Search

Abstract: FP-Growth is a recursive mining algorithm for frequent patterns. With each recursion, a new FP-tree and the corresponding header table are built. This paper proposes another approach to improve the mining efficiency: only one FP-tree is constructed, and instead of building new trees in every recursion, builds new header tables; frequent patterns are generated along the way, so we call this algorithm “header table recursion (HTR)” method. Tests show that it will get better performance for relative small datasets because the compress of the FP-tree has no significant impact on the performance yet inflicts extra time cost in tree creation operations. However on larger datasets, the traversal of the original FP-tree will become more time-consuming and its overall performance needs further testing. [Copyright &y& Elsevier]

Abstract: Electroencephalogram EEG and MEG research are the non-invasive brain imaging techniques. The main research measured by the scalp EEG or MEG brain source signals can obtain three-dimensional quantitative information. Structure to match with the actual situation is the first model to solve EEG forward, the inverse problem of the clinical effectiveness of pre-conditions and key. Model of the head and the head are on the premise for the brain magnetic resonance image contour extraction and segmentation. Since the first complex internal structure, the access to the brain volume data segmentation is more difficult. In this paper, we propose a simulated annealing-based deformable model method and realize the head T2 weighted magnetic resonance images of the cross-sectional area of the extraction of the brain. To avoid falling into the local minimum in the deformation of the deformable model, we use a simulated annealing process of thinking, making the algorithm with the global search capability. The experimental results show that the improved algorithm is effective and robust. [Copyright &y& Elsevier]

Abstract: As e-commerce is becoming more and more popular and interactive, Merchants selling products on the Web 2.0 often ask their customers to review the products which they have purchased. The number of reviews can be in hundreds or even thousands for a popular product. This makes it difficult for a potential customer to read them or to make an informed decision on whether to purchase the product. It also makes it difficult for the manufacturer of the product to keep track of customer opinions. In this research, we aim to mine and to analyze all the customer reviews of a product. The algorithm proposed in this research is more reliable on opinion identification because it is unsupervised and the accuracy of the result improves as the number of reviews increases. Our research is performed in four steps: (1) mining the frequent aspects of a product that have been commented on by customers; (2) mining the infrequent aspects of a product which have been commented by customers (3) identifying opinion words in each review and deciding whether each opinion word is positive, negative or neutral; (4) summarizing the results. This paper proposes several novel techniques to perform these tasks. Our experimental results using reviews of a number of products sold online demonstrate the effectiveness of the techniques. [Copyright &y& Elsevier]

Abstract: Clustering is a method of unsupervised learning, and a common technique for intelligent data analysis used in many fields. The enlarging volumes of information produced by the progress of technology, makes clustering of very large scale data a challenging task. Therefore, parallel processing of very large scale data set is a good solution to that problem. CLARANS is approved to be an efficient and effective clustering method for spatial data mining. It is insensitive to outliers and has the advantage of processing category data directly compared to other partition method like K-means. However, it is a serial algorithm and cannot process large scale data set efficiently. In this paper, we propose a parallel CLARANS clustering algorithm based on MapReduce, which is a current and powerful parallel programming framework. Experimental results demonstrate that the parallel CLARANS algorithm can handle large scale data set efficiently and effectively. Furthermore, the evaluation criterions of scaleup, speedup and sizeup validate its efficiency. [Copyright &y& Elsevier]

Abstract: In this paper, secure audio forensic mark algorithm to trace the illegal distribution is proposed. The proposed algorithm uses off-axis hologram when make the forensic mark for embedding into the audio. Embedding and extracting is performed in the DWT domain, and the Arnold transform of the forensic mark is used for improving the security. The hologram generator makes the hologram of the Arnold transformed mark. Because of the characteristics of the off-axis hologram, the proposed algorithm can embed high capacity information and extract without the original signal. [Copyright &y& Elsevier]

Abstract: Fuzzy association rules are easily understood by users. Simulated annealing algorithm is an effective approximation algorithm which is suitable for solving large-scale combinatorial optimization problem. In this paper we proposed a classifier design method based on simulated annealing algorithm and fuzzy association rules. It can achieve satisfied accuracy and a small number of classification rules. Experiments demonstrated the effectiveness and advantages of the proposed algorithm. [Copyright &y& Elsevier]

Abstract: In frequent neighboring class set mining, when generating candidate frequent itemset, present algorithms have some redundant candidate and repetitive computing, and so this paper proposes an algorithm of complementary mining frequent neighboring class set, which is suitable for mining any frequent neighboring class set. This algorithm adopts double search strategies to generate candidate frequent neighboring class set in the mining course. One is using down-up search strategy to generate candidate, namely, it connects two k-frequent neighboring class sets to generate (k+1)-candidate frequent neighboring class set. The other is computing complement set of (k+1)-candidate to generate next candidate frequent neighboring class set. By the two ways, the algorithm may delete redundant candidate and repetitive computing. The algorithm creates digital database of neighboring class set via neighboring class weight, according to character of database, it also uses digit logical operation to computes support. The result of experiment indicates that the algorithm is faster and more efficient than present algorithms when mining frequent neighboring class set in large spatial data. [Copyright &y& Elsevier]

Abstract: In the course of mining frequent neighboring class set, present algorithms have some redundant candidate and repetitive computing, which are only able to efficiently extract short frequent neighboring class set, and so this paper proposes an algorithm of fast duplex mining frequent neighboring class set, which is suitable for mining any frequent neighboring class set. This algorithm adopts two methods to generate candidate frequent neighboring class set in the mining course. One is using top-down search strategy to generate candidate by numerical index, the other is using anti-code of front this candidate to generate next candidate frequent neighboring class set. The course of top-down search strategy used by the algorithm isn’t different from traditional top-down search strategy, which uses numerical index to generates k-subset of (k+1)-non frequent neighboring class set. By the two methods, the algorithm may delete redundant candidate and repetitive computing. The algorithm creates digital database of neighboring class set via neighboring class weight, according to character of database, it also uses digit logical operation to computes support. The result of experiment indicates that the algorithm is faster and more efficient than present algorithms when mining frequent neighboring class set in large spatial data. [Copyright &y& Elsevier]