process for improving the quality and safety of the patient in the health care department. The researchers paid attention to assessing the performance of the health care service, where the data is usually heterogeneous to patient’s health conditions. In our study, the support vector machine (SVM) regression model is used to handle the challenge of adjusting the risk factors attached to the patients. Further, the design of exponentially weighted moving average (EWMA) control charts is proposed based on the residuals obtained through SVM regression model. Analyzing real cardiac surgery patient data, we employed the SVM method to gauge patient condition. The resulting SVM-EWMA chart, fashioned via SVM modeling, revealed superior shift detection capabilities and demonstrated enhanced efficacy compared to the risk-adjusted EWMA control chart.

Measurement of body size parameters and body weight prediction in beef cattle based on image analysis

The body size parameter of cattle is an important index reflecting the growth and development and health condition of cattle. The traditional manual contact measurement is not only a large workload and difficult to measure, but also prone to problems such as affecting the normal life habits of cattle. In this paper, we address this problem by proposing a contactless body size measurement method for cattle based on machine vision. Firstly, the cattle is confined to a fixed space using a position-limiting device, and images of the body of the cattle are taken from three directions: top, left, and right, using multiple cameras. Secondly, the image is segmented using a fuzzy clustering algorithm based on neighborhood adaptive local spatial information improvement, and the image is processed to extract the contour images of the top view and side view. The key points of body measurements were extracted using interval division and curvature calculation for the side view images, and the key point information was extracted using skeleton extraction and pruning for the top view images, which realized the measurements of body height(BH), rump height(RH), body slanting length(BSL), and abdominal circumference(AC) parameters of the cattle. The correlation between body size and weight data obtained by contactless methods was investigated and the modeled using one-factor linear regression, one-factor nonlinear regression, multivariate stepwise regression, RBF network fitting, BP neural network fitting, support vector machine, and particle swarm optimization-based support vector machine methods, respectively. Information on body size parameters was collected from 137 cattles, and the results showed that the maximum errors between the measured and actual values of BH, RH, BSL and AC were 5.0%, 4.4%, 3.6%, and 5.5%, respectively. Correlation of BH, RH, BSL and AC with weight obtained by non-contact methods was >  0.75. The BH parameter can be selected in the single-factor growth monitoring. The multi-body scale can reflect the growth status of cattle more comprehensively, in which RH, BSL and AC are important detection parameter; the multi-factor nonlinear model can reflect the growth characteristics of cattle more comprehensively. The contactless measurement method proposed in the paper can effectively improve the work efficiency and reduce the stress reaction of cattle, which is a long-term and effective monitoring method, and is of great significance in promoting accurate and welfare cattle rearing.

Altered functional connectivity strength of primary visual cortex in subjects with thyroid-associated ophthalmopathy.

Our objective was to explore the disparities in the intrinsic functional connectivity (FC) patterns of primary visual cortex (V1) between patients with thyroid-associated ophthalmopathy (TAO) and healthy controls (HCs) utilizing resting-state functional MRI. Twenty-one patients with TAO (14 males and 7 females; mean age: 54.17 ± 4.83 years) and 21 well-matched HCs (14 males and 7 females; mean age: 55.17 ± 5.37 years) underwent functional MRI scans in the resting-state. We assessed modifications in the intrinsic FC patterns of the V1 in TAO patients using the FC method. Subsequently, the identified alterations in FC regions in the analysis were selected as classification features to distinguish TAO patients from HCs through the support vector machine (SVM) method. The results indicated that, in comparison to HCs, patients with TAO exhibited notably reduced FC values between the left V1 and the bilateral calcarine (CAL), lingual gyrus (LING) and superior occipital gyrus, as well as between the right V1 and the bilateral CAL/LING and the right cerebellum. Furthermore, the SVM classification model based on FC maps demonstrated effective performance in distinguishing TAO patients from HCs, achieving an accuracy of 61.9% using the FC of the left V1 and 64.29% using the FC of the right V1. Our study revealed that patients with TAO manifested disruptions in FC between the V1 and higher visual regions during rest. This might indicate that TAO patients could present with impaired top-down modulations, visual imagery and vision-motor function. These insights could be valuable in understanding the underlying neurobiological mechanisms of vision impairment in individuals with TAO.

Prediction method of stalk content in strip particles after redrying of tobacco leaves

AbstractFour tobacco leaf modules processed by Yunnan Tobacco Redrying Co., Ltd. during the 2022 roasting season were used to investigate the method of stalk content in strip particles after redrying of tobacco leaves, effectively reducing the loss of strip particles. A total of 151 sets of experimental data were used to construct the prediction model for the stalk content in strip particles after redrying using the BP artificial neural network method, the linear regression method, and the support vector machine method. The results show that the prediction model constructed by the BP artificial neural network method has high accuracy and stability, with a relatively small absolute error of prediction (e = 0.0195%) and the root-mean-square error of interactive verification (RMSECV = 0.0227%), as well as a relatively small mean absolute error of production data validation (e = 0.0675%), while the prediction deviation ratio (RPD = 2.2435) is relatively large. Overall, the prediction model established by BP artificial neural network could provide new insight into the non-destructive detection of stalk content in strip particles of redried tobacco leaves after threshing and redrying and potentially leading to a reduction in tobacco leaf crushing by more than 112,500 kg per year.

The Forecast of Streamflow through Göksu Stream Using Machine Learning and Statistical Methods

Forecasting streamflow in stream basin systems plays a crucial role in facilitating effective urban planning to mitigate floods. In addition to employing intricate hydrological modeling systems, machine learning and statistical techniques offer an alternative means for streamflow forecasts. Nonetheless, the precision and dependability of these methods are subjects of paramount importance. This study rigorously investigates the effectiveness of three distinct machine learning techniques and two statistical approaches when applied to streamflow data from the Göksu Stream in the Marmara Region of Turkey, spanning from 1984 to 2022. Through a comparative analysis of these methodologies, this examination aims to contribute innovative advancements to the existing methodologies used in the prediction of streamflow data. The methodologies employed include machine learning methods such as Extreme Gradient Boosting (XGBoost), Random Forest (RF), and Support Vector Machine (SVM) and statistical methods such as Simple Exponential Smoothing (SES) and Autoregressive Integrated Moving Average (ARIMA) model. In the study, 444 data points between 1984 and 2020 were used as training data, and the remaining data points for the period 2021–2022 were used for streamflow forecasting in the test validation period. The results were evaluated using various metrics, such as the correlation coefficient (r), mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), coefficient of determination (R2), and Nash–Sutcliffe efficiency (NSE). Upon analyzing the results, it was found that the model generated using the XGBoost algorithm outperformed other machine learning and statistical techniques. Consequently, the models implemented in this study demonstrate a high level of accuracy in predicting potential streamflow in the river basin system.

Landscape Ecological Risk Assessment of Saihanba under the Change in Forest Landscape Pattern

Examining the Saihanba Mechanical Forest Farm, this study utilized Landsat remote sensing data from 1987, 1997, 2001, 2013, and 2020 to interpret land use from the Support Vector Machine (SVM) method, and to decipher evolving land use patterns over the last four decades. Grounded in landscape ecology theory, an innovative evaluation index for landscape ecological risk was introduced, leading to the delineation of 382 ecological risk evaluation units. Employing landscape pattern indices and a method of spatial autocorrelation, we analyzed the spatial and temporal distribution characteristics and spatial correlation patterns of landscape ecological risk across five distinct periods. Geostatistical approaches were used to explore the driving factors of landscape risk. The results indicate that since 1987, there have been significant changes in land use types, especially in forest landscapes, their proportion increasing from 23.19% to 74.55%. In 1987, the proportion of high-risk areas was 72.30%, but in 2020, high-risk areas had significantly decreased and clustered in specific locations. The landscape ecological risks in each period of the study area showed a positive spatial correlation and tended to gather in space. After comprehensive exploration using a geographic detector, we found that landscape type, temperature, and vegetation coverage are the main risk factors. Among them, landscape type has the greatest impact on the landscape and works together with slope, aspect, and precipitation. In forest farm management, only the adaptation and adjustment of single factors are often paid attention to, while the compound effects of multiple factors are ignored. The results of this study bring important reference value to the operation and development of forest farms.

Identifying a group of factors predicting cognitive impairment among older adults.

Cognitive impairment has multiple risk factors spanning several domains, but few studies have evaluated risk factor clusters. We aimed to identify naturally occurring clusters of risk factors of poor cognition among middle-aged and older adults and evaluate associations between measures of cognition and these risk factor clusters. We used data from the National Health and Nutrition Examination Survey (NHANES) III (training dataset, n = 4074) and the NHANES 2011-2014 (validation dataset, n = 2510). Risk factors were selected based on the literature. We used both traditional logistic models and support vector machine methods to construct a composite score of risk factor clusters. We evaluated associations between the risk score and cognitive performance using the logistic model by estimating odds ratios (OR) and 95% confidence intervals (CI). Using the training dataset, we developed a composite risk score that predicted undiagnosed cognitive decline based on ten selected predictive risk factors including age, waist circumference, healthy eating index, race, education, income, physical activity, diabetes, hypercholesterolemia, and annual visit to dentist. The risk score was significantly associated with poor cognitive performance both in the training dataset (OR Tertile 3 verse tertile 1 = 8.15, 95% CI: 5.36-12.4) and validation dataset (OR Tertile 3 verse tertile 1 = 4.31, 95% CI: 2.62-7.08). The area under the receiver operating characteristics curve for the predictive model was 0.74 and 0.77 for crude model and model adjusted for age, sex, and race. The model based on selected risk factors may be used to identify high risk individuals with cognitive impairment.

Combining Local Knowledge with Object-Based Machine Learning Techniques for Extracting Informal Settlements from Very High-Resolution Satellite Data

AbstractThe classification of informal settlements using very high-resolution (VHR) satellite data and expert knowledge has proven very useful for urban planning. The objective of this work was to improve the accuracy of informal settlement classification within the city of Riyadh, Saudi Arabia. The analysis incorporated the use of expert knowledge (EK). Twenty unique indicators relevant to informal settlements were identified by experts familiar with these areas, and incorporated into the image classification process. Object-based image analysis (OBIA) was then used to extract informal settlement indicators from a VHR image. These indicators were used to classify the image utilising two machine learning (ML) algorithms, random forest (RF) and support vector machine (SVM) methods. A VHR image (e.g., Worldview 3) of the city was employed. A total of 6,000 sample points were randomly generated, with 1800 used for training the VHR image. The classification process was able to clearly distinguish the formal settlement areas from informal areas, road networks, vacant blocks, shaded areas, and vegetation features. The object-based RF technique provided an overall accuracy of 96% (kappa value of 95%), while OB-SVM provided an accuracy of 95% (kappa of 91%). The results demonstrated that object-based ML methods such as RF and SVM, when combined with EK, can effectively and efficiently distinguish informal settlements from other urban features. This technique has the potential to be very useful for mapping informal settlements.

Drones Detection Using a Fusion of RF and Acoustic Features and Deep Neural Networks.

The use of drones has recently gained popularity in a diverse range of applications, such as aerial photography, agriculture, search and rescue operations, the entertainment industry, and more. However, misuse of drone technology can potentially lead to military threats, terrorist acts, as well as privacy and safety breaches. This emphasizes the need for effective and fast remote detection of potentially threatening drones. In this study, we propose a novel approach for automatic drone detection utilizing the usage of both radio frequency communication signals and acoustic signals derived from UAV rotor sounds. In particular, we propose the use of classical and deep machine-learning techniques and the fusion of RF and acoustic features for efficient and accurate drone classification. Distinct types of ML-based classifiers have been examined, including CNN- and RNN-based networks and the classical SVM method. The proposed approach has been evaluated with both frequency and audio features using common drone datasets, demonstrating better accuracy than existing state-of-the-art methods, especially in low SNR scenarios. The results presented in this paper show a classification accuracy of approximately 91% at an SNR ratio of -10 dB using the LSTM network and fused features.

The Prediction of the Remaining Useful Life of Rotating Machinery Based on an Adaptive Maximum Second-Order Cyclostationarity Blind Deconvolution and a Convolutional LSTM Autoencoder.

The prediction of the remaining useful life (RUL) is important for the conditions of rotating machinery to maintain reliability and decrease losses. This study proposes an efficient approach based on an adaptive maximum second-order cyclostationarity blind deconvolution (ACYCBD) and a convolutional LSTM autoencoder to achieve the feature extraction, health index analysis, and RUL prediction for rotating machinery. First, the ACYCBD is used to filter noise from the vibration signals. Second, based on the peak value properties, a novel health index (HI) is designed to analyze the health conditions for the denoising signal, showing a high sensitivity for the degradation of bearings. Finally, for better prognostics and health management of the rotating machinery, based on convolutional layers and LSTM, an autoencoder can achieve a transform convolutional LSTM network to develop a convolutional LSTM autoencoder (ALSTM) model that can be applied to forecast the health trend for rotating machinery. Compared with the SVM, CNN, LSTM, GRU, and DTGRU methods, our experiments demonstrate that the proposed approach has the greatest performance for the prediction of the remaining useful life of rotating machinery.

Partial label learning for automated classification of single-cell transcriptomic profiles.

Single-cell RNA sequencing (scRNASeq) data plays a major role in advancing our understanding of developmental biology. An important current question is how to classify transcriptomic profiles obtained from scRNASeq experiments into the various cell types and identify the lineage relationship for individual cells. Because of the fast accumulation of datasets and the high dimensionality of the data, it has become challenging to explore and annotate single-cell transcriptomic profiles by hand. To overcome this challenge, automated classification methods are needed. Classical approaches rely on supervised training datasets. However, due to the difficulty of obtaining data annotated at single-cell resolution, we propose instead to take advantage of partial annotations. The partial label learning framework assumes that we can obtain a set of candidate labels containing the correct one for each data point, a simpler setting than requiring a fully supervised training dataset. We study and extend when needed state-of-the-art multi-class classification methods, such as SVM, kNN, prototype-based, logistic regression and ensemble methods, to the partial label learning framework. Moreover, we study the effect of incorporating the structure of the label set into the methods. We focus particularly on the hierarchical structure of the labels, as commonly observed in developmental processes. We show, on simulated and real datasets, that these extensions enable to learn from partially labeled data, and perform predictions with high accuracy, particularly with a nonlinear prototype-based method. We demonstrate that the performances of our methods trained with partially annotated data reach the same performance as fully supervised data. Finally, we study the level of uncertainty present in the partially annotated data, and derive some prescriptive results on the effect of this uncertainty on the accuracy of the partial label learning methods. Overall our findings show how hierarchical and non-hierarchical partial label learning strategies can help solve the problem of automated classification of single-cell transcriptomic profiles, interestingly these methods rely on a much less stringent type of annotated datasets compared to fully supervised learning methods.

Design and Implementation of Machine Learning Algorithms in Automatic Grading of Students' Assignments

This study introduces machine learning methods for automatically grading students' assignments, catering to the changing demands of education in the digital age. By utilizing diverse techniques, the aim is to simplify the grading process, give prompt feedback, and improve overall efficiency in assessment practices. The research delves into the different types of assignments such as written responses and coding tasks, and presents a range of ML algorithms including NB, Decision Trees (DT), RF, SVM, Linear Regression (LR), KNN, and Ensemble Methods (EMs). The implementation involves extracting relevant features from assignments, preprocessing data to handle noise and outliers, and training models on a varied set of examples. The grading system prioritizes interpretability, accuracy, and efficiency while aligning with educational objectives and grading policies. Evaluation is done using suitable metrics like recall, accuracy, precision, F1 score for each algorithm. The study contributes to the progress of automated grading systems by providing valuable insights for educators on the potential impact of machine learning in enhancing assessment processes. Additionally, the designed algorithms demonstrate adaptability for different types of assignments and serve as a foundation for future enhancements in grading precision and efficiency.

Using the Support Vector Machine method with the HOG feature for classification of orchid types

Orchids are the most species-rich flowering plants, with approximately 750 genera and 43,000 types of orchids in the world, of which about 5,000 species have been recorded in several provinces in Indonesia. Orchids have beautiful flowers with attractive colors, making them ornamental plants that many people like. From plant morphology, orchid plants can be differentiated based on the morphology of flowers, leaves, fruit, stems, and roots. The leaves of orchid plants have their characteristics for each type of orchid, such as long, round, or lanceolate. All orchids have veins that run parallel to their leaves. The individual shapes of orchid leaves can be classified using a Support Vector Machine (SVM) and Histogram of Gradient (HOG). In this research, five types of orchids that are popular among orchid lovers were used, namely Dendrobium, Cattleya, Oncidium, Phalaenopsis, and Vanda orchids, which were taken from public data. The accuracy of this method in classifying orchid species based on leaf morphology can be measured using a confusion matrix that measures precision, recall, and accuracy. From five tests, the Oncidium orchid had the highest average accuracy with a value of 98%, the Vanda orchid had the highest average precision of 99.80%, and the Cattleya orchid had the highest average recall of 100%.

Prototype Fingerprint & Facial Attendance System Berbasis Raspberry Pi Dengan Metode Support Vector Machine

Biometrics is a part of the field of technology that approaches recognition systems designed to be able to recognize and identify characteristic parts of the human body. Parts of the body in humans that are used such as faces and fingerprints are used as patterns to carry out needs in the identification process, then these body parts are used as objects for means of identifying presence. Raspberry Pi is used as a tool that processes the integrity of biometric objects that are used with the help of a webcam and fingerprint sensor to recognize someone's face and fingerprints. Then a Raspberry Pi-based Face Attendance System Prototype was designed using the Support Vector Machine method. The Support Vector Machine method based on supervised learning is used to solve classification and regression problems. By using the Support Vector Machine algorithm in MATLAB to process facial images, it is known that the accuracy value obtained by testing 25 students with facial images obtained 375 images divided from one student consisting of 15 images. 250 images of training data were tested with a result of 65.20% and 125 images of test data with a result of 66.40%.

Evolutionary trends and analysis of the driving factors of Ulva prolifera green tides: A study based on the random forest algorithm and multisource remote sensing images

Understanding the prolonged spatiotemporal evolution and identifying the underlying causes of Ulva prolifera green tides play pivotal roles in managing such occurrences, restoring water ecology, and fostering sustainable development in marine ecosystems. Satellite remote sensing represents the primary choice for monitoring Ulva prolifera green tides due to its capability for extensive, long-term ocean monitoring. With the use of multisource ecological and environmental big data, such as GOCI Ⅰ/Ⅱ images, meteorological data, runoff data, and data on various water quality parameters (SST, ocean current speed, wind speed, precipitation, DO, PAR, Si, NO3-, PO4-, and N/P), pertaining to the highly impacted South Yellow Sea of China, an innovative remote sensing extraction model was established. This model involved a hybrid approach, merging the robustness of the random forest (RF) model and the optical algae cloud index (ACI) to map Ulva prolifera distribution patterns. Moreover, we analysed the evolutionary trends and the driving factors determining these distribution patterns. The ACI-RF method exhibited exceptional performance, with an F1 score surpassing 0.95, outperforming alternative methods such as the support vector machine (SVM) and K-nearest neighbour (KNN) methods. Over the period from 2011 to 2022, excluding 2021, there was a notable decline in the area of Ulva prolifera green tides, varying between 397 and 2689.9 km2, with an average annual reduction rate of 3%. The maximum annual biomass varied between 0.12 and 15.9 kt. Notably, more than 75% of the area of Ulva prolifera green tides exhibited northward drift, which was significantly influenced by northern currents and wind fields. The regions most susceptible to Ulva prolifera green tides were observed in areas of water eutrophication. The analysis of driving factors shows that factors such as average sea surface temperature, eastward wind speed, northward wind speed, precipitation, PO4-, and N/P/Si significantly influence the biological growth rate of Ulva prolifera. Furthermore, coastal land use change and surface runoff in Jiangsu Province, particularly surface runoff in June, significantly impacted the growth rate of Ulva prolifera, with Pearson correlation coefficients of 0.74 and 0.67, respectively. Against the background of global warming and severe deterioration in the marine environment, Ulva prolifera blooms persist. Consequently, two distinct management strategies were proposed based on the distribution patterns and cause analysis results for addressing Ulva prolifera green tides: establishing a continuous protection framework for rivers, lakes, and nearshore areas to mitigate pollutant inputs and implementing precise environmental monitoring measures in urban expansion areas and farmlands to combat overgrowth-induced green tides. Leveraging the inherent advantages of remote sensing big data for rapid mapping, this methodology could be applied in other regions affected by marine ecological disasters. Additionally, the criteria for selecting influencing factors offer a valuable reference for designing tailored measures aimed at controlling Ulva prolifera green tides.

A SVM Based Demand Response Capacity Prediction Model Considering Internal Factors under Composite Program

—Demand response capacity is a key reference for demand-side participation in peak regulation and earning profit in the electricity market, and it is influenced by not only external public factors but demand-side internal individual factors. Demand change caused by internal individual reasons cannot be recognized by public features. Single participating in response industrial and commercial consumers cannot balance this change through large numbers of participants, unlike aggregators. The prediction model only considering the external public factors will cause significant errors. However, the actual industrial and commercial data is difficult to obtain because of the cost and privacy, and the simulation platform cannot accurately describe this situation, resulting in the problem being ignored. Therefore, this paper aims to prove the importance of internal factors and proposes an improved method considering the problem. First, based on an actual dataset, the optimal responsive behavior of consumers is modeled to quantify response capacity. Second, a support vector machine (SVM) method is applied to predict response capacity. Finally, combined with raw data, this paper proves that significant errors are caused by internal factors and provides different improvement methods for internal and external operators. The case study indicates that the improved model can provide better performance.

Optimization of support vector machine for transformer core and winding looseness fault diagnosis based on improved grey wolf algorithm

To diagnose transformer core and winding looseness faults more timely and effectively, a complementary ensemble empirical mode decomposition (CEEMD) and improved grey wolf optimization-based support vector machine (IGWO-SVM) method for transformer core and winding looseness fault diagnosis are proposed. Firstly, the vibration signal is decomposed into multiple intrinsic mode functions (IMFs) through CEEMD. Secondly, the energy entropy of different IMFs is calculated, and the energy entropy of different states is formed into a feature dataset. Then, the improved GWO-optimized SVM model is used to classify and recognize the feature dataset. Finally, we establish an experimental platform for experimental verification. The results show that the proposed method can accurately and effectively diagnose transformer core and winding looseness faults, and has a high diagnostic accuracy, which is at least 3.5% higher than the existing optimal diagnostic models. The proposed method provides a theoretical reference for the development of transformer fault diagnosis strategies.