Table of Content

    20 July 2016
    Volume 51 Issue 7
    Computational humor researches and applications
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  1-10.  doi:10.6040/j.issn.1671-9352.0.2016.266
    Abstract ( 1513 )   PDF (1413KB) ( 884 )   Save
    References | Related Articles | Metrics
    Humor, as a special phenomenon of human communications, can warm up the atmosphere and eliminate embarrassment. In recent years, with the research development of artificial intelligence, research area related to how to model humorous expression using computers becomes a hot topic in natural language processing tasks, and evolves to become a new subject, called computational humor. Computational humor aims to recognize and interpret humorous expressions in context using natural language processing technologies, and construct humor based computational models. In this article, we firstly introduce the backgrounds of computational humor research and detail the reasons for which humor can be modeled using computers. After that, we review related research in two lines, humor recognition and humor generation, and give the computational procedure of them respectively. Finally, we introduce some applications of humor computing in different tasks, including chatting robots, machine translation, children teaching software and English teaching. Overall, we review the recent research work in the area of humor computing to motivate new ideas and broaden horizons for further research in this area, which can help computers understand the natural language of humans, and promote the development of artificial intelligence.
    Deduplicating search results of cloud disk resources using meta-information
    LIU Chi, YAN Hong-fei
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  11-17.  doi:10.6040/j.issn.1671-9352.1.2015.060
    Abstract ( 370 )   PDF (1623KB) ( 347 )   Save
    References | Related Articles | Metrics
    Different from classical duplicate detection methods which calculating text similarity of web pages, the multi-media cloud disk resources only have limited meta-information to deduplicate search results. The research is based on a newly established cloud disk resources search engine. This paper analyzed the characteristic of cloud disk resource meta-information, finding that besides resource names, extension filename, size and ownership are significant features to detect duplicate records. According to this, this paper proposed a feature normalization method and trained an unsupervised method to capture the task. Experiments proved that this method is able to solve the cloud disk resources search results deduplicating problem effectively.
    Study on collection statistics for parameter selection in pseudo relevance feedback
    MENG Ye, ZHANG Peng, SONG Da-wei
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  18-22.  doi:10.6040/j.issn.1671-9352.1.2015.031
    Abstract ( 324 )   PDF (1668KB) ( 373 )   Save
    References | Related Articles | Metrics
    Pseudo-relevance feedback(PRF)is an effective technique used to improve the Ad hoc retrieval performance. For PRF methods, how to optimize the balance parameter between the original query model and feedback model is an important but difficult problem. In the current feedback methods, the balance parameter is often set to a fixed value across all collections. However, due to the difference among collections, this parameter should be tuned differently. In this paper, we aim to discover some meaningful clues for the optimization of the balance parameter through analyzing the statistical features of collections. We investigates the dependency between the optimal parameter and a number of collection statistics, including the standard deviation of document length(Dev(dl)), the proportion of low frequency terms in the collection(LFT-C)and in the expansion terms. The experiments on six TREC collections demonstrate that the higher LFT-C and Dev(dl)are, the bigger weight of the original query model should be given.
    An ontology-based readability model for vertical search
    ZHANG Wen-ya, SONG Da-wei, ZHANG Peng
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  23-29.  doi:10.6040/j.issn.1671-9352.1.2015.069
    Abstract ( 456 )   PDF (1309KB) ( 341 )   Save
    References | Related Articles | Metrics
    As an emerging evaluation criteria of information retrieval(IR), readability plays an important role in accessing documents relevance, utility and quality. How to provide different users with relevant and readable documents has been an urgent problem in vertical search. In order to solve this problem, we propose a new ontology-based readability method. Based on users’ reading process, we measure documents readability from surface and conceptual levels. In this model, three readability indicator shave been introduced, i.e., Concept Topography, Concept Scope and Document Coherence. Specifically, the readability of a document that computed by individual or combined indicators can be used to re-rank the initial lists of documents which are returned by a conventional search engine. In medical domain, the user-oriented evaluations show that our model has good correlation with humans’ judgments in readability prediction. And our model is also competitive compared with one of the state-of-the-artreadability models in system-orient edevaluation.
    Construction of expert relationship network based on random walk strategy
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  30-34.  doi:10.6040/j.issn.1671-9352.1.2015.085
    Abstract ( 380 )   PDF (1444KB) ( 491 )   Save
    References | Related Articles | Metrics
    Organizing expert relationship is the core of constructing expert relationship network. A method was proposed to build an expert relationship network based on random walk strategy. Firstly, on the basis of the extracted expert entities and their relationships, the method gets a friend relationship, or a guidance relationship or a colleague relationship between two expert entities to form a simple undirected graph. From the relationships between expert nodes in these graphs, an expert relationship matrix was built. Then according to random walk strategy thought, it is made an organic combination of these simple undirected graphs which characterize expert relationships to construct a complex network of expert relationships. The experimental results show that the proposed method to construct an expert relationship network based on random walk strategy is effective.
    Personalized search based on folksonomy and category
    GUAN Yi-zhou, XU Bo, LIN Yuan, LIN Hong-fei
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  35-42.  doi:10.6040/j.issn.1671-9352.1.2015.E28
    Abstract ( 342 )   PDF (1304KB) ( 442 )   Save
    References | Related Articles | Metrics
    Web 2.0 has provided information retrieval with many useful resources. Two kinds of them are beneficial for personalized search, which are social annotations and categorical information. The annotations that user gives come from his consideration, and the categories of the documents that he annotated reflect his preference and interest. So combining these two kinds of resources will benefit the personalized search. Our work is based on a personalized searching method with only annotations, and we propose a method based on both annotations and categorical information. We use them to screen similar users in preference and interest to extend users profile, so that user extended profile will be more accurate. The experiment based on real dataset proves that our method has superiority.
    Translation model adaptation based on semantic distribution similarity
    YAO Liang, HONG Yu, LIU Hao, LIU Le, YAO Jian-min
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  43-50.  doi:10.6040/j.issn.1671-9352.1.2015.116
    Abstract ( 548 )   PDF (1611KB) ( 434 )   Save
    References | Related Articles | Metrics
    Statistical machine translation(SMT)system is trained with large-scale and domain-mixed parallel corpus, when the data for training and testing are not belonged to the same domain, the translation quality usually drops dramatically. To solve this problem, we proposed a novel approach to adapt the translation model based on semantic distribution similarity of translation pair. The approach firstly obtained word representations both in source and target language, and then built mapping to link the different vector space. With the mapping function the semantic k-nearest neighbors of source language in the target vector space can be easily obtained. Based on the semantic distribution of k neighbors in the general domain space, we computed phrases translation similarity in the domain of interest. The similarities are then integrated into the decoder engine, in order to enhance the adaption ability of common translation model. Experiments on English to Chinese translation tasks show that the optimized translation systems build on our method outperform the baseline system by 0.67 and 0.56 BLUE points on news and science-technology test sets respectively.
    Text sentimental orientation analysis based on HNC contextual framework and sentimental dictionaries
    ZHANG Ke-liang, HUANG Jin-zhu, CAO Rong, LI Feng
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  51-58.  doi:10.6040/j.issn.1671-9352.1.2015.E48
    Abstract ( 366 )   PDF (1965KB) ( 528 )   Save
    References | Related Articles | Metrics
    Based on the current sentimental dictionaries and HNC contextual framework, a method of text sentimental orientation analysis was put for ward. The sentimental analysis process covers two phases: feature words matching and feature sentence(or sentence group)finding; feature sentence or sentence group sentimental analysis based on HNC contextual framework. In the first phase, sentimental dictionary of HowNet and Valency Dictionary of English Adjective(VDEA)are applied while the feature sentence or sentence group are analyzed in the second phase. The method, through exact matching of feature words, makes the posterior processing work more effective energy-focused because only those sentences or sentence groups containing subjective sentiment can be analyzed and processed. This thought also illustrates one of the spirits of HNC: doing certain things and refraining from doing other things. The paper takes texts concerning politics, economy,sports and films’ comments as experimental data and experiment result shows that sentimental orientation recognition rate of texts concerning goods and film comments are higher than that politics and sports. The expected experimental results of the paper is reached, which tested the feasibility of the method.
    Online shopping customer service dialogue annotation and analysis
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  66-73.  doi:10.6040/j.issn.1671-9352.1.2015.007
    Abstract ( 681 )   PDF (1139KB) ( 572 )   Save
    References | Related Articles | Metrics
    There is lack of research data on real application environment for interactive question-answering research. This paper collected a large number of online shopping customer service dialogue records as real application environment interactive question-answering corpus. First, the online customer service dialogue records were statistics and analysis. Then 174 groups service dialogues were randomly selected. Those dialogues were annotated and statistics on unnormal language, question relevance and question answer matching phenomena. The annotation and statistics results show that: high frequent dialogue sentences reached to large proportion, 15% of high frequent customer dialogue sentences covered 45% of all data customer sent out; 50% of dialogue sentences contained unnormal language phenomena; Anaphora relevance, omission relevance and common word sequences are the three most important features for judging relevance of client questions; more than 60% of service dialogue sentences are cross matching question answers pairs, and more than 50% of matching question answers pairs are recessive matching.
    Research of gender prediciton based on SVM with E-commerce data
    PENG Qiu-fang, LIU Yang
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  74-80.  doi:10.6040/j.issn.1671-9352.1.2015.094
    Abstract ( 369 )   PDF (1443KB) ( 426 )   Save
    References | Related Articles | Metrics
    Different gender of Users have different view on products, particularly in appreciation of fashion related products, the gender influence is much important. This paper used seven characteristics choosed from online based e-commerce product browsing history data and used support vector machines(SVM)set model by these seven characteristics to predict users' gender. By analysing and training the model, accuracy of gender prediction reached up to 79.21%.While taking advantage of the problem, the paper discusseed the differences between online shopping and offline shopping and do the research about the kernel function of support vector machine and other performance, give the theory and practice reference for the selection of kernel functions and selection of support vector machine.
    Recognition of Chinese Micro-blog sentiment polarity and extraction of opinion target
    HU Mo-zhi, YAO Tian-fang
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  81-89.  doi:10.6040/j.issn.1671-9352.1.2015.089
    Abstract ( 693 )   PDF (1052KB) ( 761 )   Save
    References | Related Articles | Metrics
    According to the dependency and emotional words in sentences, we extract features and apply these features to the maximum entropy model to predict the polarity of a sentence(positive, negative or neutral). Using words, part of speech and composition of syntactic structure as a feature to train CRF model and extract opinion target. Experimental results shows that recognition rate of Chinese Micro-blogging sentiment polarity are increased obviously.
    Access control based on relationship strength for online social network
    CAI Hong-yun, MA Xiao-xue
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  90-97.  doi:10.6040/j.issn.1671-9352.2.2015.350
    Abstract ( 547 )   PDF (1661KB) ( 392 )   Save
    References | Related Articles | Metrics
    Access control is one of the effective methods to protect the privacy of people in OSN. However, there are some problems in the relationship-based access control scheme, such as the coarse-grained and inflexibility. So evaluating the relationship strength between users is introduced in relationship-based access control, users access can be authorized according to relationship strength. Based on the characteristics of users interaction in OSN, the users attention is acquired by analyzing interaction behavior between users, and then a new evaluation model for relationship strength is constructed by considering the following features: attention factor, interaction strength and time decay. Experimental results show that the proposed method is feasible and effective.
    Security transfer model of access control information based on TCB subsets
    TANG Qian, YANG Fei, HUANG Qi, LIN Guo-yuan
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  98-106.  doi:10.6040/j.issn.1671-9352.0.2015.571
    Abstract ( 638 )   PDF (1984KB) ( 478 )   Save
    References | Related Articles | Metrics
    A security transfer model of access control information based on TCB subsets was proposed by taking a comprehensive consideration of the security requirements for the application layer transferring the access control information to the kernel layer. One security manager in the application layer and the other security manager in the kernel layer are connected by security channel, which has been encrypted. The key is stored in the trusted platform module. The access control information must be managed by the trusted platform module before passing through the security channel. The application layer interface of the security channel transfers the access control information and the labels to the kernel layer interface of the security channel and then does random check, after the security channel has been encrypted. The kernel layer interface returns the proofs and the application layer interface judges the result. The security transfer model can not only ensure the security of the access control information, but also resist the spiteful cheat and the hostile attack, thus improving the reliability and valid of the access control.
    Cryptanalysis and improvement of two kind of certificateless aggregate signature schemes
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  107-114.  doi:10.6040/j.issn.1671-9352.0.2016.026
    Abstract ( 504 )   PDF (1012KB) ( 377 )   Save
    References | Related Articles | Metrics
    According to the existing certificateless aggregate signature schemes presented based on bilinear pairings, lots of them have the security flaws and low computational efficiency problem. The security of two certificateless aggregate signature schemes proposed was analyzed, and it is found that the verify equation in the first scheme is not right and the two schemes can not resist forgery attack under TypeⅡ. Finally, an improved scheme based on RSA without bilinear pairing was proposed in this paper. Based on the RSA assumption and the DL problem, it is proved that the new scheme is existentially unforgeable. Compared with other schemes, the new scheme is more efficient and secure.
    Research on beamforming design for multi-user full-duplex SWIPT systems
    WANG Zhi, CHEN Dong-hua, HE Yu-cheng
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  115-120.  doi:10.6040/j.issn.1671-9352.0.2015.491
    Abstract ( 377 )   PDF (1494KB) ( 555 )   Save
    References | Related Articles | Metrics
    A joint optimization scheme based on the dual goals which are the minimization of the power consumption and the maximization of the energy harvest was proposed for the multi-user full-duplex(FD)cellular communication system. The proposed scheme used zero-forcing algorithm to eliminate multi-user interference in the uplinks, in order to achieve the effective transfer of information and energy, the full-duplex base station provided users with communication services via information beamforming and energy beamforming in the downlinks. The proposed scheme made a simultaneous improvement in power efficiency and spectrum efficiency while guaranteeing the quality-of-service(Qos)of all users and power constraints. Because the problem of power is non-convex, the original problem was transformed to a convex one via semi-definite relaxation. Simulation results show that the proposed scheme provides substantial power savings over traditional scheme. Moreover, FD base station transfers energy to the downlink users through energy beamforming, which improves power efficiency of the system effectively.
    A novel small peptide ligand for detection of Alzheimer-associated neuronal thread protein
    ZHANG Yan, WEI Yu-ping, ZHANG Liang, FU Yan-kai, XU Jian-dong, XU Xia
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  121-125.  doi:10.6040/j.issn.1671-9352.0.2016.119
    Abstract ( 422 )   PDF (1592KB) ( 444 )   Save
    References | Related Articles | Metrics
    Alzheimer-associated neuronal thread protein(AD7c-NTP)as a sensitive, easily obtained and noninvasive biomarker for AD disease detection. This research is aim to detect AD7c-NTP protein. A novel peptide for detecting AD7c-NTP was developed using molecular simulations and the peptide was synthesized. The recombinant protein was expressed. The bound protein was tested using chromatography and the mimic urine was detected as well. The peptide of DEWH is able to bind AD7c-NTP with the adsorption rate of around 85% and the linearity range is 200-900 μg/L. This novel peptide could provide an easy and repeatable method for detecting AD disease, which is critical for early diagnosis and postponing disease progression with proper treatment.
    Salt-tolerant mechanism of alcohol ether carboxylate investigated by molecular dynamics simulation
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  126-130.  doi:10.6040/j.issn.1671-9352.0.2016.188
    Abstract ( 409 )   PDF (2229KB) ( 547 )   Save
    References | Related Articles | Metrics
    The aggregation behavior in solution of the surfactants, sodium dichloroisocyanurate(SDC)and dodecyl alcohol polyoxyethylene ether carboxylate(C12E3C)were investigated by unit-atoms molecular dynamic simulation. Interactions between the surfactants and Na+, Ca2+ were analyzed, and the salt-tolerant mechanism of srufacants on molecular level were explained. Salt-bridging structure, which reduced the electrostatic interaction of the surfactant micelles, thus led to the closer combination of the micelle, between Ca2+ and nearest neighbor headgroup pair in salt solution was observed. Ca2+ changed the structure of hydration layers around the headgroup of surfactant. The potential of mean force(PMF)showed that the energy barriers between the headgroup and Ca2+ and Na+ in the C12E3C system were higher than those in the SDC system, which indicated that SDC binds the ions more easily than C12E3C, and the ions have strong influences on SDC system.
    Analysis and distribution of wetland vegetation of Nansi Lake, China
    FAN Xiao-li, LIU Bo-yan, LIANG Yu, LIU Jian, FANG Yong, MENG Zhen-nong
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  131-136.  doi:10.6040/j.issn.1671-9352.0.2016.025
    Abstract ( 370 )   PDF (1485KB) ( 351 )   Save
    References | Related Articles | Metrics
    Nansi Lake is the most important wetland in Shandong Province, plant species are very rich, there are 201 species,belonging to 66 families and 147 genera. Among, fern 3 families, 3 genera and 3 species; gymnospermae 2 families, 2 genera and 2 species; Monocotyledoneae 14 families, 39 genera and 58 species; Dicotyledoneae 47 families, 103 genera and 138 species. The main plant life form are perennial herb and annual herb. Most of the genera distribution type belong to Cosmopolitan distribution, Pantropic distribution and North Temperate distribution.
    Research into the accumulative levels about Cd,Pb in Channa argus and Siniperca chuatsi from the East Dongting Lake
    SUN Xiao-chuan, WANG De-liang, WANG Yuan-lan, ZHAN Hui-ying
    JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2016, 51(7):  137-142.  doi:10.6040/j.issn.1671-9352.0.2016.035
    Abstract ( 345 )   PDF (981KB) ( 395 )   Save
    References | Related Articles | Metrics
    The concentrations and distribution characteristics of Cd,Pb in the organizations of Channa argus and Siniperca chuatsi from East Dongting lake had been measured with wet digestion by Graphite Furnace Atomic Absorption Spectrophptometry(GF-AAS), and their associations with body length and body weight were also examined, and the method of single factor pollution index and THQ were used to evaluate the pollution degree and edible security. The results showed that the concentrations of the heavy metals present tissue-specific, and the concentrations and the distributions of the Cd, Pb in the same organs and tissues were quite different.(Accumulations of Pb are higher than Cd). The contents of Cd and Pb are higher in intestines and purtenance, and lower in muscle. Correlation analysis showed that there were significant correlations between concentrations of the heavy metals in the organs of these fish(P<0.05)except the Pb in Channa argus, also the levels of Cd, Pb in each organ presented higher correlation with body length and body weight in Siniperca chuatsi, but there is lower in Channa argus. The pollution degree analysis and health risk assessment showed that in addition to the muscle tissue(Pb snakehead muscle tissue is slightly polluted), other tissues and organs are presented different pollution levels. Consumption of snakehead and mandarin fish from this area has no obvious health risks and food security is higher.