Retrieval Domain Research Articles

Enhancing comprehensibility of software clustering results

As requirements of organisations change, so do the software systems within them. When changes are carried out under tough deadlines, software developers often do not follow software engineering principles, which results in deteriorated structure of the software. A badly structured system is difficult to understand for further changes. To improve structure, re-modularisation may be carried out. Clustering techniques have been used to facilitate automatic re-modularisation. However, clusters produced by clustering algorithms are difficult to comprehend unless they are labelled appropriately. Manual assignment of labels is tiresome, thus efforts should be made towards automatic cluster label assignment. In this study, the authors focus on facilitating comprehension of software clustering results by automatically assigning meaningful labels to clusters. To assign labels, the authors use term weighting schemes borrowed from the domain of information retrieval and text categorisation. Although some term weighting schemes have been used by researchers for software cluster labelling, there is a need to analyse the term weighting schemes and related issues to identify the strengths and weaknesses of these schemes for software cluster labelling. In this context, the authors analyse the behaviour of seven well-known term weighting schemes. Also, they perform the experiments on five software systems to identify software characteristics which affect the labelling behaviour of the term weighting schemes.

Tuning user profiles based on analyzing dynamic preference in document retrieval systems

Modeling users' preferences and needs is one of the most important personalization tasks in information retrieval domain. In this paper a model for user profile tuning in document retrieval systems is considered. Methods for tuning the user profile based on analysis of user preferences dynamics are experimentally evaluated to check whether with growing history of user activity the created user profile can converge to his preferences. As the statistical analysis of series of simulations has shown, proposed methods of user profile actualization are effective in the sense that the distance between user preferences and his profile becomes smaller and smaller along with time.

Evaluating retrieval models using retrievability measurement

Evaluation is the main driving force in research, development and applications related to information retrieval (IR). In the traditional IR evaluation paradigm a list of query topics along with their relevance judgments are given. The main limitation of this kind of evaluation paradigm is that it focuses almost exclusively on a small set of judged documents and does not consider what influence the given retrieval models have on accessing all the relevant information in the collection. This is particularly important for recall oriented retrieval applications where we want to ensure that that everything relevant has been found. In this thesis we analyze the effectiveness of retrieval models from the documents retrievability point of view. We focus particularly on the retrieval bias of different retrieval models, and try to examine to what extent this bias restricts the users in retrieving relevant information. We explore this research with the help of three factors. First, we analyze the relationship between different characteristics of queries and retrievability. This is important from the query generation point of view, since in case of exhaustive queries, it is practically infeasible to complete retrievability approximation in reasonable time. The strong correlation between retrievability and query characteristics allows us to approximate the retrievability score accurately with the help of a query subset without processing an exhaustive number of queries. After this, we examine to what extent the retrievability and other IR effectiveness measures are related to each other. This specifically helps us to understand to what extent it is possible to automatically rank the effectiveness of retrieval models on the basis of their retrieval bias. This also offers a basis for optimizing retrieval systems for specific collections without the need to provide manually annotated ground truth. This is particularly useful for those retrieval domains where it is difficult to obtain a sufficient amount of relevance judgments. At the end we investigate and devise different retrieval strategies for mitigating the effect of low retrievability of documents. These include collection partitioning and query expansion on the basis of improved pseudo relevance feedback selection. The work present in this thesis provides an a novel approach for the evaluation and optimization of retrieval models particularly for recall oriented retrieval domains, where the focus is on retrieving all relevant information but not just retrieving a subset of relevant information. Available online at: http://www.ifs.tuwien.ac.at/~bashir/shariqbashir_phd_thesis.pdf.

Topic modelling of clickthrough data in image search

In this paper we explore the benefits of latent variable modelling of clickthrough data in the domain of image retrieval. Clicks in image search logs are regarded as implicit relevance judgements that express both user intent and important relations between selected documents. We posit that clickthrough data contains hidden topics and can be used to infer a lower dimensional latent space that can be subsequently employed to improve various aspects of the retrieval system. We use a subset of a clickthrough corpus from the image search portal of a news agency to evaluate several popular latent variable models in terms of their ability to model topics underlying queries. We demonstrate that latent variable modelling reveals underlying structure in clickthrough data and our results show that computing document similarities in the latent space improves retrieval effectiveness compared to computing similarities in the original query space. These results are compared with baselines using visual and textual features. We show performance substantially better than the visual baseline, which indicates that content-based image retrieval systems that do not exploit query logs could improve recall and precision by taking this historical data into account.

ATOB algorithm: an automatic ontology construction for Thai legal sentences retrieval

Ontology plays an important role in knowledge representation, especially in the domain of information retrieval. However, building ontology remains a challenging problem because it is a time-consuming task for experts. To overcome these drawbacks, we propose a novel approach called the Automatic Thai Legal Ontology Building (ATOB) algorithm for automatic legal ontology building and to improve the court sentences retrieval process. The ATOB can automatically generate seed ontology and expand the ontology using Thai legal terminology, i.e. TLlexicon. The expansion process is terminated automatically by the threshold parameter. Moreover, the ATOB applies the concept of the ant colony algorithm to improve the court sentences retrieval process. We conclude that the effective ontology should be weight-embedded. The empirical results demonstrate that the performance of the ATOB algorithm is better than that of the traditional search method. The performance figures for the ATOB framework measured in terms of precision, recall, F-measure and diversity are 0.90, 0.91, 0.90 and 0.39, respectively.

AGRIX: An Ontology Based Agricultural Expertise Retrieval Framework

Generally, information retrieval (IR) performs keyword search based on the user query to find a set of relevant documents. In the domain of agricultural expertise retrieval, the goal is to find a group of experts who has knowledge in agriculture (by using publications as the evidence) specified by the input query. Typical publication IR systems could sometimes return the search result sets, which consist of a huge amount of publications. Some of the returned publications are not relevant to the individual user’s information need. In this paper, an ontology based agricultural expertise retrieval framework called AGRIX is proposed with the focus on the ontology creation to cover three following aspects: (1) expert profiles and publications, (2) type of plants and (3) problem solving. To build the ontology model, we used a set of publications (1,249 records) which was collected from the Thai national AGRIS center, Bureau of Library Kasetsart University. In addition, a set of inference rules is created to support the expertise retrieval task. By using AGRIX to implement an agricultural expertise retrieval, users can search for experts in two perspectives, plant (e.g., rice, sugar canes) and problem solving (e.g., plant diseases, fertilizers).

A Review on the Cross and Multilingual Information Retrieval

In this paper we explore some of the most important areas of information retrieval. In particular, Crosslingual Information Retrieval (CLIR) and Multilingual Information Retrieval (MLIR). CLIR deals with asking questions in one language and retrieving documents in different language. MLIR deals with asking questions in one or more languages and retrieving documents in one or more different languages. With an increasingly globalized economy, the ability to find information in other languages is becoming a necessity. We also presented the evaluation initiatives of information retrieval domain. Finally we have presented the overall review of the research works in Indian and Foreign languages.

ELDAMeth: An agent-oriented methodology for simulation-based prototyping of distributed agent systems

In application domains, such as distributed information retrieval, content management and distribution, e-Commerce, the agent-based computing paradigm has been demonstrated to be effective for the analysis, design and implementation of distributed software systems. In particular, several agent-oriented methodologies, incorporating suitable agent models, frameworks and tools, have been to date defined to support the development lifecycle of distributed agent systems (DAS). However, few of them provide effective validation methods to analyze design objects at different degrees of refinement before their actual implementation and deployment. In this paper, ELDAMeth, a simulation-based methodology for DAS, which enables rapid prototyping based on visual programming, validation, and automatic code generation for JADE-based DAS, is presented. ELDAMeth can be used both stand-alone for the modeling and evaluation of DAS and coupled with other agent-oriented methodologies for enhancing them with simulation-based validation. In particular, the proposed methodology, which is based on the ELDA (Event-driven Lightweight Distilled StateCharts-based Agents) agent model, provides key programming abstractions (event-driven computation, multi-coordination, and coarse-grained strong mobility) very suitable for highly dynamic distributed computing and is supported by a CASE tool-driven iterative process seamlessly covering the detailed design, simulation, and implementation phases of DAS. A simple yet effective case study in the distributed information retrieval domain is used to illustrate the proposed methodology.

A query expansion framework in image retrieval domain based on local and global analysis

We present an image retrieval framework based on automatic query expansion in a concept feature space by generalizing the vector space model of information retrieval. In this framework, images are represented by vectors of weighted concepts similar to the keyword-based representation used in text retrieval. To generate the concept vocabularies, a statistical model is built by utilizing Support Vector Machine (SVM)-based classification techniques. The images are represented as “bag of concepts” that comprise perceptually and/or semantically distinguishable color and texture patches from local image regions in a multi-dimensional feature space. To explore the correlation between the concepts and overcome the assumption of feature independence in this model, we propose query expansion techniques in the image domain from a new perspective based on both local and global analysis. For the local analysis, the correlations between the concepts based on the co-occurrence pattern, and the metrical constraints based on the neighborhood proximity between the concepts in encoded images, are analyzed by considering local feedback information. We also analyze the concept similarities in the collection as a whole in the form of a similarity thesaurus and propose an efficient query expansion based on the global analysis. The experimental results on a photographic collection of natural scenes and a biomedical database of different imaging modalities demonstrate the effectiveness of the proposed framework in terms of precision and recall.

Aerosol single scattering albedo retrieved from ground-based measurements in the UV and visible region

Abstract. Estimates of Aerosol Single Scattering Albedo (SSA) from ground-based spectral measurements in the UV-visible are conducted at Villeneuve d'Ascq (VdA) in France. In order to estimate this parameter, measurements of global and diffuse UV-visible solar irradiances performed under cloud-free conditions since 2003 with a spectroradiometer operated by the Laboratoire d'Optique Atmosphérique (LOA) are used. The technique consists in comparing the measured irradiance values to modelled irradiances computed for various SSA. The retrieval is restricted to the 330–450 nm range to avoid ozone influence. For validation purpose, the retrieved values of SSA at 440 nm are compared to the ones obtained from sunphotometer measurements of the AERONET/PHOTONS network available on the LOA site. The results are rather satisfying: for the period 2003–2006 the Root Mean Square (RMS) of the differences is about 0.05, this value is within the uncertainty domain of retrieval of both products. Distinction between days characterized by different aerosol content, by means of the aerosol optical thickness (AOT) retrieved from ground-based measurements at the same wavelength, shows that the comparisons between both products are better when AOT are higher. Indeed in case AOT are greater than 0.2, the RMS is 0.031 compared to 0.060 for days with an AOT lower than 0.2. The SSA estimated at 340 and 380 nm from ground-based spectra are also studied, though no validation can be carried out with sunphotometer data (440 nm is the shortest wavelength at which the SSA is provided by the network). The good comparisons observed at 440 nm can let assume that the SSA retrieved from spectroradiometer measurements at the two other wavelengths are also obtained with a good confidence level. Thus these values in the UV range can be used to complete aerosol data provided by AERONET/PHOTONS at VdA. Moreover they can be used for a best knowledge of the aerosol absorption that is necessary to quantify the error on surface UV irradiances estimated from satellites.

Personalised video retrieval

A challenging problem in the user profiling domain is to create profiles of users of retrieval systems. This problem even exacerbates in the multimedia domain. Due to the Semantic Gap, the difference between low-level data representation of videos and the higher concepts users associate with videos, it is not trivial to understand the content of multimedia documents and to find other documents that the users might be interested in. A promising approach to ease this problem is to set multimedia documents into their semantic contexts. The semantic context can lead to a better understanding of the personal interests. Knowing the context of a video is useful for recommending users videos that match their information need. By exploiting these contexts, videos can also be linked to other, contextually related videos. From a user profiling point of view, these links can be of high value to recommend semantically related videos, hence creating a semantic-based user profile. This thesis introduces a semantic user profiling approach for news video retrieval, which exploits a generic ontology to put news stories into its context. Major challenges which inhibit the creation of such semantic user profiles are the identification of user's long-term interests and the adaptation of retrieval results based on these personal interests. Most personalisation services rely on users explicitly specifying preferences, a common approach in the text retrieval domain. By giving explicit feedback, users are forced to update their need, which can be problematic when their information need is vague. Furthermore, users tend not to provide enough feedback on which to base an adaptive retrieval algorithm. Deviating from the method of explicitly asking the user to rate the relevance of retrieval results, the use of implicit feedback techniques helps by learning user interests unobtrusively. The main advantage is that users are relieved from providing feedback. A disadvantage is that information gathered using implicit techniques is less accurate than information based on explicit feedback.

The effect of G72 genotype on neural correlates of memory encoding and retrieval

Polymorphisms in the G72 (also named d-amino acid oxidase activator, DAOA) gene increase the vulnerability for schizophrenia and affective psychosis. Three recent genetic neuroimaging studies showed that variation in G72 influences the brain activity in the medial temporal lobe (MTL), supporting the hypothesis that G72 might play a modulatory role on brain activity in MTL structures. In the present study we therefore investigated the effect of G72 on the neural correlates of long-term memory encoding and retrieval in a large sample of healthy subjects (n=83) using functional magnetic resonance imaging. A face encoding and a face retrieval memory task were chosen because on the one hand they specifically activate MTL structures and on the other hand they tap into memory processes that are compromised in patients with schizophrenia and affective disorder. Despite a strong a-priori hypothesis of genotype group activation differences in the MTL along with a large sample size we did neither find an effect of G72 genotype status on brain activity in the MTL nor in any other brain regions. The present data therefore do not support the view of a general modulatory role of G72 on MTL brain activity, at least not in the domain of long-term memory encoding and retrieval. Our results highlight the importance of replication studies in genetic neuroimaging.

Structured-image retrieval invariant to rotation, scaling and translation

In Content Based Image Retrieval (CBIR) domain a new methodology is proposed in this paper; this methodology is based on statistical features such as Hu invariant moments (invariant to scale and translation) and a correlation measure (between 2 images). In order to attack image rotation problem, we use principal components analysis. Image comparison is done by correlation. Proposed scheme was tested with hundred of structured images like: electronic circuits, cell phones, cars, trees, leaves, grass, glasses, tables, etc. Features characteristics are extracted from invariant moments, taken from window size estimation. From a query image, a set of features were estimated in order to compare to a set of images. Correlation function is applied to get image similarity, it is obtained a believe percentage value. As a methodology conclusion, we found that the invariant moments combined with principal component analysis gives excellent results in image retrieval task. An exhaustive study was performed with 1000 images. We also evaluated the impact of noise on the images testing additive Gaussian random noise.

A NEW METHOD OF MEDICAL IMAGE RETRIEVAL BASED ON COLOR–TEXTURE CORRELOGRAM AND GTI MODEL

This paper presents a method for endoscopic image retrieval based on color–texture correlogram and Generalized Tversky's Index (GTI) model. First we define a new image feature named color–texture correlogram, which is the extension of color correlogram. The texture image extracted by texture spectrum algorithm is combined with color feature vector, and then we calculate the spatial correlation of color–texture feature vector. Similarity metric is also the key technology during domain of image retrieval, GTI model is used in medical image retrieval for similarity metric, and the technique of relevance feedback is used in the algorithm to enhance the efficiency of retrieval. Experimental results show that the method discussed in this paper is much more effective.

Students’ attitudes towards e‐books in a Scottish Higher Education Institute

PurposeThe purpose of this paper is to present the third of three inter‐related experiments investigating the use and usability of e‐books in Higher Education based on experiments conducted at the University of Strathclyde. This study has looked in greater detail at user interactions with e‐books for reference purposes by focusing on searching and browsing tasks using three search tools: back‐of the‐book index (BoBI), table of contents (ToC) and full text search (FTS).Design/methodology/approachThis study was carried out by subject‐specific users and using a between‐subjects approach. The target population was MSc and research students in the Department of Computer and Information Sciences, at the University of Strathclyde and involved a total of 45 responses.FindingsThe study found that a BoBI was more efficient compared to a ToC and FTS tool for finding information in an e‐book environment. A BoBI was found to perform the best for accurately finding relevant content in e‐books. The usability evaluation also found that a BoBI was more useful compared to a ToC for finding information in an e‐book environment.Research limitations/implicationsThe study was focused only on the usability of e‐books, and in particular on retrieval performance, user satisfaction and preferences regarding BoBI, ToC and FTS, and not on other features such as the user interface. The e‐book usability evaluation was constrained in so far as the e‐books used were: non‐fiction; in the domain of information retrieval; e‐books that already had BoBIs with hyperlinks from the BoBI to the text; e‐books that had ToCs with hyperlinks; e‐books that had FTS tools; and e‐books that were available in PDF format.Practical implicationsThe study is important in gaining a better understanding of the retrieval performance of three search tools (BoBI, ToC and FTS) for browsing for relevant, and searching for specific, information in e‐books. This will be of value for designing better e‐books and access systems.Originality/valueThe strengths and novelty of this study are the methodology that was used, the comprehensive inter‐comparison of tools, and the size of the population. The findings have supported empirically – through an assessment of the performances of BoBIs and ToCs – the need for an enhanced library catalogue system in order to improve users’ browsing and searching capabilities for relevant book content.

A Comparative Performance Evaluation Model of Mobile Agent Versus Remote Method Invocation for Information Retrieval

The development of distributed systems has been affected by the need to accommodate an increasing degree of flexibility, adaptability, and autonomy. The Mobile Agent technology is emerging as an alternative to build a smart generation of highly distributed systems. In this work, we investigate the performance aspect of agent-based technologies for information retrieval. We present a comparative performance evaluation model of Mobile Agents versus Remote Method Invocation by means of an analytical approach. We demonstrate the effectiveness of mobile agents for dynamic code deployment and remote data processing by reducing total latency and at the same time producing minimum network traffic. We argue that exploiting agent-based technologies significantly enhances the performance of distributed systems in the domain of information retrieval.

The sixth Dutch-Belgian Information Retrieval workshop

The 6th issue of the Dutch-Belgian Information Retrieval workshop took place at March 13 and 14, 2006 and was hosted by TNO Information and Communication Technology in Delft, The Netherlands. The primary aim of the DIR workshops is to provide an international meeting place where researchers from the domain of information retrieval and related disciplines, can exchange information and present new research developments. This year, there was a special focus on contributions focusing on domain-specific retrieval tasks.

Using string matching to detect video transitions

The detection of shot boundaries in videos captures the structure of the image sequences by the identification of transitional effects. This task is important in the video indexing and retrieval domain. The video slice or visual rhythm is a single two-dimensional image sampling that has been used to detect several types of video events, including transitions. We use the longest common subsequence (LCS) between two strings to transform the video slice into one-dimensional signals obtaining a highly simplified representation of the video content. We also developed a chain of mathematical morphology operations over these signals leading to the detection of the most frequent video transitions, namely, cut, fade, and wipe. The algorithms are tested with success with various genres of videos.

Handling temporal heterogeneous data for content-based management of large video collections

Video document retrieval is now an active part of the domain of multimedia retrieval. However, unlike for other media, the management of a collection of video documents adds the problem of efficiently handling an overwhelming volume of temporal data. Challenges include balancing efficient content modeling and storage against fast access at various levels. In this paper, we detail the framework we have built to accommodate our developments in content-based multimedia retrieval. We show that not only our framework facilitates the development of processing and indexing algorithms but it also opens the way to several other possibilities such as rapid interface prototyping or retrieval algorithm benchmarking. Here, we discuss our developments in relation to wider contexts such as MPEG-7 and the TREC Video Track.

사회적 네비게이션 기반 사회적 검색

웹기반 교육 자료들이 폭발적으로 증가함에 따라 적합한 자료들에 보다 효과적으로 접근할 수 있는 방법이 요구되고 있다. 이러한 새로운 방법들 중의 하나로 사회적 네비게이션(social navigation) 기반의 사회적 검색(social searching)이 정보 검색 분야에서 제시되었는데, 이는 동료 이용자들로부터 제공된 정보를 바탕으로 검색 결과의 향상을 추구하는 기법이다. 본 연구에서는 개인화와 사회적 네비게이션에 근거한 웹 기반 사회적 검색 시스템을 구축하였으며 이용자 연구를 통해 이용자에게 적합하고 필수적인 정보를 제공할 수 있는 방법이라는 것을 검증하려 하였다.