Table of Content

  

20 December 2014
Volume 49 Issue 12

Topic sentiment analysis of Chinese news based on emotional dependency tuple

ZHOU Wen, ZHANG Shu-qing, OUYANG Chun-ping, LIU Zhi-ming, YANG Xiao-hua

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  1-6.  doi:10.6040/j.issn.1671-9352.3.2014.159

Abstract ( 1501 )   PDF (1681KB) ( 1246 )   Save

References | Related Articles | Metrics

Taking the emotional dependency tuple (EDT) as the basic structure of Chinese emotional expression, the news text theme emotion recognition task was divided into three progressive sub-tasks: topics identification, emotional tendentiousness analysis, subjective and objective classification. TF-IDF method was improved before identifying the topic, and then the cross-entropy-based method was combined to extract themes feature words. The topic representation of the news title was taken into consideration at the same time, and the title words were put into the theme feature set. The similarity between sentence and the topic feature vector was calculated based on the vector space model. Some statistical rules such as sentence position, sentence length and sentence's similarity with title were added on this foundation to get topic sentences. Finally, the emotional dependency tuple discriminant model was established to calculate sentences emotion and the subjective and objective judgment rule were used to filter out the tendency key sentence. The approaching to the best results of experiment based on COAE 2014 evaluation data shows that the classification method based on the EDT has high classification performance.

Approximate reasoning research of customer's brand switching intention based on online reviews

SONG Shuang, ZHAO Narisa, ZHANG Yang

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  7-11.  doi:10.6040/j.issn.1671-9352.3.2014.132

Abstract ( 1186 )   PDF (1197KB) ( 780 )   Save

References | Related Articles | Metrics

Based on consumer psychology and experiential marketing theory, a model of customer switching intention was established from four dimensions: perceived quality, perceived value, brand reputation and expectation. Then semantic analysis on the valuation words and emotional words extracted from the consumer online reviews was made. Furthermore, combined with fuzzy sentiment corpus and fuzzy inference, the consumer's brand switching intention was computed, and the customer's switching intention types was identified. Finally, an experimental study with the case of four types of cosmetics on the TaoBao site was carried out, and reasonable conclusions and some marketing advices were obtained, which all verify the validity of the method.

Event inference for semi-supervised Chinese event extraction

XU Xia, LI Pei-feng, ZHENG Xin, ZHU Qiao-ming

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  12-17.  doi:10.6040/j.issn.1671-9352.3.2014.182

Abstract ( 1338 )   PDF (805KB) ( 1302 )   Save

References | Related Articles | Metrics

The performance semi-supervised Chinese event extraction depends on the quality of seed patterns. However, the expression styles and coverage of those seed patterns, which are extracted automatically, is limited and that leads to lots of event mentions cannot be identified for their contexts. To solve this issue, an event inference mechanism based on co-reference events and relevant events was proposed, which follows the theory of event consistency in a topic. This mechanism can infer those event mentions which have the co-reference or relevance relations with the extracted event mentions in the same document, and then the performance of semi-supervised Chinese event extraction was further improved. The experimental results on the ACE 2005 Chinese corpus show that our approach outperforms the baseline significantly.

Short text classification based on word embedding similarity

MA Cheng-long, JIANG Ya-song, LI Yan-ling, ZHANG Yan, YAN Yong-hong

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  18-22.  doi:10.6040/j.issn.1671-9352.3.2014.295

Abstract ( 1629 )   PDF (2161KB) ( 2949 )   Save

References | Related Articles | Metrics

As the short length of the Web short text and less shared words, a lot of out of vocabulary (OOV) words would appear, and these words make the task of text classification more difficult. To solve this problem, a new general framework based on word embedding similarity was proposed. First, get the word embedding file with unsupervised learning method based on unlabeled data. Second, extend the OOVs with the similar words in training data through computing the similarities of different word embeddings. The comparison with the baseline system shows that the proposed method gets better 1%-2% rate and outperforms more 10% rate on small training data set.

Feature extraction method based on Bootstrapping in English product comment

WANG Hui, CHEN Guang

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  23-29.  doi:10.6040/j.issn.1671-9352.3.2014.123

Abstract ( 1479 )   PDF (1246KB) ( 827 )   Save

References | Related Articles | Metrics

An feature extraction method based on Bootstrapping in English product comment was proposed. By this method, starting with a set of extraction patterns as seeds, and then applying an incremental iterative procedure to find new features. During the process of the each iteration, the system ranks the new features by score, which is calculated by the intimacy relationship between the candidate features and patterns. This is useful for prevent topic drift. After extracting features, WordNet is used to calculate the similarity between features. Then clustering the features by the similarity score, get different aspects of the product features, then filtering out the low score of the class clusters, remove noise. What's more, to improve the portability of the system, the seed features are replaced by seed patterns. Experimental results show that extracting features by this method has a good result, the precision, recall and F-measure reach 0.799, 0.779, 0.789 and it has good extraction performance.

An variable length code algorithm compression inverted index

MAO Fu-lin, QU You-li

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  30-35.  doi:10.6040/j.issn.1671-9352.3.2014.230

Abstract ( 1297 )   PDF (1462KB) ( 1605 )   Save

References | Related Articles | Metrics

The efficiency of text search engines relies on data structure: inverted index. And the more large space is need to storage the inverted index. A new compression algorithm was proposed. For the given document collections. Terrier, the information retrival tool, was used to build inverted index, and the state-of-the-art compression techniques was used to compress inverted file. Then the compress ratio was confirmed by comparing the file size. Experiments show that the new compression techniques can get much better compress ratio.

Coreference wordlist construction of Russian weapon

ZHANG Ming, TANG Hui-feng, LI Zhu-feng

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  36-42.  doi:10.6040/j.issn.1671-9352.3.2014.139

Abstract ( 1454 )   PDF (2286KB) ( 797 )   Save

References | Related Articles | Metrics

In the automatic processing of Russian military texts, it is necessary to deal with coreference resolution of weapon names. In order to solve this problem, the coreference words and the patterns were taken out from the Infobox of entries in Wikipedia by using the pattern matching method. These words and patterns were combined into new patterns. And then these new patterns were iteratively put back into the entries of Wikipedia to find more coreference words. A coreference wordlist was generated as the result of this method. The correct rate of result shows that the pattern matching method can accurately and efficiently identify coreference words of weapon names from the entries of Russian Wikipedia.

Contour matching method based on curvature scale space

JIN Yong-gang, WANG Fan, HU Xiao-peng

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  43-48.  doi:10.6040/j.issn.1671-9352.3.2014.124

Abstract ( 1484 )   PDF (2667KB) ( 1473 )   Save

References | Related Articles | Metrics

An approach for matching and recognizing affine-distorted planar shapes was proposed, which is based on the theory of curvature scale space (CSS). First, the contours were extracted by edge detection algorithm, and then use the PCA whitening method to eliminate the impact of scale, translation and shear. Second, the contour was resampled and the local maximum point was extracted. Final, the array of points was used as descriptor to match. The proposed method was evaluated on MCD shape database, and the experimental results show that the method not only has the higher retrieval ratio but also is robust to affine transformation.

Tail dependence analysis based on tail sample data

LI Shu-shan

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  49-54.  doi:10.6040/j.issn.1671-9352.0.2014.156

Abstract ( 1998 )   PDF (710KB) ( 2474 )   Save

References | Related Articles | Metrics

Tail dependence is the nature of the tail to the joint distribution of the two random variables. Be directed against tail dependence analysis, two concepts of order statistics for 2-dimension random vector were proposed and an idea was advanced to estimate the tail dependence coefficient by fitting Copula function using tail sample. Then the method of parameter estimation and fitting tests for the tail based on tail sample date and the corresponding estimate method for tail dependence coefficient were also discussed. A Monte Carlo simulation was given to illustrate the validity of the method given. Finally the tail dependence between the SHI and the SZI was analyzed.

Weight of evidence method and credit risk control

GAN Xin-jun, YANG Wei-qiang

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  55-59.  doi:10.6040/j.issn.1671-9352.0.2014.269

Abstract ( 1998 )   PDF (1030KB) ( 1015 )   Save

References | Related Articles | Metrics

The weight of evidence (WOE) method was adopted on the credit risk control in commercial bank and a full description of the algorithm was proposed. By employing this WOE approach on the real financial data of commercial bank, the risk control model was constructed such that the probability of default can be estimated more precisely. The significant advantages of WOE method over conventional methods were verified, especially in statistical power.

Mean square exponential stability analysis of stochastic generalized cellular neural networks with S-type distributed delays

ZHEN Yan, WANG Lin-shan

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  60-65.  doi:10.6040/j.issn.1671-9352.0.2014.181

Abstract ( 1287 )   PDF (941KB) ( 581 )   Save

References | Related Articles | Metrics

Global exponential stability of generalized cellular neural networks with S-type distributed delays was discussed by using Lyapunov functional method and stochastic analysis technique. A criterion for the stability in mean square was obtained. Simulation shows the validity of the conclusions.

The freeness of Shi-Catalan arrangements of type G₂

GAO Rui-mei

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  66-70.  doi:10.6040/j.issn.1671-9352.0.2014.187

Abstract ( 1128 )   PDF (1324KB) ( 669 )   Save

References | Related Articles | Metrics

The Shi-Catalan arrangements of type G₂ are multiarrangements in the 2-dimensional vector space, which are obtained by giving the same multiplicity to the hyperplanes in the same orbit in Weyl arrangements of type G₂. Four concrete forms of Shi-Catalan arrangements of type G₂ are given. By counting the intersections of hyperplanes in reflection planes, we prove the conclusion that the cones over the Shi-Catalan arrangements of type G₂ are all free.

Totally real minimal submanifolds in quasi-complex projective space

LIU Min, SONG Wei-dong

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  71-75.  doi:10.6040/j.issn.1671-9352.0.2013.595

Abstract ( 1303 )   PDF (710KB) ( 582 )   Save

References | Related Articles | Metrics

The totally real minimal submanifolds in quasi-complex projective space was studied.Some pinching results were obtained by method of moving frams and improve the related results.

A new general method making up n order magic square

FENG Bao-chu, FENG Yuan-yuan

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  76-80.  doi:10.6040/j.issn.1671-9352.7.2014.003

Abstract ( 1285 )   PDF (695KB) ( 883 )   Save

References | Related Articles | Metrics

A new method making up magic square is presented by analyzing the relation between the structure of n×n natural numbers array and the basic characteristic of any order magic square. The method is as follow: n/2 digitals of all dual rows or dual columns will be interchanged. If n is odd, adjusting the structure of the array is necessary, that is, the ordered series of numbers on the reverse first diagonal angle line, on the second diagonal line, on the central column and on the central row should be replaced respectively by the central column, the central row, the first diagonal angel line and the second diagonal angel line in the new array. Finally, some examples are given to show the advantage of this method, and many new kinds of magic squares will be generated by using this method.

Generalized singular value decompositions with respect to Semi-definite inner product

ZHENG Chan, LI Han-yu

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  81-86.  doi:10.6040/j.issn.1671-9352.0.2014.159

Abstract ( 1373 )   PDF (731KB) ( 753 )   Save

References | Related Articles | Metrics

The generalized inverses of matrices were used to study the weighted singular value decomposition with one of the weight matrices being positive semi-definite and the other being positive definite. Some conditions for the existence of the generalized singular value decompositions with respect to a semi-definite inner product were also presented.

Fuzzy ultra ⊙-ideals in regular residuated lattices

LIU Chun-hui

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2014, 49(12):  87-94.  doi:10.6040/j.issn.1671-9352.0.2014.325

Abstract ( 1395 )   PDF (846KB) ( 782 )   Save

References | Related Articles | Metrics

Firstly, the notion of fuzzy ultra ⊙-ideal is introduced and its properties are investigated in regular residuated lattices, and some equivalent characterizations of fuzzy ultra ⊙-ideals are obtained. Secondly, the lattice operations ∨, ∧ and the order-reversing involution on the set FU(L) of all fuzzy ultra ⊙-ideals in a regular residuated lattice L are defined. It is proved that (FU(L),∨,∧,,0_L,1_L) formes a De Morgan algebra if L satisfies the condition (P). Finally, the adjoint pair () on FU(L) is defined. It is also proved that (FU(L),,0_L,1_L) formes a residuated lattice if L satisfies the condition (P).