Improvement of building field association term dictionary using passage retrieval

Field Association (FA) terms are a limited set of discriminating terms that can specify document fields. Document fields can be decided efficiently if there are many relevant FA terms in that documents. An earlier approach built FA terms dictionary using a WWW search engine, but there were irrelevant selected FA terms in that dictionary because that approach extracted FA terms from the whole documents. This paper proposes a new approach for extracting FA terms using passage (portions of a document text) technique rather than extracting them from the whole documents. This approach extracts FA terms more accurately than the earlier approach. The proposed approach is evaluated for 38,372 articles from the large tagged corpus. According to experimental results, it turns out that by using the new approach about 24% more relevant FA terms are appending to the earlier FA term dictionary and around 32% irrelevant FA terms are deleted. Moreover, precision and recall are achieved 98% and 94% respectively using the new approach.     

A comparison of feature selection methods for an evolving RSS feed corpus

Previous researchers have attempted to detect significant topics in news stories and blogs through the use of word frequency-based methods applied to RSS feeds. In this paper, the three statistical feature selection methods: χ², Mutual Information (MI) and Information Gain (I) are proposed as alternative approaches for ranking term significance in an evolving RSS feed corpus. The extent to which the three methods agree with each other on determining the degree of the significance of a term on a certain date is investigated as well as the assumption that larger values tend to indicate more significant terms. An experimental evaluation was carried out with 39 different levels of data reduction to evaluate the three methods for differing degrees of significance. The three methods showed a significant degree of disagreement for a number of terms assigned an extremely large value. Hence, the assumption that the larger a value, the higher the degree of the significance of a term should be treated cautiously. Moreover, MI and I show significant disagreement. This suggests that MI is different in the way it ranks significant terms, as MI does not take the absence of a term into account, although I does. I, however, has a higher degree of term reduction than MI and χ². This can result in loosing some significant terms. In summary, χ² seems to be the best method to determine term significance for RSS feeds, as χ² identifies both types of significant behavior. The χ² method, however, is far from perfect as an extremely high value can be assigned to relatively insignificant terms.     

Wikipedia bi-linear link (WBLM) model: A new approach for measuring semantic similarity and relatedness between linguistic concepts using Wikipedia link structure

Wikipedia links its articles by manually defined semantic relations called the Wikipedia hyperlink (link) structure. The existing Wikipedia link-based semantic similarity (SS) and semantic relatedness (SR) computation models, such as Wikipedia one-way link (WOLM) model and Wikipedia two-way link (WTLM) model, do not assess the strengths of the relationships between a candidate concept and its links (out-links or in-links). These models treat all the links as equally important even though some links are semantically more influential than others and should be given more importance. This phenomenon reduces the accuracy of these models. This paper presents the Wikipedia bi-linear link (WBLM) model that extends the previously proposed WOLM and WTLM models. The WBLM model explores the Wikipedia link structure as a semantic graph and discovers the strongly (bi-linear links) and weakly (out-links or in-links) connected links of a candidate concept. It improves the link-based vector representations of concepts by assigning weights to their connected links according to the strengths of their semantic associations. The experimental results demonstrate that the proposed WBLM model significantly improves the SS and SR computation accuracy of the WOLM model (6.9%, 8%, 24%, 17.3%, 31.2%, 30.6%, 26.5%, and 35.4%) and WTLM model (1.2%, 3.9%, 7.1%, 9.9%, 11%, 6.3%, 12.7%, and 13%), in terms of linear correlations with human judgments on gold standard benchmarks, including MC30, RG65, WS203, SimLex, 353All, MTurk287, MTurk771, and MEN3000, respectively. Moreover, this research offers a deep insight into the Wikipedia link structure and provides an adequate base for understanding it as a semantic graph.     

Deriving the sentiment polarity of term senses using dual-step context-aware in-gloss matching

Vital to the task of Sentiment Analysis (SA), or automatically mining sentiment expression from text, is a sentiment lexicon. This fundamental lexical resource comprises the smallest sentiment-carrying units of text, words, annotated for their sentiment properties, and aids in SA tasks on larger pieces of text. Unfortunately, digital dictionaries do not readily include information on the sentiment properties of their entries, and manually compiling sentiment lexicons is tedious in terms of annotator time and effort. This has resulted in the emergence of a large number of research works concentrated on automated sentiment lexicon generation. The dictionary-based approach involves leveraging digital dictionaries, while the corpus-based approach involves exploiting co-occurrence statistics embedded in text corpora. Although the former approach has been exhaustively investigated, the majority of works focus on terms. The few state-of-the-art models concentrated on the finer-grained term sense level remain to exhibit several prominent limitations, e.g., the proposed semantic relations algorithm retrieves only senses that are at a close proximity to the seed senses in the semantic network, thus prohibiting the retrieval of remote sentiment-carrying senses beyond the reach of the ‘radius’ defined by number of iterations of semantic relations expansion. The proposed model aims to overcome the issues inherent in dictionary-based sense-level sentiment lexicon generation models using: (1) null seed sets, and a morphological approach inspired by the Marking Theory in Linguistics to populate them automatically; (2) a dual-step context-aware gloss expansion algorithm that ‘mines’ human defined gloss information from a digital dictionary, ensuring senses overlooked by the semantic relations expansion algorithm are identified; and (3) a fully-unsupervised sentiment categorization algorithm on the basis of the Network Theory. The results demonstrate that context-aware in-gloss matching successfully retrieves senses beyond the reach of the semantic relations expansion algorithm used by prominent, well-known models. Evaluation of the proposed model to accurately assign senses with polarity demonstrates that it is on par with state-of-the-art models against the same gold standard benchmarks. The model has theoretical implications in future work to effectively exploit the readily-available human-defined gloss information in a digital dictionary, in the task of assigning polarity to term senses. Extrinsic evaluation in a real-world sentiment classification task on multiple publically-available varying-domain datasets demonstrates its practical implication and application in sentiment analysis, as well as in other related fields such as information science, opinion retrieval and computational linguistics.     

Applying semantic knowledge to the automatic processing of temporal expressions and events in natural language

This paper addresses the problem of the automatic recognition and classification of temporal expressions and events in human language. Efficacy in these tasks is crucial if the broader task of temporal information processing is to be successfully performed. We analyze whether the application of semantic knowledge to these tasks improves the performance of current approaches. We therefore present and evaluate a data-driven approach as part of a system: TIPSem. Our approach uses lexical semantics and semantic roles as additional information to extend classical approaches which are principally based on morphosyntax. The results obtained for English show that semantic knowledge aids in temporal expression and event recognition, achieving an error reduction of 59% and 21%, while in classification the contribution is limited. From the analysis of the results it may be concluded that the application of semantic knowledge leads to more general models and aids in the recognition of temporal entities that are ambiguous at shallower language analysis levels. We also discovered that lexical semantics and semantic roles have complementary advantages, and that it is useful to combine them. Finally, we carried out the same analysis for Spanish. The results obtained show comparable advantages. This supports the hypothesis that applying the proposed semantic knowledge may be useful for different languages.     

Image searching on the Excite Web search engine

A growing body of research is beginning to explore the information-seeking behavior of Web users. The vast majority of these studies have concentrated on the area of textual information retrieval (IR). Little research has examined how people search for non-textual information on the Internet, and few large-scale studies has investigated visual information-seeking behavior with general-purpose Web search engines. This study examined visual information needs as expressed in users’ Web image queries. The data set examined consisted of 1,025,908 sequential queries from 211,058 users of Excite, a major Internet search service. Twenty-eight terms were used to identify queries for both still and moving images, resulting in a subset of 33,149 image queries by 9855 users. We provide data on: (1) image queries – the number of queries and the number of search terms per user, (2) image search sessions – the number of queries per user, modifications made to subsequent queries in a session, and (3) image terms – their rank/frequency distribution and the most highly used search terms. On average, there were 3.36 image queries per user containing an average of 3.74 terms per query. Image queries contained a large number of unique terms. The most frequently occurring image related terms appeared less than 10% of the time, with most terms occurring only once. We contrast this to earlier work by P.G.B. Enser, Journal of Documentation 51 (2) (1995) 126–170, who examined written queries for pictorial information in a non-digital environment. Implications for the development of models for visual information retrieval, and for the design of Web search engines are discussed.     

Unsupervised graph-based rank aggregation for improved retrieval

This paper presents a robust and comprehensive graph-based rank aggregation approach, used to combine results of isolated ranker models in retrieval tasks. The method follows an unsupervised scheme, which is independent of how the isolated ranks are formulated. Our approach is able to combine arbitrary models, defined in terms of different ranking criteria, such as those based on textual, image or hybrid content representations.We reformulate the ad-hoc retrieval problem as a document retrieval based on fusion graphs, which we propose as a new unified representation model capable of merging multiple ranks and expressing inter-relationships of retrieval results automatically. By doing so, we claim that the retrieval system can benefit from learning the manifold structure of datasets, thus leading to more effective results. Another contribution is that our graph-based aggregation formulation, unlike existing approaches, allows for encapsulating contextual information encoded from multiple ranks, which can be directly used for ranking, without further computations and post-processing steps over the graphs. Based on the graphs, a novel similarity retrieval score is formulated using an efficient computation of minimum common subgraphs. Finally, another benefit over existing approaches is the absence of hyperparameters.A comprehensive experimental evaluation was conducted considering diverse well-known public datasets, composed of textual, image, and multimodal documents. Performed experiments demonstrate that our method reaches top performance, yielding better effectiveness scores than state-of-the-art baseline methods and promoting large gains over the rankers being fused, thus demonstrating the successful capability of the proposal in representing queries based on a unified graph-based model of rank fusions.     

Automatic building of new Field Association word candidates using search engine

With increasing popularity of the Internet and tremendous amount of on-line text, automatic document classification is important for organizing huge amounts of data. Readers can know the subject of many document fields by reading only some specific Field Association (FA) words. Document fields can be decided efficiently if there are many FA words and if the frequency rate is high. This paper proposes a method for automatically building new FA words. A WWW search engine is used to extract FA word candidates from document corpora. New FA word candidates in each field are automatically compared with previously determined FA words. Then new FA words are appended to an FA word dictionary. From the experiential results, our new system can automatically appended around 44% of new FA words to the existence FA word dictionary. Moreover, the concentration ratio 0.9 is also effective for extracting relevant FA words that needed for the system design to build FA words automatically.     

A framework for pre-processing of social media feeds based on integrated local knowledge base

Most of the previous studies on the semantic analysis of social media feeds have not considered the issue of ambiguity that is associated with slangs, abbreviations, and acronyms that are embedded in social media posts. These noisy terms have implicit meanings and form part of the rich semantic context that must be analysed to gain complete insights from social media feeds. This paper proposes an improved framework for pre-processing of social media feeds for better performance. To do this, the use of an integrated knowledge base (ikb) which comprises a local knowledge source (Naijalingo), urban dictionary and internet slang was combined with the adapted Lesk algorithm to facilitate semantic analysis of social media feeds. Experimental results showed that the proposed approach performed better than existing methods when it was tested on three machine learning models, which are support vector machines, multilayer perceptron, and convolutional neural networks. The framework had an accuracy of 94.07% on a standardized dataset, and 99.78% on localised dataset when used to extract sentiments from tweets. The improved performance on the localised dataset reveals the advantage of integrating the use of local knowledge sources into the process of analysing social media feeds particularly in interpreting slangs/acronyms/abbreviations that have contextually rooted meanings.     

基于词共现与词向量的概念层次关系自动抽取模型——以学术论文评价领域为例

【目的/意义】通过概念层次关系自动抽取可以快速地在大数据集上进行细粒度的概念语义层次自动划分, 为后续领域本体的精细化构建提供参考。【方法/过程】首先,在由复合术语和关键词组成的术语集上,通过词频、篇 章频率和语义相似度进行筛选,得到学术论文评价领域概念集;其次,考虑概念共现关系和上下文语义信息,前者 用文献-概念矩阵和概念共现矩阵表达,后者用word2vec词向量表示,通过余弦相似度进行集成,得到概念相似度 矩阵;最后,以关联度最大的概念为聚类中心,利用谱聚类对相似度矩阵进行聚类,得到学术论文评价领域概念层 次体系。【结果/结论】经实验验证,本研究提出的模型有较高的准确率,构建的领域概念层次结构合理。【创新/局限】 本文提出了一种基于词共现与词向量的概念层次关系自动抽取模型,可以实现概念层次关系的自动抽取,但类标 签确定的方法比较简单,可以进一步探究。     

基于本体的信息系统引论

Since Tim Bemers-Lee, current W3C chairman, first proposed the concept of Semantic Web, it is be-coming a hot topic in computer information processing area. Ontologies are playing a key role in the Semantic Web, ex-tending syntactic interoperability to semantic intemperability by providing a source of shared and precisely defined terms.The paper analyzes the requirement of information systems for ontology languages. The current popular ontology languages are also discussed.     

Residue matrices and related properties in two-variable reactance functions

In this paper, the characterization of a two-variable reactance polynomial φ(λ,μ) is given in terms of the residue matrices of a single variable reactance matrix, Y(λ). Specificially, if Y(λ) is expressed in terms of its partial fraction as Y(λ)=λH_∞+Σ$\text{H}{}_{\mathrm{i}}\text{+jβ}{}_{\mathrm{i}}\text{λ+jω}{}_{\mathrm{i}}$+G where the residue matrices in general are p.s.d. Hermitian, then the ranks of these residue matrices are fixed in relation to the construction of φ(λ,μ) as the determinant of the two-variable reactance matrix μ1+Y(λ). Three theorems concerning these ranks—one each corresponding to the finite poles, poles at ∞ and the behaviour at λ=0 of Y(λ) are stated and proved. Several properties following from these theorems are studied. Also, implications of these theorems from a network theoretic point of view, like the minimum number of gyrators required to synthesize Y(λ) to yield the specific type of φ(λ,μ) etc., are studied. In the sequel, the concept of “generalized compact pole conditions” is introduced. Finally, these results are applied for the generation of two-variable reactance functions and matrices.     

Neural embedding-based indices for semantic search

Traditional information retrieval techniques that primarily rely on keyword-based linking of the query and document spaces face challenges such as the vocabulary mismatch problem where relevant documents to a given query might not be retrieved simply due to the use of different terminology for describing the same concepts. As such, semantic search techniques aim to address such limitations of keyword-based retrieval models by incorporating semantic information from standard knowledge bases such as Freebase and DBpedia. The literature has already shown that while the sole consideration of semantic information might not lead to improved retrieval performance over keyword-based search, their consideration enables the retrieval of a set of relevant documents that cannot be retrieved by keyword-based methods. As such, building indices that store and provide access to semantic information during the retrieval process is important. While the process for building and querying keyword-based indices is quite well understood, the incorporation of semantic information within search indices is still an open challenge. Existing work have proposed to build one unified index encompassing both textual and semantic information or to build separate yet integrated indices for each information type but they face limitations such as increased query process time. In this paper, we propose to use neural embeddings-based representations of term, semantic entity, semantic type and documents within the same embedding space to facilitate the development of a unified search index that would consist of these four information types. We perform experiments on standard and widely used document collections including Clueweb09-B and Robust04 to evaluate our proposed indexing strategy from both effectiveness and efficiency perspectives. Based on our experiments, we find that when neural embeddings are used to build inverted indices; hence relaxing the requirement to explicitly observe the posting list key in the indexed document: (a) retrieval efficiency will increase compared to a standard inverted index, hence reduces the index size and query processing time, and (b) while retrieval efficiency, which is the main objective of an efficient indexing mechanism improves using our proposed method, retrieval effectiveness also retains competitive performance compared to the baseline in terms of retrieving a reasonable number of relevant documents from the indexed corpus.     

A conceptual framework for fuzzy query processing—A step toward very intelligent database systems

This paper is concerned with techniques for fuzzy query processing in a database system. By a fuzzy query we mean a query which uses imprecise or fuzzy predicates (e.g. AGE = “VERY YOUNG”, SALARY = “MORE OR LESS HIGH”, YEAR-OF-EMPLOYMENT = “RECENT”, SALARY ? 20,000, etc.). As a basis for fuzzy query processing, a fuzzy retrieval system based on the theory of fuzzy sets and linguistic variables is introduced. In our system model, the first step in processing fuzzy queries consists of assigning meaning to fuzzy terms (linguistic values), of a term-set, used for the formulation of a query. The meaning of a fuzzy term is defined as a fuzzy set in a universe of discourse which contains the numerical values of a domain of a relation in the system database.The fuzzy retrieval system developed is a high level model for the techniques which may be used in a database system. The feasibility of implementing such techniques in a real environment is studied. Specifically, within this context, techniques for processing simple fuzzy queries expressed in the relational query language SEQUEL are introduced.     

Ranking of field association terms using Co-word analysis

Information retrieval involves finding some desired information in a store of information or a database. In this paper, Co-word analysis will be used to achieve a ranking of a selected sample of FA terms. Based on this ranking a better arranging of search results can be achieved. Experimental results achieved using 41 MB of data (7660 documents) in the field of sports. The corpus was collected from CNN newspaper, sports field. This corpus was chosen to be distributed over 11 sub-fields of the field sports from the experimental results, the average precision increased by 18.3% after applying the proposed arranging scheme depending on the absolute frequency to count the terms weights, and the average precision increased by 17.2% after applying the proposed arranging scheme depending on a formula based on “TF∗IDF” to count the terms weights.     

Document indexing: a concept-based approach to term weight estimation

Traditional index weighting approaches for information retrieval from texts depend on the term frequency based analysis of the text contents. A shortcoming of these indexing schemes, which consider only the occurrences of the terms in a document, is that they have some limitations in extracting semantically exact indexes that represent the semantic content of a document. To address this issue, we developed a new indexing formalism that considers not only the terms in a document, but also the concepts. In this approach, concept clusters are defined and a concept vector space model is proposed to represent the semantic importance degrees of lexical items and concepts within a document. Through an experiment on the TREC collection of Wall Street Journal documents, we show that the proposed method outperforms an indexing method based on term frequency (TF), especially in regard to the few highest-ranked documents. Moreover, the index term dimension was 80% lower for the proposed method than for the TF-based method, which is expected to significantly reduce the document search time in a real environment.     

Improving relational similarity measurement using symmetries in proportional word analogies

Measuring the similarity between the semantic relations that exist between words is an important step in numerous tasks in natural language processing such as answering word analogy questions, classifying compound nouns, and word sense disambiguation. Given two word pairs (A, B) and (C, D), we propose a method to measure the relational similarity between the semantic relations that exist between the two words in each word pair. Typically, a high degree of relational similarity can be observed between proportional analogies (i.e. analogies that exist among the four words, A is to B such as C is to D). We describe eight different types of relational symmetries that are frequently observed in proportional analogies and use those symmetries to robustly and accurately estimate the relational similarity between two given word pairs. We use automatically extracted lexical-syntactic patterns to represent the semantic relations that exist between two words and then match those patterns in Web search engine snippets to find candidate words that form proportional analogies with the original word pair. We define eight types of relational symmetries for proportional analogies and use those as features in a supervised learning approach. We evaluate the proposed method using the Scholastic Aptitude Test (SAT) word analogy benchmark dataset. Our experimental results show that the proposed method can accurately measure relational similarity between word pairs by exploiting the symmetries that exist in proportional analogies. The proposed method achieves an SAT score of 49.2% on the benchmark dataset, which is comparable to the best results reported on this dataset.     

An information-pattern-based approach to novelty detection

In this paper, a new novelty detection approach based on the identification of sentence level information patterns is proposed. First, “novelty” is redefined based on the proposed information patterns, and several different types of information patterns are given corresponding to different types of users’ information needs. Second, a thorough analysis of sentence level information patterns is elaborated using data from the TREC novelty tracks, including sentence lengths, named entities (NEs), and sentence level opinion patterns. Finally, a unified information-pattern-based approach to novelty detection (ip-BAND) is presented for both specific NE topics and more general topics. Experiments on novelty detection on data from the TREC 2002, 2003 and 2004 novelty tracks show that the proposed approach significantly improves the performance of novelty detection in terms of precision at top ranks. Future research directions are suggested.     

Strong delay-independent stability of linear delay systems

This paper presents a new necessary and sufficient condition for testing the strong delay-independent stability of linear systems subject to a single delay. The proposed method follows from the use of matrix polynomials constraints and the Kalman–Yakubovich–Popov lemma. The resulting condition can be checked exactly by solving a feasibility problem in terms of a linear matrix inequality (LMI). Simple numerical examples are given to show the effectiveness of the proposed method.