Lexical Matching Research Articles

MATCHING ALTERNATIVE ADDRESSES: A SEMANTIC WEB APPROACH

Abstract. Rapid development of crowd-sourcing or volunteered geographic information (VGI) provides opportunities for authoritatives that deal with geospatial information. Heterogeneity of multiple data sources and inconsistency of data types is a key characteristics of VGI datasets. The expansion of cities resulted in the growing number of POIs in the OpenStreetMap, a well-known VGI source, which causes the datasets to outdate in short periods of time. These changes made to spatial and aspatial attributes of features such as names and addresses might cause confusion or ambiguity in the processes that require feature’s literal information like addressing and geocoding. VGI sources neither will conform specific vocabularies nor will remain in a specific schema for a long period of time. As a result, the integration of VGI sources is crucial and inevitable in order to avoid duplication and the waste of resources. Information integration can be used to match features and qualify different annotation alternatives for disambiguation. This study enhances the search capabilities of geospatial tools with applications able to understand user terminology to pursuit an efficient way for finding desired results. Semantic web is a capable tool for developing technologies that deal with lexical and numerical calculations and estimations. There are a vast amount of literal-spatial data representing the capability of linguistic information in knowledge modeling, but these resources need to be harmonized based on Semantic Web standards. The process of making addresses homogenous generates a helpful tool based on spatial data integration and lexical annotation matching and disambiguating.

A framework for ontology-based question answering with application to parasite immunology.

BackgroundLarge quantities of biomedical data are being produced at a rapid pace for a variety of organisms. With ontologies proliferating, data is increasingly being stored using the RDF data model and queried using RDF based querying languages. While existing systems facilitate the querying in various ways, the scientist must map the question in his or her mind to the interface used by the systems. The field of natural language processing has long investigated the challenges of designing natural language based retrieval systems. Recent efforts seek to bring the ability to pose natural language questions to RDF data querying systems while leveraging the associated ontologies. These analyze the input question and extract triples (subject, relationship, object), if possible, mapping them to RDF triples in the data. However, in the biomedical context, relationships between entities are not always explicit in the question and these are often complex involving many intermediate concepts.ResultsWe present a new framework, OntoNLQA, for querying RDF data annotated using ontologies which allows posing questions in natural language. OntoNLQA offers five steps in order to answer natural language questions. In comparison to previous systems, OntoNLQA differs in how some of the methods are realized. In particular, it introduces a novel approach for discovering the sophisticated semantic associations that may exist between the key terms of a natural language question, in order to build an intuitive query and retrieve precise answers. We apply this framework to the context of parasite immunology data, leading to a system called AskCuebee that allows parasitologists to pose genomic, proteomic and pathway questions in natural language related to the parasite, Trypanosoma cruzi. We separately evaluate the accuracy of each component of OntoNLQA as implemented in AskCuebee and the accuracy of the whole system. AskCuebee answers 68 % of the questions in a corpus of 125 questions, and 60 % of the questions in a new previously unseen corpus. If we allow simple corrections by the scientists, this proportion increases to 92 %.ConclusionsWe introduce a novel framework for question answering and apply it to parasite immunology data. Evaluations of translating the questions to RDF triple queries by combining machine learning, lexical similarity matching with ontology classes, properties and instances for specificity, and discovering associations between them demonstrate that the approach performs well and improves on previous systems. Subsequently, OntoNLQA offers a viable framework for building question answering systems in other biomedical domains.Electronic supplementary materialThe online version of this article (doi:10.1186/s13326-015-0029-x) contains supplementary material, which is available to authorized users.

Retrieving taxa names from large biodiversity data collections using a flexible matching workflow

In the domain of biological classification there are several taxon name matching services that can search for a species scientific name in a large collection of taxonomic names. Many of these services are available online, and many others run on computers of individual scientists. While these systems may work very well, most suffer from the fact that the list of names used as a reference, and the criteria to decide on a match, are hard-coded in the engine that performs the name matching. In this paper we present BiOnym, a taxon name matching system that separates reference namelists, search criteria and matching engine. The user is offered a choice of several taxonomic reference lists, including the option to upload his/her own list onto the system. Furthermore, BiOnym is a flexible workflow, which embeds and combines techniques using lexical matching algorithms as well as expert knowledge. It is also an open platform allowing developers to contribute with new techniques. In this paper we demonstrate the benefits brought by this approach in terms of the efficiency and effectiveness of the information retrieval process with respect to other solutions.

BiobankConnect: software to rapidly connect data elements for pooled analysis across biobanks using ontological and lexical indexing.

Objective Pooling data across biobanks is necessary to increase statistical power, reveal more subtle associations, and synergize the value of data sources. However, searching for desired data elements among the thousands of available elements and harmonizing differences in terminology, data collection, and structure, is arduous and time consuming.Materials and methods To speed up biobank data pooling we developed BiobankConnect, a system to semi-automatically match desired data elements to available elements by: (1) annotating the desired elements with ontology terms using BioPortal; (2) automatically expanding the query for these elements with synonyms and subclass information using OntoCAT; (3) automatically searching available elements for these expanded terms using Lucene lexical matching; and (4) shortlisting relevant matches sorted by matching score.Results We evaluated BiobankConnect using human curated matches from EU-BioSHaRE, searching for 32 desired data elements in 7461 available elements from six biobanks. We found 0.75 precision at rank 1 and 0.74 recall at rank 10 compared to a manually curated set of relevant matches. In addition, best matches chosen by BioSHaRE experts ranked first in 63.0% and in the top 10 in 98.4% of cases, indicating that our system has the potential to significantly reduce manual matching work.Conclusions BiobankConnect provides an easy user interface to significantly speed up the biobank harmonization process. It may also prove useful for other forms of biomedical data integration. All the software can be downloaded as a MOLGENIS open source app from http://www.github.com/molgenis, with a demo available at http://www.biobankconnect.org.

The Research and Implementation of Keyword Extraction Technology

Keyword extraction plays an important role in abstract, information retrieval, data mining, text clustering etc. Extracting the keywords from a document can increases the efficiency of retrieval, thus provide great help to efficiently organize the resource. Few writers on the Internet have given the keywords of a document. Artificially extracting the keywords of a document is a great deal of work, so we need a method of extracting the keywords automatically. The paper constructing a verb, function words, stop words etc. small library from the perspective of the Chinese part of speech, realize rapid word segmentation based on the research, analysis, improvement of traditional lexical maximum matching points, and analyze, realize extracting the keywords based on TFIDF(Term Frequency Inverse Document Frequency).

A system for formative assessment and monitoring of students' progress

Assessment plays a central role in any educational process as a way of evaluating the students' knowledge on the concepts associated with learning objectives. The assessment of free-text answers is a process that, besides being very costly in terms of time spent by teachers, may lead to inequities due to the difficulty in applying the same evaluation criteria to all answers. This paper describes a system composed by several modules whose main goal is to work as a formative assessment tool for students and to help teachers creating and assessing exams as well monitoring students' progress. The system automatically creates training exams for students to practice based on questions from previous exams and assists teachers in the creation of evaluation exams with various kinds of information about students' performance. The system automatically assesses training exams to give automatic feedback to students. The correction of free-text answers is based on the syntactic and semantic similarity between the student answers and various reference answers, thus going beyond the simple lexical matching. For this, several pre-processing tasks are performed in order to reduce each answer to its more manageable canonical form. Besides the syntactic and semantic similarity between answers, the way the teacher evaluates the answers is also acquired. To accomplish that, the assessment is done using sub scores defined by the teacher concerning parts of the answer or its subgoals. The system has been trained and tested on exams manually graded by History teachers. There is a good correlation between the evaluation of the instructors and the evaluation performed by our system.

DNorm: disease name normalization with pairwise learning to rank

Motivation: Despite the central role of diseases in biomedical research, there have been much fewer attempts to automatically determine which diseases are mentioned in a text—the task of disease name normalization (DNorm)—compared with other normalization tasks in biomedical text mining research.Methods: In this article we introduce the first machine learning approach for DNorm, using the NCBI disease corpus and the MEDIC vocabulary, which combines MeSH® and OMIM. Our method is a high-performing and mathematically principled framework for learning similarities between mentions and concept names directly from training data. The technique is based on pairwise learning to rank, which has not previously been applied to the normalization task but has proven successful in large optimization problems for information retrieval.Results: We compare our method with several techniques based on lexical normalization and matching, MetaMap and Lucene. Our algorithm achieves 0.782 micro-averaged F-measure and 0.809 macro-averaged F-measure, an increase over the highest performing baseline method of 0.121 and 0.098, respectively.Availability: The source code for DNorm is available at http://www.ncbi.nlm.nih.gov/CBBresearch/Lu/Demo/DNorm, along with a web-based demonstration and links to the NCBI disease corpus. Results on PubMed abstracts are available in PubTator: http://www.ncbi.nlm.nih.gov/CBBresearch/Lu/Demo/PubTatorContact: zhiyong.lu@nih.gov

Modularity, Phase-Phase Faithfulness and prosodification of function words in English

This paper investigates the interface of syntax and phonology in a fully modular view of language, deriving the effects of (morpho)syntactic structure on prosodification without referring to that structure in the phonological computation, contra the use of constraints that map (morpho)syntactic edges or constituents to prosodic ones. The data focus is on function words in English, which receive different prosodic treatment from lexical words. The approach presented here adopts the view of the ‘syntax-all-the-way-down’ approaches, specifically Nanosyntax, which erase the traditional distinction between lexical and functional categories. The paper argues that phonological computation needs to proceed in phases in order to achieve domain mapping while maintaining an input consisting of purely phonological information, and offers a formalization within the Optimality Theory framework by introducing Phase-Phase Faithfulness constraints. Spell-out is attempted at each merge, and is successful when lexical matching is successful. The paper argues that spell-out cannot proceed in chunks but in concentric circles, producing cumulative cyclic input to phonology. An analysis is provided deriving prosodic domains from phases by phonological computation being faithful to the prosodification output of the previous phase. The prosodic word status of lexical words is derived from their status as phase 1 in the derivation.

Learning to Recognize Textual Entailment in Japanese Texts with the Utilization of Machine Translation

Recognizing Textual Entailment (RTE) is a fundamental task in Natural Language Understanding. The task is to decide whether the meaning of a text can be inferred from the meaning of another one. In this article, we conduct an empirical study of recognizing textual entailment in Japanese texts, in which we adopt a machine learning-based approach to the task. We quantitatively analyze the effects of various entailment features, machine learning algorithms, and the impact of RTE resources on the performance of an RTE system. This article also investigates the use of machine translation for the RTE task and determines whether machine translation can be used to improve the performance of our RTE system. Experimental results achieved on benchmark data sets show that our machine learning-based RTE system outperforms the baseline methods based on lexical matching and syntactic matching. The results also suggest that the machine translation component can be utilized to improve the performance of the RTE system.

Quantitative comparison of mapping methods between Human and Mammalian Phenotype Ontology

Researchers use animal studies to better understand human diseases. In recent years, large-scale phenotype studies such as Phenoscape and EuroPhenome have been initiated to identify genetic causes of a species' phenome. Species-specific phenotype ontologies are required to capture and report about all findings and to automatically infer results relevant to human diseases. The integration of the different phenotype ontologies into a coherent framework is necessary to achieve interoperability for cross-species research.Here, we investigate the quality and completeness of two different methods to align the Human Phenotype Ontology and the Mammalian Phenotype Ontology. The first method combines lexical matching with inference over the ontologies' taxonomic structures, while the second method uses a mapping algorithm based on the formal definitions of the ontologies. Neither method could map all concepts. Despite the formal definitions method provides mappings for more concepts than does the lexical matching method, it does not outperform the lexical matching in a biological use case. Our results suggest that combining both approaches will yield a better mappings in terms of completeness, specificity and application purposes.

A supervised framework for resolving coreference in clinical records

A method for the automatic resolution of coreference between medical concepts in clinical records. A multiple pass sieve approach utilizing support vector machines (SVMs) at each pass was used to resolve coreference. Information such as lexical similarity, recency of a concept mention, synonymy based on Wikipedia redirects, and local lexical context were used to inform the method. Results were evaluated using an unweighted average of MUC, CEAF, and B(3) coreference evaluation metrics. The datasets used in these research experiments were made available through the 2011 i2b2/VA Shared Task on Coreference. The method achieved an average F score of 0.821 on the ODIE dataset, with a precision of 0.802 and a recall of 0.845. These results compare favorably to the best-performing system with a reported F score of 0.827 on the dataset and the median system F score of 0.800 among the eight teams that participated in the 2011 i2b2/VA Shared Task on Coreference. On the i2b2 dataset, the method achieved an average F score of 0.906, with a precision of 0.895 and a recall of 0.918 compared to the best F score of 0.915 and the median of 0.859 among the 16 participating teams. Post hoc analysis revealed significant performance degradation on pathology reports. The pathology reports were characterized by complex synonymy and very few patient mentions. The use of several simple lexical matching methods had the most impact on achieving competitive performance on the task of coreference resolution. Moreover, the ability to detect patients in electronic medical records helped to improve coreference resolution more than other linguistic analysis.

Lexical Noise Analysis and Removal in Intelligent Search Engines

In the field of intelligent information retrieval (IR), latent semantic indexing (LSI) is a popular technique used to retrieve information related more in meaning than in lexical matching. This technique overcomes the problems associated with synonymy and polysemy (common causes of inaccuracy in matching algorithms). A core component in the process is the use of the singular value decomposition (SVD) which acts as a mathematical model for the lexical noise in the term document matrix (TDM). This paper investigates various aspects of LSI from the viewpoint of noise modeling and removal in image processing. A discussion about and an investigation into, mathematical modeling for lexical noise in the TDM is presented. The work addresses a definition for noise in text processing and seeks to determine the best structure of the TDM. In other words, the structure of the TDM that would facilitate efficient searching within the LSI.

Mapping the Terms of Medicinal Material and Formula Classification to International Standard Terminology

The current study aims to analyze the acceptance of International Standard Terminology (IST) related to herbs and formulas used in Korea. It also intends to examine limitations of each term source by linking texts for herbal medicine research and formula research used in schools of oriental medicine with medicinal substance-formula classification names within the IST framework. This study examined 64 medicinal classification names of IST, including synonyms, 41 formula classification names, 65 classification names of “Herbal Medicine Study,” 89 medicinal classification names of “Shins Clinical Herbal Medicine Study,” and lastly 83 formula classification names of “Formula Study.” Data on their chief virtue, efficacy and characteristics as medicinal substances were extracted from their definitions, and such data were used to perform Chinese character-English mapping using the IST. The outcomes of the mapping were then analyzed in terms of both lexical matching and semantic matching. In terms of classification names for medicinal substances, “Herbal Medicine Study” had 60.0% lexical matching, whereas “Shins Clinical Herbal Medicine Study” had 48.3% lexical matching. When semantic matching was also applied, “Herbal Medicine Study” showed a value of 87.7% and “Shins Clinical Herbal Medicine Study” 74.2%. In terms of formula classification names, lexical matching was 28.9% of 83 subjects, and when semantic matching was also considered, the value was 30.1%. When the conceptual elements of this study were applied, some IST terms that are classified with other codes were found to be conceptually consistent, and some terms were not accepted due to different depths in the classification systems of each source.

The mental lexicon is fully specified: Evidence from eye-tracking.

Four visual-world experiments, in which listeners heard spoken words and saw printed words, compared an optimal-perception account with the theory of phonological underspecification. This theory argues that default phonological features are not specified in the mental lexicon, leading to asymmetric lexical matching: Mismatching input (pin) activates lexical entries with underspecified coronal stops (tin), but lexical entries with specified labial stops (pin) are not activated by mismatching input (tin). The eye-tracking data failed to show such a pattern. Although words that were phonologically similar to the spoken target attracted more looks than did unrelated distractors, this effect was symmetric in Experiment 1 with minimal pairs (tin-pin) and in Experiments 2 and 3 with words with an onset overlap (peacock-teacake). Experiment 4 revealed that /t/-initial words were looked at more frequently if the spoken input mismatched only in terms of place than if it mismatched in place and voice, contrary to the assumption that /t/ is unspecified for place and voice. These results show that speech perception uses signal-driven information to the fullest, as was predicted by an optimal perception account.

SOLAR versus SERIOL revisited

This paper compares two approaches to modelling orthographic processing, the Self-Organising Lexical Acquisition and Recognition (SOLAR; Davis, 1999, in press) and the Sequential Encoding Regulated by Inputs to Oscillating Letter units (SERIOL; Whitney, 2001, 2004) models, following up on a previous analysis by Whitney (2008). I provide a brief overview of the SOLAR model, and its key similarities to and differences from the SERIOL model, focusing in particular on the different mechanisms underlying the formation of the positional gradient in the two models. I also discuss the neural implementation of the SOLAR model's lexical matching algorithm, and its plausibility. In the final part of the paper I discuss empirical attempts to adjudicate between the two models, focusing on the masked form priming paradigm, as well as the use of theoretical match values to test model predictions. It is concluded that the SOLAR model provides an account of visual word identification that is neurally plausible and that succeeds in explaining critical orthographical similarity data, but that the SERIOL model does not satisfy these constraints.

The spatial coding model of visual word identification.

Visual word identification requires readers to code the identity and order of the letters in a word and match this code against previously learned codes. Current models of this lexical matching process posit context-specific letter codes in which letter representations are tied to either specific serial positions or specific local contexts (e.g., letter clusters). The spatial coding model described here adopts a different approach to letter position coding and lexical matching based on context-independent letter representations. In this model, letter position is coded dynamically, with a scheme called spatial coding. Lexical matching is achieved via a method called superposition matching, in which input codes and learned codes are matched on the basis of the relative positions of their common letters. Simulations of the model illustrate its ability to explain a broad range of results from the masked form priming literature, as well as to capture benchmark findings from the unprimed lexical decision task.

To show or not to show: The effects of item stems and answer options on performance on a multiple-choice listening comprehension test

Abstract The purpose of this study is to examine whether the choice between three multiple-choice listening comprehension test formats results in any difference in listening comprehension test performance. The three formats entail (a) allowing test takers to preview both the question stem and answer options prior to listening; (b) allowing test takers to preview only the answer options; and (c) allowing preview of the question stems, but not answer options. Account is taken of additional, intervening factors relating to the content of the items that were tested; and we report on the extent to which they too may have contributed to difficulty. The results indicate that format (b) produces significantly fewer correct answers than the other two formats. No significant difference in test scores is found between the other two formats. The analysis suggests that the differences may be associated with reliance on a lexical matching strategy on the part of test takers. Implications are drawn for test developers and test takers.

Preorthographical constraints on visual word recognition: Evidence from a case study of developmental surface dyslexia

We investigated the visual word recognition ability of M.T., a young boy with surface dyslexia, by means of a paradigm that measures performance as a function of the eye fixation position within the word, known as the “viewing-position effect” paradigm. In well-achieving readers, the viewing-position effect is mainly determined by factors affecting letter visibility and by lexical constraints on word recognition. We further quantified M.T.'s sensory limitations on letter visibility by computing visual-span profiles—that is, the number of letters recognizable at a glance. Finally, in an ideal-observer's perspective, M.T.'s performance was compared with a parameter-free model combining M.T.'s letter visibility data with a simple lexical matching rule. The results showed that M.T. did not use the whole visual information available on letter identities to recognize words and that preorthographical factors constrained his word recognition performance. The results can be best accounted for by a reduction of the number of letters processed in parallel.

Information Retrieval by Semantic Similarity

Semantic Similarity relates to computing the similarity between conceptually similar but not necessarily lexically similar terms. Typically, semantic similarity is computed by mapping terms to an ontology and by examining their relationships in that ontology. We investigate approaches to computing the semantic similarity between natural language terms (using WordNet as the underlying reference ontology) and between medical terms (using the MeSH ontology of medical and biomedical terms). The most popular semantic similarity methods are implemented and evaluated using WordNet and MeSH. Building upon semantic similarity, we propose the Semantic Similarity based Retrieval Model (SSRM), a novel information retrieval method capable for discovering similarities between documents containing conceptually similar terms. The most effective semantic similarity method is implemented into SSRM. SSRM has been applied in retrieval on OHSUMED (a standard TREC collection available on the Web). The experimental results demonstrated promising performance improvements over classic information retrieval methods utilizing plain lexical matching (e.g., Vector Space Model) and also over state-of-the-art semantic similarity retrieval methods utilizing ontologies.

Optimising the Heuristics in Latent Semantic Indexing for Effective Information Retrieval

Latent Semantic Indexing (LSI) is a famous Information Retrieval (IR) technique that tries to overcome the problems of lexical matching using conceptual indexing. LSI is a variant of vector space model and proved to be 30% more effective. Many studies have reported that good retrieval performance is related to the use of various retrieval heuristics. In this paper, we focus on optimising two LSI retrieval heuristics: term weighting and rank approximation. The results obtained demonstrate that the LSI performance improves significantly with the combination of optimised term weighting and rank approximation.