Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2022-06-06 Cooperative journals: 《计算机应用研究》
Abstract： In the large data environment, there are many problems in the parallel deep convolution neural network (DCNN) algorithm, such as excessive data redundancy, slow convolution layer operation and poor convergence of loss function. This paper proposed a parallel deep convolution neural network optimization algorithm based on the Im2col method. First, the algorithm proposed a parallel feature extraction strategy based on Marr-Hildreth operator to extract target features from data as input of convolution neural network, which can effectively avoid the problem of excessive data redundancy. Secondly, the algorithm designed a parallel model training strategy based on the Im2col method. The redundant convolution kernel is removed by designing the Mahalanobis distance center value, and the convolution layer operation speed is improved by combining the MapReduce and Im2col methods. Finally, the algorithm proposed an improved small-batch gradient descent strategy, which eliminates the effect of abnormal data on the batch gradient and solves the problem of poor convergence of the loss function. The experimental results show that IA-PDCNNOA algorithm performs well in deep convolution neural network calculation under large data environment and is suitable for parallel DCNN model training of large datasets.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2022-05-11 Cooperative journals: 《计算机应用研究》
Abstract： Aiming at the problems that the Raft algorithm cannot resist the attacks of Byzantine nodes and the logs are easy to tamper with, this paper proposes an RB-Raft (Resist Byzantine-Raft) algorithm that resists Byzantine nodes. Firstly, this paper uses the method of hash chain to iteratively hash each log. At the same time, verification of the log through the dynamic verification mechanism, so that the malicious behavior of the leader node has a certain fault tolerance rate, which solves the problem of log forgery and verification. Secondly, this paper proposes a "Legacy" mechanism based on threshold encryption, which makes it legal for Candidate nodes to pull votes. This mechanism can prevent Byzantine nodes from randomly the attack of pulling votes to replace leader nodes, and solves the problem that Byzantine nodes affect the system consistency. The experimental results show that the proposed RB-Raft algorithm has the ability to resist Byzantine nodes, and its log recognition rate can reach 100%. At the same time, compared with PBFT, the consensus latency of the algorithm in this paper is reduced by 53.3%, and the throughput is increased by 61.8%. The algorithm proposed in this paper is suitable for consensus in untrusted consortium chains.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2022-04-07 Cooperative journals: 《计算机应用研究》
Abstract： Classic flower pollination algorithm (FPA) can be easily exposed to the shortcomings of local optimal solution and slow convergence velocity. In view of these shortcomings, this paper proposed an FPA with an enhanced lens imaging strategy and random neighborhood-based mutation strategy. The lens imaging strategy can help the algorithm avoid the shortcoming of local optimal solution by expanding the search space of FPA to increase the diversity of the solution. The introduction of random neighborhood-based mutation strategy can enhance the convergence accuracy and search speed of the algorithm by guiding algorithm search with information in the neighborhood. A comparison of the improved FPA with four other improved algorithms on CEC2013 test function found that the improved multi-strategy FPA performs better than the comparison algorithms in both convergence accuracy and search speed. To study its practical utility, this paper applies the multi-strategy FPA into the automobile transmission parameter model and the results indicate that multi-strategy FPA is better than the comparison algorithm in optimization of automobile transmission parameters.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Aiming at the problem of excessive time, space complexity and unbalanced load for each node based on the parallel frequent itemset mining algorithm MRPrePost, this paper proposed an optimization parallel frequent itemset mining algorithm based on MapReduce, named PFIMD. Firstly, this algorithm adopted a data structure called DiffNodeset, which effectively avoid the defect that the N-list cardinality got very large in the MRPrePost algorithm. Secondly, in order to reduce the time complexity of this algorithm, it designed the T-wcs (2-way Comparison Strategy) to avoid the invalid calculation in the procession of two DiffNodesets connection. Finally, considering the impact of cluster load on the efficiency of parallel algorithm, it proposed the LBSBDG (Load Balancing Strategy Based on Dynamic Grouping) , which decreased the size of PPC-Tree on each computing node and reduced the amount of time required to traverse the PPC-Tree by evenly grouping each item in the F-list. The experimental results show that the modified algorithm has better performance on mining frequent itemset in a big data environment.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-05-10 Cooperative journals: 《计算机应用研究》
Abstract： To solve the problem of low efficiency of large data clustering, this paper proposes a fast grid density clustering algorithm SGBSCAN(Square-neighborhood and Grid-based DBSCAN) . Firstly, this paper gave the definition of square neighborhood density clustering , and used the square neighborhood instead of the circular neighborhood to reduce the time complexity. Secondly, this paper proposed the concept of grid of square neighborhood density clustering , to determine the density relationship between core points and data points in high density region quickly. Finally, this paper proposed the Grid density cluster, the method used the relationship between the grid to accelerate the formation of density clusters. The algorithm made 16 data sets and compared with the existing literature algorithms. The results shows that the algorithm has a significant improvement in clustering efficiency. The larger the data volume, the more obvious the efficiency of the algorithm, and the algorithm is suitable for multidimensional data clustering.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-04-01 Cooperative journals: 《计算机应用研究》
Abstract： Aiming at the problem that the accuracy and recall of the protein complexes identification algorithm based on ant colony and fuzzy C-means (FCM) clustering are not high and the running efficiency is low, this paper proposed a novel protein complex recognition algorithm named FAC-PC (algorithm for identifying weighted protein complexes based on fuzzy ant colony clustering) . Firstly, combing with the Pearson correlation coefficient and edge aggregation coefficient, it constructed the weighted protein network. Secondly, in order to overcome the defects of massive merger, filter, repeated pick-up and drop-down operations in ant colony clustering algorithm, it designed the EPS (essential protein selection) metric to select essential protein, and designed the PFC (protein fitness calculation) metric to traverse neighbors of essential proteins to obtain essential group proteins, then the essential group protein replaced the seed node in the process of ant colony clustering, which improved results that the accuracy and time performance. Furthermore, it proposed the SI (similarity improvement) metric to optimize the probability of picking and dropping operations of ant colony to obtain the number of clustering. Finally, according to the improved ant colony algorithm, it obtained the essential protein and the number of clustering to initialize the FCM algorithm, and designed the membership update strategy to optimize the membership update, at the same time, a new FCM objective function which took a balance between intra-clustering and proposed inter-clustering variation, finally identified the protein complex by improved FCM algorithm. It used FAC-PC algorithm to identify protein complexes on DIP data. The experimental results show that FAC-PC algorithm has better performance on accuracy and recall, which is more reasonable to identify protein complexes.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-01-03 Cooperative journals: 《计算机应用研究》
Abstract： The recommendation system needs to utilize a large amount of user data, which may expose the user's preferences and pose a huge challenge to the privacy concerns. To ensure the accuracy of recommendation and user privacy, this paper proposed a matrix factorization recommendation model combining differential privacy and tag information. Firstly, the model added the tag information to the process of calculating item similarity, then integrated it into the recommendation model to improve the recommendation accuracy. Finally, this paper solved the model optimal value by the stochastic gradient descent method. For protecting users from privacy threats, the proposed approach divided Laplace noise into two parts, which are added to the process of item similarity and gradient solution respectively, so that the whole recommendation process satisfied the differential privacy, and analyzed the validity of the algorithm on a real data set. Experimental results show that the proposed method has high recommendation accuracy while protecting users’ privacy.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-01-03 Cooperative journals: 《计算机应用研究》
Abstract： There exists several issues in traditional collaborative filtering algorithms: a) It has the sparsity of user rating matrix; b) It ignores the relationship between item attributes. Considering all these problems, this paper proposed a novel collaborative filtering algorithm combining score ratio factor and item attribute. The algorithm used the scoring matrix to obtain the ratio matrix of common and non-common score users between items. Therefore, it increased the influence degree of the users of the item common score, and reduced the error caused by the sparsity of the user-item scoring matrix on the item similarity calculation. quantifying the item attribute could obtain the weight of the item similarity, and it also improved the accuracy of the item similarity calculation. According to the above two points, an algorithm combining scoring scale factor and item attribute weight as item similarity weight is proposed. Experimental results show that, it improved the recall rate and accuracy of the algorithm by 5.1% and 4.7% respectively compared with the existing methods. The algorithm is suitable for personalized recommendation of e-commerce websites.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-12-13 Cooperative journals: 《计算机应用研究》
Abstract： Since static PPI networks are difficult to truly reflect the dynamic character of cells, the convergence speed is slow, cluster precision and recall is low in mining protein complex based on ant colony clustering, this paper proposes an ant colony clustering algorithm based on fuzzy granular and closeness degree to mine protein complexes in dynamic weighted PPI network, named FGCDACC-DPC. First, based on the topological and biological characteristics of the PPI network, a comprehensive weight metric (CWM) is designed to accurately describe the interaction between proteins. Second, this method constructs a series of dense and highly co-expressed complex core based on the basic characteristic of the complexes, then it employs the picking and dropping operations, which based on fuzzy granular and closeness degree, to cluster the nodes in PPI networks, in order to reduce effectively the computational complexity and randomness, speed up the clustering speed. Finally, this algorithm designs a local and global strategy founded on function transmission and timing functional relevance theory for weight’s update, which achieve the function transmission between different generations of ant colonies and networks at different times to effectively improve clustering accuracy. FGCDACC-DPC algorithm is used to mine protein complexes on DIP data. Experimental results demonstrate that this algorithm has better performance on precision and recall, which is more reasonable to identify protein complexes.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-08-13 Cooperative journals: 《计算机应用研究》
Abstract： Many applications have more functions than their types, and they need to acquire more permissions. Excessive permissions may bring some security risks. To address these issues, this paper proposes an Android malware detection method based on meta information. First, through the description of Android application of LDA theme extraction, the implementation of data dimensionality reduction, using the k-means clustering algorithm in accordance with the functional type of the application group; Then, for all applications belonging to the same functional type, extract their permission information, and take the permission features as the research object, using KNN algorithm to classify and detect the malicious software of Android. The experimental results obtained the average accuracy of 94.81% and proved the validity and high accuracy of the method.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-20 Cooperative journals: 《计算机应用研究》
Abstract： Since the rainfall and other uncertainties are difficult to obtain and effectively deal with in landslide hazard prediction, and the existence of setting density threshold and high time complexity in the OPTICS-PLUS algorithms, in order to improve the prediction accuracy, this paper proposed an uncertainty NNSB-OPTICS clustering algorithm and applied to landslide prediction. Firstly, the expansion strategy of OPTICS-PLUS algorithm is optimized, which avoids the manual setting of density threshold and improves the efficiency of the algorithm. Then, according to the distribution characteristics of rainfall data, combined with EW distance formula and cloud model theory, this paper puts forward EC distance formula, can deal with the uncertain rainfall data effectively. Finally, the uncertain NNSB-OPTICS clustering algorithm is applied to predict landslide hazard in Baota district of Yan’an city and the landslide prediction accuracy reaches into 87.9%. The experimental results show that this method can effectively improve the accuracy of landslide prediction and has high feasibility.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-20 Cooperative journals: 《计算机应用研究》
Abstract： For a low capacity of the watermarking algorithm based on the gray scale symbiosis matrix, this paper would introduce the image watermark algorithm which combines texture complexity with DCT-domain JND model. Firstly, it divided the original image into several blocks, and the blocks the texture complexity would be calculated by four texture features of GLCM of each block. The blocks would be sorted based on the texture complexity to locate the blocks of watermark embedding. After that, the prime matrix of original image would be transformed into DCT domain and, combined with new partition way, JND value of each block would be calculated. The mode of watermark embedding would be determined on the basis of JND value and new rules of embedding. The experimental results show that In the same capacity of watermark embedding, the average peak sinal-to-noise ratio(PSNR) of the image in this method is 4.27% higher than existing mathods, when embedding a watermark that is twice the upper limit of the capacity of the original method, the average PSNR of the image is still 53.4498 dB.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-04-19 Cooperative journals: 《计算机应用研究》
Abstract： In the clustering study of landslide prediction, the difficulties of determining the number of clusters which traditional clustering algorithm needs to set in advance and accurately measuring the important factor of Landslide induced-rainfall leads to bad prediction effect. Therefore, this paper proposes a new clustering algorithm-Uncertain PAHT algorithm , the algorithm introduces a kind of uncertain data model called M-D distance, which effectively measure the uncertain rainfall; and based on the hierarchical clustering thinking, through finding the best threshold p* to determine the k value. Contrast experiment in Yenan Baota district as an example, the experimental results verified the effectiveness of uncertain M-D distance and PAHT algorithm and the feasibility of uncertain PAHT algorithm on the landslide hazard prediction.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-04-19 Cooperative journals: 《计算机应用研究》
Abstract： The traditional particle swarm optimization algorithm uses constant learning constants for social and self -cognition to limit the population 's global coordination ability. In the late convergence of the algorithm, the diversity of the population is lost and all the individuals converge to one point in search space, which can trigger the precocious convergence. In view of this defect, this paper proposed a chaotic map particle swarm optimization algorithm based on variable learning factor. In the early stage of the algorithm, the emphases should focus on the best location of self-recording. At the later period of the algorithm, it should design a coordinated dynamic learning factor to converge on the best position of population. In order to overcome the premature phenomenon and determine the variance of population diversity below the set value, using chaotic mapping updated the optimal individual location of the generation and utilizing a new way to optimized. The experimental show the new algorithm has better performance in convergence speed and precision.
Hits
Hits
Downloads
Comment