scispace - formally typeset
Search or ask a question

Answers from top 6 papers

More filters
Papers (6)Insight
Open accessJournal ArticleDOI
Wen Fenghua, Xiao Jihong, HE Zhifang, Gong Xu 
50 Citations
The empirical evidence shows that, compared with the SVM without these price features, the combination predictive methods-the EEMD-SVM and the SSA-SVM, which combine the price features into the SVMs perform better, with the best prediction to the SSA-SVM.
Experimental results show that SVM model is marginally superior to CART with DA, being more robust than its other counterparts.
The results of experimental transactions show the advantages of using SVM model compared to the transactions without using SVM model.
Proceedings ArticleDOI
Thuy Nguyen Thi Thu, Vuong Dang Xuan 
11 May 2018
19 Citations
The experimental results show the advantages of use SVM compared to the transactions without use SVM ones.
Accordingly, this study compares the ANN, LR, SVM, Bagging SVM, Boosting SVM techniques and experience shows that the new SVM based ensemble model can be used as an alternative method for credit assessing.
The results show that the proposed SVM model outperforms NN and PCA and the merits of using SVM to mitigate the limitations of using NN are elaborated.

Related Questions

How safe is the use of e-wallet payment method?5 answersThe safety of using e-wallets as a payment method is a crucial concern addressed in various research papers. E-wallets offer security features such as encryption keys for card data transmission, which significantly enhance the safety of electronic transactions. Factors influencing the security of digital transactions include ease of use, history, and notification, impacting the intention to use QRIS. Additionally, the evolution from cash payments to digital wallets has made transactions more secure against theft, although challenges like wrong payments and link failures persist. Integrating digital wallets with blockchain technology can provide a more secure payment system by ensuring privacy, permission, and secure data transmission through a peer-to-peer network. Overall, advancements in technology and security measures contribute to making e-wallet payments a safe and reliable option for users.
How can payment apps be made more secure?5 answersPayment apps can be made more secure by implementing various measures. Firstly, it is important to have standardized protocols and secure deployment guidelines. This ensures that all payment apps follow a consistent and secure approach. Additionally, incorporating strong authentication methods such as fingerprint identification, face recognition, and iris matching can enhance security. Matching authenticated user information associated with SIM cards can also prevent unauthorized access. Furthermore, addressing implementation flaws and vulnerabilities, such as insecure backend SDKs, is crucial. Regular security analysis and testing can help identify and fix any exploits or weaknesses. Finally, providing users with secure payment guidelines and steps can help them make informed decisions and protect their credentials. By implementing these measures, payment apps can enhance security and protect user information and transactions.
Is this website safe?4 answersThe invention provides a website password safe input method that utilizes the safety of telecom voice communication networks to avoid password leakage and enhance security. The invention also discloses a website safety detection method and system that allows users to enjoy safety detection services and ensures effective website safety. An interactive website filter based on heuristics is presented, which considers the disparity between a website's true identity and its observed identity to detect malicious websites with high accuracy. A method for certifying a website to be malware-free and owned by a trusted entity is described, where URLs are validated periodically and a network-accessible document certifies the website's safety. Additionally, a method for safely logging on to a website is disclosed, which verifies the security of the server based on browsing request information and deciphered content.
Why SVM cannot handle long sequential data?4 answersSVM cannot handle long sequential data due to its high training complexity and the choice of kernel function for non-separable data. The training time for SVM is long, making it impractical for large datasets. Additionally, SVMs are not suitable for classification of large datasets because of their high training complexity. The computational complexity of SVMs is O(n^2), which further limits their ability to handle very large datasets. To address these issues, various approaches have been proposed, such as weighted Euclidean distance, radial integral kernel function SVM, and dimensionality reduction algorithms. These approaches aim to reduce the training time and improve the classification accuracy of SVMs for large datasets.
How do I make SVM more accurate?9 answers
How is SVM solved?7 answers

See what other people are reading

What are the most common machine learning algorithms used in walking behavior analysis?
4 answers
Machine learning algorithms commonly used in walking behavior analysis include logistic regression, radial-basis function support vector machine, eXtreme Gradient Boosting (XGBoost), multilayered perceptron (MLP), decision tree, and random forest. These algorithms have been applied to predict walking behavior within the next 3 hours based on previous steps-per-minute data, showing varying levels of performance. Among these algorithms, the multilayered perceptron (MLP) demonstrated the highest overall performance in terms of accuracy, Mathew correlation coefficient, sensitivity, and specificity in predicting walking behavior. Additionally, machine learning algorithms such as kNN, Random Forest, and boosting have been utilized in classifying emotions during walking sessions, showing promising results for emotion recognition based on human gait patterns.
How important is stagger angle in tonal noise predictions in fan noise?
5 answers
The stagger angle plays a crucial role in tonal noise predictions in fan noise studies. Research indicates that variations in static stagger can significantly impact the dynamic untwist behavior of fan assemblies, affecting aerodynamic performance, vibratory response, and noise production. Additionally, the interaction of fan-blade wakes with guide vanes and the fan self-noise, which are key contributors to tonal noise, are influenced by the distortion caused by circumferentially nonuniform flow near the fan. Implementing a two stagger-angle method to account for vane camber effects in noise prediction models has shown clear improvements in the analytical prediction of modal acoustic power radiated within the duct, particularly for fan tonal noise where camber effects are significant. Furthermore, studies have demonstrated that even small variations in vane stagger angle can lead to substantial changes in acoustic power and modal distribution of the sound field in annular ducts.
What are the most common challenges faced by students and professionals in Computer-Aided Design (CAD) education and training?
4 answers
Students and professionals in Computer-Aided Design (CAD) education and training encounter various challenges. For students, transitioning from face-to-face to online learning poses difficulties in design learning, although it also presents opportunities for peer learning and active class involvement. On the other hand, professionals face challenges related to collaborative design, communication, data management, and permissioning when working with distributed CAD tools, hindering effective collaboration in CAD teams. The pandemic further exacerbated these challenges, forcing educators to adapt quickly to online teaching methods, impacting the practices of design education significantly. Overall, both students and professionals grapple with adapting to new learning environments, communication barriers, and managing collaborative workflows in CAD education and training.
How does the relationship between creative self-efficacy and flexible thinking vary across different cultural contexts and demographic groups?
5 answers
The relationship between creative self-efficacy and flexible thinking varies across different cultural contexts and demographic groups. Research indicates that self-efficacy is positively correlated with task enjoyment, effort, and creative performance. Additionally, individuals with higher self-efficacy tend to exhibit greater flexibility in their thinking, which in turn enhances originality. Moreover, cultural factors such as individualism and collectivism can mediate the creative mindset between different countries, influencing creative self-efficacy. Furthermore, children demonstrate remarkable flexibility in transferring external strategies across contexts as they age, showcasing individual differences in the strategies devised. These findings suggest that cultural contexts and demographic factors play a significant role in shaping the relationship between creative self-efficacy and flexible thinking.
Are women more likely to pose at gym than men?
5 answers
Women are less likely to engage in weightlifting compared to men due to cultural dissociation and evaluation concerns associated with strength-related exercise goals. However, when it comes to posing at the gym, research on Instagram fitness influencers reveals that there are no significant gender differences in the objectification of males and females in terms of muscularity and sultry poses. Artificial intelligence and machine learning technologies are increasingly being used in the fitness industry for human pose estimation, aiding in activities like counting repetitions during weightlifting or CrossFit events. These technologies can be applied in gym trainer trackers to assist individuals in achieving their fitness goals by accurately assessing their movements and poses.
What papers have implemented RRT on dynamic targets?
5 answers
The research papers that have implemented Rapidly-exploring Random Trees (RRT) on dynamic targets include papers on radar cross sections (RCS) modeling for dynamic aircraft, dynamic target following control for autonomous vehicles, and regression-based speech end-pointing models for interactive voice assistants. These papers utilize RRT in various contexts such as simulating dynamic RCS data efficiently, controlling autonomous vehicles to follow dynamic targets smoothly, and improving speech end-pointing accuracy and latency for voice assistants. Each paper demonstrates the effectiveness of using RRT in addressing the challenges associated with dynamic targets in their respective domains.
What are some effective strategies for addressing data quality and availability challenges in policymaking?
5 answers
Effective strategies for addressing data quality and availability challenges in policymaking involve implementing quality metrics to analyze datasets, ensuring data reliability, relevance, and sufficiency, and leveraging big data resources. By focusing on the quality of open data, governments can better use, manipulate, and distribute information to citizens for informed decision-making. Additionally, utilizing big data can help fill information gaps in ecosystem services evaluation, although careful management is necessary due to certain limitations. A data-driven policy-making framework, such as the RULER criteria (Reliable, Uncoupled, Lifecycle, Environment, Relevant), can aid in evaluating and filtering data for waste management and policy-making, ensuring the collection of necessary and reliable data for informed decision-making. Policymakers must recognize the critical importance of data quality in shaping policies, as policies are direct products of the data considered during their formulation.
How does the quality of audio equipment affect one's ability to interpret verbal cues?
5 answers
The quality of audio equipment significantly impacts one's ability to interpret verbal cues. Studies have shown that audio quality affects users' evaluation of multimedia communication, with noise in audio decreasing the quality of user experience. Additionally, in the context of speech perception, audiovisual cues and sentence context can facilitate speech perception in noise, particularly in older adults. Electroencephalography (EEG) measurements have been utilized to assess cortical functional connectivity in response to perceived audio quality, demonstrating the importance of high-quality audio in brainwave response patterns. Furthermore, the design of anchor signals for assessing audio codecs' quality relies on understanding the impairments and dimensions of audio quality, emphasizing the multidimensional nature of audio quality assessment. Overall, high-quality audio equipment is crucial for accurate interpretation of verbal cues in various contexts.
How can both tDCS and tACS be used in biopsychology?
5 answers
Both transcranial direct current stimulation (tDCS) and transcranial alternating current stimulation (tACS) can be utilized in biopsychology research to modulate brain activity and cognition. tDCS is effective for behavioral investigations due to its subthreshold neuromodulation effects, allowing for the study of behavior alongside brain activity. On the other hand, tACS and tDCS have been shown to be equally effective in targeting both hemispheres and enhancing brain dynamics in high-definition configurations, which could be clinically relevant. Additionally, tDCS and tACS have been compared in terms of their effects on large-scale functional networks, with tDCS increasing functional connectivity within the default-mode network (DMN) while tACS decreased it, showcasing their differential impacts on brain activity and network dynamics. These findings highlight the versatility and potential of both tDCS and tACS in advancing biopsychological research.
How LLM can do complex human tasks?
5 answers
Large Language Models (LLMs) can handle complex human tasks by bridging the gap between humans and machines through innovative interaction frameworks. These models translate natural language task descriptions into intermediary representations for effective planning and execution. By incorporating low-code visual programming interactions, LLMs offer controllable and stable responses, enhancing user-friendly interactions and applicability across various scenarios. LLMs can replicate human-like behavior in crowdsourcing tasks, showcasing potential in complex task domains when combined with human expertise. Leveraging autoregressive re-prompting, LLMs can correct errors and improve task completion rates, outperforming traditional LLM-based planning methods in intricate environments. Overall, LLMs demonstrate promise in efficiently handling complex tasks by enabling structured planning workflows and joint task-and-motion planning algorithms.
What is the advantages of the knowledge ditillation technique of the edge deployement?
4 answers
Knowledge distillation offers significant advantages in edge deployment scenarios. By employing knowledge distillation, lightweight models with high accuracy can be generated based on full models trained in the cloud, addressing the challenge of limited onboard resources and high-performance requirements in UAV delivery services. Additionally, in Federated Edge Learning (FEL), knowledge distillation plays a crucial role in overcoming challenges related to resources, personalization, and network environments, enhancing the efficiency and effectiveness of collaborative on-device machine learning model training without compromising data privacy. Moreover, in personalized federated learning, self-knowledge distillation enables the acceleration of recalling personalized knowledge for individual clients, leading to improved model performance and robustness in edge computing environments.