2,979 publications from this institution
In this paper, we propose a novel architecture of polynomial neural network classifier (PNNC) with the aid of data preprocessing technique and space search optimization, which adopts accelerated convergence mechanism instead of purely random search. Two type of polynomials are adopted for constructing discriminate functions in the PNNC to alleviate the limitation of relatively simple geometry using linear discriminate function in the conventional neural network classifiers. Space search optimization is exploited here to realize structure optimizes and parameter optimize in the design of PNNC. Moreover, data preprocessing techniques are used to reduce the dimension of training data. The proposed PNNC is compared with some well-known classifiers based on several benchmark data sets. Experimental results illustrate the effectiveness of PNNCs.
Fuzzy cognitive maps (FCMs) are a powerful and convenient tool for describing and analysing dynamic systems. Their generic design is performed manually, exploits expert knowledge and is quite tedious, especially in the case of larger systems. This shortcoming is alleviated by completing the design of FCMs through learning carried out on experimental data. Comprehensive experiments reveal that this approach helps design models of required accuracy in an automated manner.
Purpose – In a risk analysis system, different underlying indices often play different roles in identifying the risk scale of the total target in a system, so a concept of discriminatory weight is introduced first. With the help of discriminatory weight and membership functions, a new method for information security risk analysis is proposed. The purpose of this paper is to discuss the above issues. Design/methodology/approach – First, a concept of discriminatory weight is introduced. Second, with the help of fuzzy sets, risk scales are captured in terms of fuzzy sets (namely their membership functions). Third, a new risk analysis method involving discriminatory weights is proposed to realize a transformation from the membership degrees of the underlying indices to the membership degrees of the total target. At last, an example of information security risk analysis shows the effectiveness and feasibleness of the new method. Findings – The new method generalizes the weighted-average method. The comparative analysis done with respect to other two methods show that the proposed method exhibits higher classification accuracy. Therefore, the proposed method can be applied to other risk analysis system with a hierarchial. Originality/value – This paper proposes a new method for information security risk analysis with the help of membership functions and the concept of discriminatory weight. The new method generalizes the weighted-average method. Comparative analysis done with respect to other two methods show that the proposed method exhibits higher classification accuracy in E-government information security system. What is more, the proposed method can be applied to other risk analysis system with a hierarchial.
Efficient and robust data clustering remains a challenging task in the field of data analysis. Recent efforts have explored the integration of granular-ball (GB) computing with clustering algorithms to address this challenge, yielding promising results. However, existing methods for generating GBs often rely on single indicators to measure GB quality and employ threshold-based or greedy strategies, potentially leading to GBs that do not accurately capture the underlying data distribution. To address these limitations, this article introduces a novel GB generation method. The originality of this method lies in leveraging the principle of justifiable granularity to measure the quality of a GB for clustering tasks. To be precise, we define the coverage and specificity of a GB and introduce a comprehensive measure for assessing GB quality. Utilizing this quality measure, the method incorporates a binary tree pruning-based strategy and an anomaly detection method to determine the best combination of sub-GBs for each GB and identify abnormal GBs, respectively. Compared to previous GB generation methods, the new method maximizes the overall quality of generated GBs while ensuring alignment with the data distribution, thereby enhancing the rationality of the generated GBs. Experimental results obtained from both synthetic and publicly available datasets underscore the effectiveness of the proposed GB generation method, showcasing improvements in clustering accuracy and normalized mutual information.
Most existing data preprocessing is done at the CPU. Although some studies use techniques such as multi-processing and double buffering to accelerate CPU preprocessing, CPU computational speed and storage bandwidth still limit the processing speed. Other studies try to use intelligent data storage devices, such as computational storage devices, to complete data preprocessing instead of CPUs. The current studies use only one device to complete data preprocessing operations, which cannot fully overlap data preprocessing and accelerator computation time. To fully exploit the independence and high bandwidth of the novel CSD, this paper proposes an advanced, highly parallel dual-pronged data preprocessing algorithm (DDLP) that significantly improves the execution efficiency and computational overlap between heterogeneous devices. DDLP enables the CPU and CSD to start data preprocessing operations from both ends of the dataset separately. Meanwhile, we propose two adaptive dynamic selection strategies to make DDLP control the GPU to automatically read data from different sources. We achieve sufficient computational overlap between CSD data preprocessing and CPU preprocessing, GPU computation, and GPU data reading. In addition, DDLP leverages GPU Direct Storage technology to enable efficient SSD-to-GPU data transfer. DDLP reduces the usage of expensive CPU and DRAM resources, reduces the number of SSD-to-GPU data transfers, and improves the energy efficiency of preprocessing while reducing the overall preprocessing and training time. Extensive experimental results show that DDLP can improve learning speed by up to 23.5% on ImageNet Dataset while reducing energy consumption by 19.7% and CPU and DRAM usage by 37.6%. DDLP also improve learning speed by up to 27.6% on Cifar-10 Dataset.
