The spectrum of a graph has been widely used to characterize the properties of a graph and extract information from its structure. In this paper, we investigate the performance of Laplacian spectrum and multidimension...
详细信息
The spectrum of a graph has been widely used to characterize the properties of a graph and extract information from its structure. In this paper, we investigate the performance of Laplacian spectrum and multidimensional scaling (MDS) as shape recognition and clustering. Firstly, we extract boundary points to characterize the shape and to construct the Laplacian matrix. Secondly, the structural information about graph is described by using the eigenvalues of the Laplacian matrix. Finally, the given shapes are projected onto the low-dimensional space by performing MDS. Meanwhile, the clustering is achieved via analyzing the distribution of shapes. Comparative experiments on the public data sets demonstrate the validation of the proposed algorithm.
Current system combination methods usually use confusion networks to find consensus translations among different systems. Requiring one-to-one mappings between the words in candidate translations, confusion networks h...
详细信息
Bandelet transform is an efficient image sparse representation approach which can adaptively approximate the geometrical regularity of image structures. In this paper, a multi-bandelets based method for SAR image comp...
详细信息
The rapid development of the Web technology makes the Web mining become the focus of the current data mining, and the XML technology also becomes the standard of the data exchange on the Web. The paper introduces the ...
详细信息
Current statistical machine translation systems usually extract rules from bilingual corpora annotated with 1-best alignments. They are prone to learn noisy rules due to alignment mistakes. We propose a new structure ...
详细信息
In this paper, an improved method of calculating ontology semantic similarity is proposed to enhance the information retrieval recall and precision. To filter out the document which have smaller related degree with or...
详细信息
In this paper, an improved method of calculating ontology semantic similarity is proposed to enhance the information retrieval recall and precision. To filter out the document which have smaller related degree with original query, the scores of search results document is re-calculated by use of ontology semantic similarity. A new definition of the iterative query expansion parameters is put forward which can reduce the number of expansion and further improve the efficiency of the query. The use of open source tools for text semantic retrieval test, i.e., Jena and Lucene, has verified the feasibility and effectiveness of the proposed method.
Current tree-to-tree models suffer from parsing errors as they usually use only 1-best parses for rule extraction and decoding. We instead propose a forest-based tree-to-tree model that uses packed forests. The model ...
详细信息
Current SMT systems usually decode with single translation models and cannot benefit from the strengths of other models in decoding phase. We instead propose joint decoding, a method that combines multiple translation...
详细信息
A multiuser detector based on Schur algorithm is studied in this paper. Because the computational complexity of the conventional decorrelating detector is high while computing the inverse of system matrices, especiall...
详细信息
A multiuser detector based on Schur algorithm is studied in this paper. Because the computational complexity of the conventional decorrelating detector is high while computing the inverse of system matrices, especially when the system is asynchronous and the number of users is huge. The simulation results show that the performance of the multiuser detector based on Schur algorithm is similar to that of decorrelating detector, but it's computational complexity is much lower than decorrelating detector's.
Hyper Surface Classification (HSC), which is based on Jordan Curve Theorem in Topology, has been proven to be a simple and effective method for classifying a large database in our previous work. In this paper, through...
详细信息
Hyper Surface Classification (HSC), which is based on Jordan Curve Theorem in Topology, has been proven to be a simple and effective method for classifying a large database in our previous work. In this paper, through theoretical analysis, we find that different scales may affect the training process of HSC, which influences its classification performance. To investigate the impact and find a suitable scale, the scale transformation of HSC is studied. The experimental results show that the accuracy increases with the shrinkage of the scale, but the effect is tiny. Furthermore, we find that some samples become inconsistent and repetitious when the scale is adequately small, because of the powerlessly providing enough precision by the data type of computer. Fortunately, HSC can get a high performance with common scales as experiments exhibit.
暂无评论