SVC Model Research Articles

Safer Floors in Public Service Buildings Based on Machine Learning

Abstract This study investigates the effects of different floor surfaces on slip safety in public service buildings (PSBs) with heavy pedestrian traffic. The K-means clustering method is used to classify various floor types and slip safety risks. The dynamic friction coefficient (DCOF) for floor coverings, such as natural stone, ceramic, laminate, and PVC, were measured in both dry and wet conditions across thirty public institutions. These measurements were obtained using the GMG 200 and Wessex S885 pendulum testers, providing a comprehensive assessment of the slip resistance of these surfaces. The machine learning models employed in the study were XGBoost, K-Nearest Neighbors (KNN), and SVC. The models were evaluated using 5-fold cross-validation. The analysis revealed that the most significant parameter in DCOF predictions for the XGBoost model was environmental conditions (EC). Performance analysis showed that the SVC model achieved the highest F1 score (0.75 ± 0.01) and AUC value (0.83), outperforming the other models. Additionally, DCOF values from slip tests were grouped into five clusters using the K-means method, and a slip safety risk scale was developed. Statistically significant differences were observed in DCOF values based on usage areas, environmental conditions, test methods, and surface materials. For instance, hospital floors were found to be generally safe in dry conditions but posed a risk in wet conditions. Based on these findings, actionable safety measures were suggested, such as applying anti-slip coatings in high-risk areas, selecting flooring materials with higher DCOF values for moisture-prone environments, and implementing regular slip resistance testing to maintain safety standards. In conclusion, this study demonstrates that machine learning models can effectively assess the slip resistance of floor surfaces. The findings offer valuable guidance for construction industry professionals and researchers in improving safety measures and minimizing slip risks. Future research with larger datasets and diverse conditions could enhance the understanding of this issue and further improve model performance.

Enhancing healthcare with AI: Sustainable AI and IoT-Powered ecosystem for patient aid and interpretability analysis using SHAP

The healthcare industry has blossomed into one of the most pivotal and technologically advanced sectors in the past decade. Individuals grapple with the peril of untimely demise from diverse ailments as patients suffer from delayed treatment. The paramount objective is to forge a dependable patient care system utilizing the Internet of Things (IoT), enabling physicians to monitor patients' well-being within medical facilities or even the confines of their homes. The system aids in tracking the patient's SpO2 level, body temperature, pulse rate (beats per minute), room temperature, and humidity, then trains the data with machine learning algorithms for the patient and finally monitors it through the Blynk IoT system. The cloud-stored data can be harnessed to ascertain and supervise one's health and predict forthcoming perils. This study unveils an efficacious decision-making model custom-tailored for Internet of Things (IoT) ventures, and the proposed trained algorithm satiates these requirements, offering efficiency and precision, rendering it appropriate for numerous IoT applications. Finally, the Shapley Additive Explanations (SHAP) is used here for finding out the most influential parameters, and Explainable AI (XAI) is utilized with the help of SHAP values for enhanced information on affecting parameters. The SVC model's hyperparameter is properly adjusted, yielding a testing accuracy of 98.83 % and a training accuracy of 98.71 %. On cross-validation, the lightweight Sklearn model achieved a mean accuracy of almost 99 %. And with a SHAP weightage magnitude of 1.38, 0.91, and 0.44 for Class ‘Good’, Class ‘Poor’, and Class ‘Bad’, respectively, patients SpO2 level is the most significant feature.

Machine Learning Predicts Peripherally Inserted Central Catheters-Related Deep Vein Thrombosis Using Patient Features and Catheterization Technology Features.

This study aims to use patient feature and catheterization technology feature variables to train the corresponding machine learning (ML) models to predict peripherally inserted central catheters-deep vein thrombosis (PICCs-DVT) and analyze the importance of the two types of features to PICCs-DVT from the aspect of "input-output" correlation. To comprehensively and systematically summarize the variables used to describe patient features and catheterization technical features, this study combined 18 literature involving the two types of features in predicting PICCs-DVT. A total of 21 variables used to describe the two types of features were summarized, and feature values were extracted from the data of 1,065 PICCs patients from January 1, 2021 to August 31, 2022, to construct a data sample set. Then, 70% of the sample set is used for model training and hyperparameter optimization, and 30% of the sample set is used for PICCs-DVT prediction and feature importance analysis of three common ML classification models (i.e. support vector classifier [SVC], random forest [RF], and artificial neural network [ANN]). In terms of prediction performance, this study selected four metrics to evaluate the prediction performance of the model: precision (P), recall (R), accuracy (ACC), and area under the curve (AUC). In terms of feature importance analysis, this study chooses a single feature analysis method based on the "input-output" sensitivity principle-Permutation Importance. For the mean model performance, the three ML models on the test set are P = 0.92, R = 0.95, ACC = 0.88, and AUC = 0.81. Specifically, the RF model is P = 0.95, R = 0.96, ACC = 0.92, AUC = 0.86; the ANN model is P = 0.92, R = 0.95, ACC = 0.88, AUC = 0.81; the SVC model is P = 0.88, R = 0.94, ACC = 0.85, AUC = 0.77. For feature importance analysis, Catheter-to-vein rate (RF: 91.55%, ANN: 82.25%, SVC: 87.71%), Zubrod-ECOG-WHO score (RF: 66.35%, ANN: 82.25%, SVC: 44.35%), and insertion attempt (RF: 44.35%, ANN: 37.65%, SVC: 65.80%) all occupy the top three in the ML models prediction task of PICCs-DVT, showing relatively consistent ranking results. The ML models show good performance in predicting PICCs-DVT and reveal a relatively consistent ranking of feature importance from the data. The important features revealed might help clinical medical staff to better understand and analyze the formation mechanism of PICCs-DVT from a data-driven perspective.

Prediction models of locally advanced rectal cancer prognosis incorporating perioperative longitudinal clinical data: A retrospective longitudinal cohort study.

e15624 Background: For locally advanced rectal cancer, neoadjuvant chemoradiation(CCRT) followed by surgery is a standard treatment. Clinical T/N stage, whether pathological complete remission (pCR) after neoadjuvant CCRT or not are well recognized tumor features that predict disease prognosis. Although the parameters that determine pathologic stage are the strongest predictors of postoperative outcome, other clinical, molecular, and histologic features may influence prognosis independent of stage. The aim is to compare the predictive efficiency, between different models for overall survival in rectal cancer patients. Methods: This retrospective study is approved by TMU-JIRB(Protocol Number: N201809032).This is also a single institute study. 175 newly-diagnosed patients with locally advanced rectal cancer were retrospectively enrolled between Jan.2011 and Dec.2016. Basic information include sex, age, body weight, body height, previous medical history were recorded. CEA level, chemotherapy regimen, type of surgery, and distance between tumor and anal verge were also collected. Chemotherapy regimen of CCRT including 5FU+LV, capecitabine (Xeloda), FOLFOX, and FOLFIRI are recorded. Surgical resection was then followed 4-8 weeks after completion of CCRT. The 2010 AJCC tumor response grading (TRG) system was used to report treatment response of proctectomy samples. The progression free survival (PFS) and overall survival (OS) are then calculated. PFS is defined as survival without any recurrence or distant metastasis before Dec 31st, 2023. A comprehensive evaluation was conducted by incorporating all these machine learning models together. The aim was to compare their predictive efficiency, with models such as Random Forest, Naïve Bayes, KNeighbor, SVM, Decision Tree, Logistic Regression, XGBoost, and LightGBM. Results: 175 patients (104 men and 71 women) were enrolled in this study. The median follow up time was 7.1 yrs. The BMI was 24.05 ± 0.03 for men and 24.20 ± 0.07 for women. 32 (17.3%) patients achieved pCR, TRG 0. Clinical staging cT3 (78.8%) and nodal positive (78.8%) are the most common tumor feature in this study group. The median age of male and female patient is 58.6 and 56.1 respectively. Upon applying various machine learning models, different predictive accuracies emerged. The Random Forest, KNeighbor, and SVM models all achieved an accuracy of 0.9250. Naïve Bayes had an accuracy of 0.8620, while the SVC model registered at 0.9310, and XGBoost was slightly behind at 0.9286. SVC model is marking the best-performing model in this evaluation. Conclusions: This multicenter retrospective data-driven deep learning prediction model informed that SVC model yielded the best predictive results. Further prospective clinical trials and external validation of this prediction model are needed to further reinforce our findings.

Enhancing Invoice Number Retrieval in Influencer Agency Transaction Records Using TF-IDF and SVC Model

This research explore the need for efficient invoice management in influencer marketing. The proposed solution using TF-IDF and SVC Model to effectively locate and handle crucial transaction information, including dates, brand names, invoice numbers, amounts, frequency and category. By using this approach to significantly improve the accuracy of identifying invoice numbers in influencer agency records. For the result this method are evaluated using precision accuracy and F1 Score metrics, to emphasize its effectiveness in current challenges of manual extraction. Overall, this paper offers practical solution to enhanced influencer agency transactions

Sleep Quality Analysis Based on Logistical, SVC and Xgboost

As a matter of fact, a healthy lifestyle requires sleep quality to be one of the main ingredients. It's a common practice to spend roughly a third of the lives sleeping, or at the very least, one should. Regrettably, that goal is not met by many of human beings in recent years. With this in mind, this study will investigate to build three models LogisticRegression Model, XGBClassifier Model, SVC Model using python technique. From the acquired data the model is used to find out some lifestyle effects on sleep. XGBClassifier Model is selected as the best model for the study. The model concluded that Occupation has the greatest effect on insomnia. Sleep apnea is most affected by BMI Category. BMI Category, Stress Level, Blood Pressure, Sleep Duration have more influence on sleep. These results show that the quality of sleep can be improved in terms of BMI Category and Blood Pressure.

Neck stabilization through sensory integration of vestibular and visual motion cues.

To counteract gravity, trunk motion, and other perturbations, the human head-neck system requires continuous muscular stabilization. In this study, we combine a musculoskeletal neck model with models of sensory integration (SI) to unravel the role of vestibular, visual, and muscle sensory cues in head-neck stabilization and relate SI conflicts and postural instability to motion sickness. A 3D multisegment neck model with 258 Hill-type muscle elements was extended with postural stabilization using SI of vestibular (semicircular and otolith) and visual (rotation rate, verticality, and yaw) cues using the multisensory observer model (MSOM) and the subjective vertical conflict model (SVC). Dynamic head-neck stabilization was studied using empirical datasets, including 6D trunk perturbations and a 4 m/s2 slalom drive inducing motion sickness. Recorded head translation and rotation are well matched when using all feedback loops with MSOM or SVC or assuming perfect perception. A basic version of the model, including muscle, but omitting vestibular and visual perception, shows that muscular feedback can stabilize the neck in all conditions. However, this model predicts excessive head rotations in conditions with trunk rotation and in the slalom. Adding feedback of head rotational velocity sensed by the semicircular canals effectively reduces head rotations at mid-frequencies. Realistic head rotations at low frequencies are obtained by adding vestibular and visual feedback of head rotation based on the MSOM or SVC model or assuming perfect perception. The MSOM with full vision well captures all conditions, whereas the MSOM excluding vision well captures all conditions without vision. The SVC provides two estimates of verticality, with a vestibular estimate SVCvest, which is highly effective in controlling head verticality, and an integrated vestibular/visual estimate SVCint which can complement SVCvest in conditions with vision. As expected, in the sickening drive, SI models imprecisely estimate verticality, resulting in sensory conflict and postural instability. The results support the validity of SI models in postural stabilization, where both MSOM and SVC provide credible results. The results in the sickening drive show imprecise sensory integration to enlarge head motion. This uniquely links the sensory conflict theory and the postural instability theory in motion sickness causation.

Text Classification from positive and unlabeled examples using Support Vector Machine (SVM)

Support Vector Machines (SVMs) are a powerful machine learning algorithm that can be used for text classification. Traditional SVMs require both positive and negative examples to train the model. However, in many real-world scenarios, it can be difficult or expensive to obtain negative examples. This study explores the application of SVMs in textclassification when only positive and unlabeled examples are available. Theresults showed that the proposed approach achieved competitive performance compared to traditional supervised methods, even when trained on limited labeled examples. The utilization of SVC in the proposed approach is twofold. First, the SVC model is used to classify theunlabeled examples as positive or negative. Second, the SVC model is used to select the positive examples that are added to the training set. Thisiterative process of training and selecting examples helps to improve the classification accuracy of the SVM model. The proposed approach is a promising method for text classification when only positive and unlabeled examples are available. Theapproach is effective in achieving competitive performance compared to traditional supervised methods, even when trained on limited labeled examples. This work contributes to enhancing text classification techniques, particularly in situationswith resource constraints and challenging label acquisition. Keywords: Support Vector Machine(S VM), Text Classifications ,Text Mining, SVC, Supervised Methods

Comparative Performance of Data Mining Techniques for Cyberbullying Detection of Arabic Social Media Text

Cyberbullying has spread like a virus on social media platforms and is getting out of control. According to psychological studies on the subject, the victims are increasingly suffering, sometimes to the point of committing suicide among the victims. The issue of cyberbullying on social media is spreading around the world. Social media use is growing, and it can have useful and negative implications when you take into account how social media platforms are abused through different forms of cyberbullying. Although there is a lot of cyberbullying detection in English, there are few studies in the Arabic language. Data Mining techniques are often used to solve and detect this problem. In this study, different data mining algorithms were used to detect cyberbullying in Arabic texts.. Our study was conducted The Bullying datasets consisted of 26,000 comments written in Arabic and were collected from kaggle.com, the Cyber_2021 dataset consisted of 13,247 comments collected via github.com, and the Data 2022 dataset consisted of 47,224 comments collected via Instagram. Various extraction features CountVectorizer and Tf-Idf were used Accuracy, precision, recall, and the F1 score were used to evaluate classifier performance. In the study, Bagging Classifier achieve high results of Bullying dataset from Kaggle Accuracy 96.04, F1-Score 95.98, Recall 96.04, Precision 95.95, SVC model gave the highest results of Cyber_2021 dataset from Github an Accuracy 98.49, F1-Score 98.49, Recall 98.49, Precision 98.50, while Data 2022 dataset from (Instagram) achieving an Accuracy of 77.51, F1-Score 76.60, Recall 77.51, and Precision 77.24. Were achieved for Tf-Idf Vectorizer. Tf-Idf Vectorizer the best to all results than count Vectorizer .

Breast Cancer Forecasting Using Machine Learning Algorithms

Some of the most prevalent and significant causes for malignancies in women is breast cancer. It is presently a widespread health problem, and it has recently become more frequent. The greatest method for managing breast cancer symptoms is early identification. The only kind of cancer that primarily affects women globally is breast cancer, which has the potential to be a major cause of mortality. Early detection of breast cancer is crucial in order to properly treat it and save many lives. This paper covers the results and analyses of several machine learning algorithms for identifying breast cancer. Several machine learning models used the information once it was analyzed. In this paper the Random forests and SVC algorithms were applied and compare the performance of these algorithms. The dataset was taken from UCI repository. Analyze and compare the classifiers' performance in terms of accuracy, precision, and f1-Score in addition. For implementing the ML algorithms, the dataset was split among training and testing phases. The notebook application Jupyter was used to implement these models. When compared to the other two models, it was successfully proven that the SVC model offers the best results. SVC's accuracy of 93% is greater than the method described earlier in that regard. The method used by this model, which will classify cancer into benign and malignant categories, yields the best results.

Influence of power oscillation damping assets reactive power capacity on damping low-frequency power system oscillations

Over the past ten years, the paradigm has shifted from conventional power generation to renewable generation. Large integration of these renewable energy sources (RES) into the power system poses challenges to system operators, leading them to put stringent requirements for their grid connection. Displacement of synchronous generators by RES reduces system inertia and consequently decreases the system damping capability of electromechanical oscillations. Poorly damped interarea oscillations reduce the transmission line's capacities and may damage power system components. Hence, future grid codes will require wind and solar power plants to provide damping to the system. Several papers have proposed adding an auxiliary damping controller to the wind turbine control algorithm to damp the low-frequency oscillations (LFOs) by modulating active or reactive power. However, these studies have not mentioned if small power plants can damp LFOs in a multimachine system. Therefore, this paper investigates the influence of reactive power capacity on the damping of LFOs and its effects on optimal controller parameters using a simplified SVC model connected at the midpoint of the tie line of a two-area test system. A local feedback signal is selected as the input signal to the SVC damping controller. Controller parameters are optimized using the particle swarm optimization algorithm. Time-domain simulations performed in PowerFactory software demonstrate the damping behavior of the controller at different SVC ratings. The results show a minimum reactive power capacity is required for effective damping of power system oscillations.

Collateral circulation detected by arterial spin labeling predicts outcome in acute ischemic stroke.

Robust collateral circulation is strongly associated with good outcomes in acute ischemic stroke (AIS). To determine whether collateral circulation detected by arterial spin labeling (ASL) magnetic resonance imaging could predict good clinical outcome in AIS patients with 90 days follow-up. Total 58 AIS patients with anterior circulation stroke were recruited. Collateral circulation was defined as arterial transit artifact in ASL images. Modified Rankin Scale (mRS), the Barthel Index, and National Institutes of Health Stroke Scale (NIHSS) were employed to evaluate neurological function for the baseline and 90 days follow-up. The percent changes of these scores were also calculated, respectively. Finally, a support vector classifier model of machine learning and receiver operating characteristic curve were employed to estimate the power of ASL collaterals (ASLcs) predicting the clinical outcome. Patients with ASLcs represented higher rate of good outcome (83.30% vs. 31.25%, p < .001) and lower follow-up mRS scores (p < .001), when compared to patients without ASLcs. There were significant differences for percent changes of mRS scores and NIHSS scores between these two groups. Further, the presence of ASLcs could predict good clinical outcome (OR, 1.54; 95% CI, 1.10-2.16), even after controlling for baseline NIHSS scores. The SVC model incorporating baseline NIHSS scores and ASLcs had significant predictive effect (accuracy, 79.3%; AUC, 0.806) on clinical prognosis for AIS patients. We targeted on the non-invasive assessment of collateral circulation using ASL technique and found that patients with ASLcs were more likely to have a good clinical outcome after AIS. This finding is of guiding significance for treatment selection and prognostic prediction. Early ASLcs assessment provides a good powerful tool to predict clinical outcome for AIS patients with 90 days follow-up.

Exploring regional and sociodemographic disparities associated with unenrollment for the disease management program for type 2 Diabetes Mellitus using Bayesian spatial modelling

BackgroundThe disease management program (DMP) for type 2 Diabetes Mellitus (T2DM) is the largest DMP in Germany. Our goal was to analyze regional differences in unenrollment rates, suggest areas for intervention and provide background information, which population groups in which locations are currently not enrolled in the DMP for T2DM.MethodsIn this study, we used data of the 1.7 mil. insurants of the AOK Nordost health insurance. For the visualization of enrollment potential, we used the Besag-York-Mollie model (BYM). The spatial scan statistic (SaTScan) was used to detect areas of unusually high rates of unenrolled diabetics to prioritize areas for intervention. To explore sociodemographic associations, we used Bayesian spatial global regression models. A Spatially varying coefficient model (SVC) revealed in how far the detected associations vary over space.ResultsThe proportion of diabetics currently not enrolled in the DMP T2DM was 36.8% in 2019 and varied within northeastern Germany. Local clusters were detected mainly in Mecklenburg-West-Pomerania and Berlin. The main sociodemographic variables associated with unenrollment were female sex, younger age, being unemployed, foreign citizenship, small household size and the proportion of persons commuting to work outside their residential municipality. The SVC model revealed important spatially varying effects for some but not all associations.ConclusionLower socioeconomic status and foreign citizenship had an ubiquitous effect on not being enrolled. The DMP T2DM therefore does currently not reach those population groups, which have a higher risk for secondary diseases and possible avoidable hospitalizations. Logically, future interventions should focus on these groups. Our methodology clearly suggests areas for intervention and points out, which population group in which locations should be specifically approached.

An ultra-short-term wind speed forecasting model based on time scale recognition and dynamic adaptive modeling

The forecast of wind speed is prerequisite for wind power prediction, which is one of the most effective means of promoting wind power absorption. However, when modeling for wind speed sequences with different fluctuations, most existing researches ignore the influence of time scale of wind speed fluctuation period, let alone the low compatibility between training and testing samples that severely limit the training performance of forecasting model. To improve the accuracy of wind speed and wind power forecasting, an ultra-short-term wind speed forecasting model based on time scale recognition and dynamic adaptive modeling is proposed in this paper. First, a series of wind processes are divided from the historical wind speed sequence according to the natural variation characteristics of wind speed. Second, we divide all the wind processes into two patterns based on their time scale, and an SVC model with input features extracted from meteorological data is built to identify the time scale of the current wind process. Third, for a specifically identified wind process, the complex network algorithm is applied in data screening to select high compatible training samples to train the forecast model dynamically for current input. Simulation indicates that the proposed approach presents higher accuracy than benchmark models using the same forecasting algorithms but without considering the time scale and data screening.

Performance evaluation of a state-of-the-art automotive radar and corresponding modeling approaches based on a large labeled dataset

Radar is a key sensor to achieve a reliable environment perception for advanced driver assistance system and automated driving (ADAS/AD) functions. Reducing the development efforts for ADAS functions and eventually enabling AD functions demands the extension of conventional physical test drives with simulations in virtual test environments. In such a virtual test environment, the physical radar unit is replaced by a virtual radar model. Driving datasets, such as the nuScenes dataset, containing large amounts of annotated sensor measurements, help understand sensor capabilities and play an important role in sensor modeling. This article includes a thorough analysis of the radar data available in the nuScenes dataset. Radar properties, such as detection thresholds, and detection probabilities depending on object, environment, and radar parameters, as well as object properties, such as reflection behavior depending on object type, are investigated quantitatively. The overall detection probability of the considered radar (Continental ARS-408-21) was found to be 27.81%. Four radar models on object level with different complexity levels and different parametrisation requirements are presented: a simple RCS-based radar model with an accuracy of 51%, a linear SVC model with an accuracy of 70%, a Random Forest model with an accuracy of 83%, and a Gradient Boost model with an accuracy of 86%. The feature importance analysis of the machine learning algorithms revealed that object class, object size, and object visibility are the most important parameters for the presented radar models. In contrast, daytime and weather conditions seem to have only minor influence on the modeling results.

Closed Loop Reactive Power Compensation on a Single-Phase Transmission Line

Nowadays the power consumption is increasing significantly. Due to this, the transmission lines are heavily loaded and in turn results in instability of the line. So orderly transmission of power is essential. For the proper operation of the load the voltage has to be maintained within the acceptable limit. Due to the changes in load, the voltage level of the line also changes. The voltage levels can be improved by using the reactive power compensator like capacitor banks, series compensator, STAT COM and svc. Here, we considered the SVC as reactive power compensator. This paper presents the reactive power compensation for a 230v transmission line model with variable inductive load using SVC and designed in Simulink. Here we considered the SVC model which contains of anti-thyristors in series to the inductor and a shunt capacitor. The required reactance is fed from the inductor to the transmission line by changing the gate triggering pulse to the thyristor pair. So, the main disadvantage of the open loop var compensation using svc is the triggering to the thyristor pair has to be changed every time depending on the load. This can overcome by implementing it using the feedback loop. Here for the feedback, we used the PI controller. under The PI controller gives the required gating pulse to the thyristor pair. The performance of this model is studied variable load conditions.

373 Can we identify intra-hepatic cholestasis of pregnancy by clinical parameters using machine learning techniques?

To apply machine learning models on clinical and laboratory features of women with intrahepatic cholestasis of pregnancy (ICP) and to create an algorithm to identify these patients without bile acid measurements. This retrospective study included 336 women with low-risk pregnancies and a chief complaint of pruritis without rash during the second and third trimesters. Data extracted included: (1) Demographic: age, BMI, ethnicity; (2) Obstetric: gravidity, parity, gestational week, singleton/multiple pregnancy; (3) Clinical: Season, duration and location of pruritus, previous ICP; (4) Laboratory: GOT, GPT levels, bile acid levels. The primary outcome was an elevated bile acid measurement ≥10 μmol/L, regardless of liver enzyme levels (GOT/GPT). We used different machine learning models and statistical regression to predict elevated bile acid levels. Among 336 women who complained about pruritis, 167 had bile acids ≥10 μmol/L and 169 had normal bile acid levels. Women with elevated bile acids tended to be older than those with normal levels (33.1±6.1 vs. 30.3±5.3 years), with higher parity (p=0.001) and higher GOT (77±79 vs. 29±25, p=0.001) and GPT levels (115±140, vs. 32±43, p=0.001). Using machine learning models, the XGB Classifier model was the most accurate (AUC, 0.9) followed by the K-neighbors model (AUC, 0.86); and then the SVC model (AUC, 0.81). The model with the lowest predicative ability was the logistic regression (AUC, 0.76). The maximum sensitivity of the XGB model was 86% and specificity 75%. The best predictive parameters of the XGB model were elevated GOT (Importance 0.17), elevated GPT (Importance 0.16), family history of bile disease (0.16) and previous pregnancy with ICP (0.13). Machine learning models using clinical data may predict ICP more accurately than logistic regression does. Using detection algorithms derived from these techniques may improve identification of ICP, especially when bile acid testing is not available.

Extended SVC Modeling for Frequency Regulation

This letter proposes an extended SVC model with thyristor firing angle law to improve transient performances of frequency and voltage of multi-machine power systems. A firing angle control synthesis framework for imposing any desired SVC susceptance value without linearization around a fixed operation point is proposed. In light of this method, the designed firing angle controller only requires local bus measures and parameters. The performance of the extended firing angle-controlled SVC model is evaluated on a 150-bus 27-machine power system and compared with conventional SVC model.

A semismooth Newton method for support vector classification and regression

Support vector machine is an important and fundamental technique in machine learning. In this paper, we apply a semismooth Newton method to solve two typical SVM models: the L2-loss SVC model and the \epsilon-L2-loss SVR model. The semismooth Newton method is widely used in optimization community. A common belief on the semismooth Newton method is its fast convergence rate as well as high computational complexity. Our contribution in this paper is that by exploring the sparse structure of the models, we significantly reduce the computational complexity, meanwhile keeping the quadratic convergence rate. Extensive numerical experiments demonstrate the outstanding performance of the semismooth Newton method, especially for problems with huge size of sample data (for news20.binary problem with 19996 features and 1355191 samples, it only takes three seconds). In particular, for the \epsilon-L2-loss SVR model, the semismooth Newton method significantly outperforms the leading solvers including DCD and TRON.

The Rapid Determination Method of Performance Parameters and PG Grade for SBS Modified Asphalt

This paper demonstrates the application of support vector machine (SVM) applied to ATR-FTIR data to solve classification and regression problems associated with rapid determination of PG grade for SBS modified asphalt. The modified asphalt samples were produced by mixing three kinds of matrix asphalt, two kinds of SBS and five kinds of SBS content from 2% to 6% under the same manufacture process. Therefore a total of 150 data sets were evaluated with five parallel tests for each sample. In the SVR model, the ATR-FTIR data were parameter for the input layer whereas the PG parameters of asphalt such as, rut factor, creep stiffness and creep rate were output layer. While in the SVC model, high temperature grade and low temperature grade were output layer. This new method allows rapid determination of multi-properties from a single spectrum for SBS modified asphalt and it is promising for online material monitoring of specific project.