Search (151 results, page 8 of 8)

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 4144) [ClassicSimilarity], result of:
      0.013336393 = score(doc=4144,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 4144, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=4144)
  0.33333334 = coord(1/3)

Abstract

Indexing documents with descriptors from a multilingual thesaurus is an approach to multilingual information retrieval. However, manual indexing is expensive. Automazed indexing methods in general use terms found in the document. Thesaurus descriptors are complex terms that are often not used in documents or have specific meanings within the thesaurus; therefore most weighting schemes of automated indexing methods are not suited to select thesaurus descriptors. In this paper a linear associative system is described that uses similarity values extracted from a large corpus of manually indexed documents to construct a rank ordering of the descriptors for a given document title. The system is adaptive and has to be tuned with a training sample of records for the specific task. The system was tested on a corpus of some 80.000 bibliographic records. The results show a high variability with changing parameter values. This indicated that it is very important to empirically adapt the model to the specific situation it is used in. The overall median of the manually assigned descriptors in the automatically generated ranked list of all 3.631 descriptors is 14 for the set used to adapt the system and 11 for a test set not used in the optimization process. This result shows that the optimization is not a fitting to a specific training set but a real adaptation of the model to the setting

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 6029) [ClassicSimilarity], result of:
      0.013336393 = score(doc=6029,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 6029, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=6029)
  0.33333334 = coord(1/3)

Abstract

Over the past few years, temporal information processing and temporal database management have increasingly become hot topics. Nevertheless, only a few researchers have investigated these areas in the Chinese language. This lays down the objective of our research: to exploit Chinese language processing techniques for temporal information extraction and concept reasoning. In this article, we first study the mechanism for expressing time in Chinese. On the basis of the study, we then design a general frame structure for maintaining the extracted temporal concepts and propose a system for extracting time-dependent information from Hong Kong financial news. In the system, temporal knowledge is represented by different types of temporal concepts (TTC) and different temporal relations, including absolute and relative relations, which are used to correlate between action times and reference times. In analyzing a sentence, the algorithm first determines the situation related to the verb. This in turn will identify the type of temporal concept associated with the verb. After that, the relevant temporal information is extracted and the temporal relations are derived. These relations link relevant concept frames together in chronological order, which in turn provide the knowledge to fulfill users' queries, e.g., for question-answering (i.e., Q&A) applications

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 601) [ClassicSimilarity], result of:
      0.013336393 = score(doc=601,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 601, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=601)
  0.33333334 = coord(1/3)

Abstract

This article describes an evaluation of the Kea automatic keyphrase extraction algorithm. Document keyphrases are conventionally used as concise descriptors of document content, and are increasingly used in novel ways, including document clustering, searching and browsing interfaces, and retrieval engines. However, it is costly and time consuming to manually assign keyphrases to documents, motivating the development of tools that automatically perform this function. Previous studies have evaluated Kea's performance by measuring its ability to identify author keywords and keyphrases, but this methodology has a number of well-known limitations. The results presented in this article are based on evaluations by human assessors of the quality and appropriateness of Kea keyphrases. The results indicate that, in general, Kea produces keyphrases that are rated positively by human assessors. However, typical Kea settings can degrade performance, particularly those relating to keyphrase length and domain specificity. We found that for some settings, Kea's performance is better than that of similar systems, and that Kea's ranking of extracted keyphrases is effective. We also determined that author-specified keyphrases appear to exhibit an inherent ranking, and that they are rated highly and therefore suitable for use in training and evaluation of automatic keyphrasing systems.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 1845) [ClassicSimilarity], result of:
      0.013336393 = score(doc=1845,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 1845, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=1845)
  0.33333334 = coord(1/3)

Abstract

It may be possible to improve the quality of automatic indexing systems by using complex descriptors, for example, phrases, in addition to the simple descriptors (words or word stems) that are normally used in automatically constructed representations of document content. This study is directed toward the goal of developing effective methods of identifying phrases in natural language text from which good quality phrase descriptors can be constructed. The effectiveness of one method, a simple nonsyntactic phrase indexing procedure, has been tested on five experimental document collections. The results have been analyzed in order to identify the inadequacies of the procedure, and to determine what kinds of information about text structure are needed in order to construct phrase descriptors that are good indicators of document content. Two primary conclusions have been reached: (1) In the retrieval experiments, the nonsyntactic phrase construction procedure did not consistently yield substantial improvements in effectiveness. It is therefore not likely that phrase indexing of this kind will prove to be an important method of enhancing the performance of automatic document indexing and retrieval systems in operational environments. (2) Many of the shortcomings of the nonsyntactic approach can be overcome by incorporating syntactic information into the phrase construction process. However, a general syntactic analysis facility may be required, since many useful sources of phrases cannot be exploited if only a limited inventory of syntactic patterns can be recognized. Further research should be conducted into methods of incorporating automatic syntactic analysis into content analysis for document retrieval.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 3311) [ClassicSimilarity], result of:
      0.013336393 = score(doc=3311,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 3311, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=3311)
  0.33333334 = coord(1/3)

Abstract

Tools for automatic subject assignment help deal with scale and sustainability in creating and enriching metadata, establishing more connections across and between resources and enhancing consistency. Although some software vendors and experimental researchers claim the tools can replace manual subject indexing, hard scientific evidence of their performance in operating information environments is scarce. A major reason for this is that research is usually conducted in laboratory conditions, excluding the complexities of real-life systems and situations. The article reviews and discusses issues with existing evaluation approaches such as problems of aboutness and relevance assessments, implying the need to use more than a single "gold standard" method when evaluating indexing and retrieval, and proposes a comprehensive evaluation framework. The framework is informed by a systematic review of the literature on evaluation approaches: evaluating indexing quality directly through assessment by an evaluator or through comparison with a gold standard, evaluating the quality of computer-assisted indexing directly in the context of an indexing workflow, and evaluating indexing quality indirectly through analyzing retrieval performance.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 3667) [ClassicSimilarity], result of:
      0.013336393 = score(doc=3667,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 3667, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=3667)
  0.33333334 = coord(1/3)

Abstract

Descriptive metadata play a key role in finding relevant search results in large amounts of unstructured data. However, current scientific audiovisual media are provided with little metadata, which makes them hard to find, let alone individual sequences. In this paper, the TIB / AV-Portal is presented as a use case where methods concerning the automatic generation of metadata, a semantic search and cross-lingual retrieval (German/English) have already been applied. These methods result in a better discoverability of the scientific audiovisual media hosted in the portal. Text, speech, and image content of the video are automatically indexed by specialised GND (Gemeinsame Normdatei) subject headings. A semantic search is established based on properties of the GND ontology. The cross-lingual retrieval uses English 'translations' that were derived by an ontology mapping (DBpedia i. a.). Further ways of increasing the discoverability and reuse of the metadata are publishing them as Linked Open Data and interlinking them with other data sets.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 3810) [ClassicSimilarity], result of:
      0.013336393 = score(doc=3810,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 3810, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=3810)
  0.33333334 = coord(1/3)

Abstract

Nowadays, online data shows an astonishing increase and the issue of semantic indexing remains an open question. Ontologies and knowledge bases have been widely used to optimize performance. However, researchers are placing increased emphasis on internal relations of ontologies but neglect latent semantic relations between ontologies and documents. They generally annotate instances mentioned in documents, which are related to concepts in ontologies. In this paper, we propose an Ontology-based Latent Semantic Indexing approach utilizing Long Short-Term Memory networks (LSTM-OLSI). We utilize an importance-aware topic model to extract document-level semantic features and leverage ontologies to extract word-level contextual features. Then we encode the above two levels of features and match their embedding vectors utilizing LSTM networks. Finally, the experimental results reveal that LSTM-OLSI outperforms existing techniques and demonstrates deep comprehension of instances and articles.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 4005) [ClassicSimilarity], result of:
      0.013336393 = score(doc=4005,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 4005, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=4005)
  0.33333334 = coord(1/3)

Abstract

Subject indexing is an intellectually intensive process that has many inherent uncertainties. Existing manual subject indexing systems generally produce binary outcomes for whether or not to assign an indexing term. This does not sufficiently reflect the extent to which the indexing terms are associated with the documents. On the other hand, the idea of probabilistic or weighted indexing was proposed a long time ago and has seen success in capturing uncertainties in the automatic indexing process. One hurdle to overcome in implementing weighted indexing in manual subject indexing systems is the practical burden that could be added to the already intensive indexing process. This study proposes a method to infer automatically the associations between subject terms and documents through text mining. By uncovering the connections between MeSH descriptors and document text, we are able to derive the weights of MeSH descriptors manually assigned to documents. Our initial results suggest that the inference method is feasible and promising. The study has practical implications for improving subject indexing practice and providing better support for information retrieval.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 5268) [ClassicSimilarity], result of:
      0.013336393 = score(doc=5268,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 5268, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=5268)
  0.33333334 = coord(1/3)

Abstract

Automatic subject indexing addresses problems of scale and sustainability and can be at the same time used to enrich existing metadata records, establish more connections across and between resources from various metadata and resource collec-tions, and enhance consistency of the metadata. In this work, au-tomatic subject indexing focuses on assigning index terms or classes from established knowledge organization systems (KOSs) for subject indexing like thesauri, subject headings systems and classification systems. The following major approaches are dis-cussed, in terms of their similarities and differences, advantages and disadvantages for automatic assigned indexing from KOSs: "text categorization," "document clustering," and "document classification." Text categorization is perhaps the most wide-spread, machine-learning approach with what seems generally good reported performance. Document clustering automatically both creates groups of related documents and extracts names of subjects depicting the group at hand. Document classification re-uses the intellectual effort invested into creating a KOS for sub-ject indexing and even simple string-matching algorithms have been reported to achieve good results, because one concept can be described using a number of different terms, including equiv-alent, related, narrower and broader terms. Finally, applicability of automatic subject indexing to operative information systems and challenges of evaluation are outlined, suggesting the need for more research.

0.0044454644 = product of:
  0.013336393 = sum of:
    0.013336393 = weight(_text_:on in 1024) [ClassicSimilarity], result of:
      0.013336393 = score(doc=1024,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.121501654 = fieldWeight in 1024, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=1024)
  0.33333334 = coord(1/3)

Abstract

There are several ways to employ machine learning for automating subject indexing. One popular strategy is to utilize a supervised learning algorithm to train a model on a set of documents that have been manually indexed by subject matter using a standard vocabulary. The resulting model can then predict the subject of new and previously unseen documents by identifying patterns learned from the training data. To do this, the first step is to gather a large dataset of documents and manually assign each document a set of subject keywords/descriptors from a controlled vocabulary (e.g., from Agrovoc). Next, the dataset (obtained from Agris) can be divided into - i) a training dataset, and ii) a test dataset. The training dataset is used to train the model, while the test dataset is used to evaluate the model's performance. Machine learning can be a powerful tool for automating the process of subject indexing. This research is an attempt to apply Annif (http://annif. org/), an open-source AI/ML framework, to autogenerate subject keywords/descriptors for documentary resources in the domain of agriculture. The training dataset is obtained from Agris, which applies the Agrovoc thesaurus as a vocabulary tool (https://www.fao.org/agris/download).

0.0035563714 = product of:
  0.010669114 = sum of:
    0.010669114 = weight(_text_:on in 3650) [ClassicSimilarity], result of:
      0.010669114 = score(doc=3650,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.097201325 = fieldWeight in 3650, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.03125 = fieldNorm(doc=3650)
  0.33333334 = coord(1/3)

Abstract

The attempt to computerize a process, such as indexing, abstracting, classifying, or retrieving information, begins with an analysis of the process into its intellectual and nonintellectual components. That part of the process which is amenable to computerization is mechanical or algorithmic. What is not is intellectual or creative and requires human intervention. Gerard Salton has been an innovator, experimenter, and promoter in the area of mechanized information systems since the early 1960s. He has been particularly ingenious at analyzing the process of information retrieval into its algorithmic components. He received a doctorate in applied mathematics from Harvard University before moving to the computer science department at Cornell, where he developed a prototype automatic retrieval system called SMART. Working with this system he and his students contributed for over a decade to our theoretical understanding of the retrieval process. On a more practical level, they have contributed design criteria for operating retrieval systems. The following selection presents one of the early descriptions of the SMART system; it is valuable as it shows the direction automatic retrieval methods were to take beyond simple word-matching techniques. These include various word normalization techniques to improve recall, for instance, the separation of words into stems and affixes; the correlation and clustering, using statistical association measures, of related terms; and the identification, using a concept thesaurus, of synonymous, broader, narrower, and sibling terms. They include, as weIl, techniques, both linguistic and statistical, to deal with the thorny problem of how to automatically extract from texts index terms that consist of more than one word. They include weighting techniques and various documentrequest matching algorithms. Significant among the latter are those which produce a retrieval output of citations ranked in relevante order. During the 1970s, Salton and his students went an to further refine these various techniques, particularly the weighting and statistical association measures. Many of their early innovations seem commonplace today. Some of their later techniques are still ahead of their time and await technological developments for implementation. The particular focus of the selection that follows is an the evaluation of a particular component of the SMART system, a multilingual thesaurus. By mapping English language expressions and their German equivalents to a common concept number, the thesaurus permitted the automatic processing of German language documents against English language queries and vice versa. The results of the evaluation, as it turned out, were somewhat inconclusive. However, this SMART experiment suggested in a bold and optimistic way how one might proceed to answer such complex questions as What is meant by retrieval language compatability? How it is to be achieved, and how evaluated?