Search (8 results, page 1 of 1)

0.0039819763 = product of:
  0.011945928 = sum of:
    0.011945928 = weight(_text_:a in 3390) [ClassicSimilarity], result of:
      0.011945928 = score(doc=3390,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.22931081 = fieldWeight in 3390, product of:
          4.2426405 = tf(freq=18.0), with freq of:
            18.0 = termFreq=18.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=3390)
  0.33333334 = coord(1/3)

Abstract

The automated categorisation (or classification) of texts into topical categories has a long history, dating back at least to 1960. Until the late '80s, the dominant approach to the problem involved knowledge-engineering automatic categorisers, i.e. manually building a set of rules encoding expert knowledge an how to classify documents. In the '90s, with the booming production and availability of on-line documents, automated text categorisation has witnessed an increased and renewed interest. A newer paradigm based an machine learning has superseded the previous approach. Within this paradigm, a general inductive process automatically builds a classifier by "learning", from a set of previously classified documents, the characteristics of one or more categories; the advantages are a very good effectiveness, a considerable savings in terms of expert manpower, and domain independence. In this tutorial we look at the main approaches that have been taken towards automatic text categorisation within the general machine learning paradigm. Issues of document indexing, classifier construction, and classifier evaluation, will be touched upon.

0.003793148 = product of:
  0.011379444 = sum of:
    0.011379444 = weight(_text_:a in 5003) [ClassicSimilarity], result of:
      0.011379444 = score(doc=5003,freq=12.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.21843673 = fieldWeight in 5003, product of:
          3.4641016 = tf(freq=12.0), with freq of:
            12.0 = termFreq=12.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0546875 = fieldNorm(doc=5003)
  0.33333334 = coord(1/3)

Abstract

Automatic text classification (ATC) is a discipline at the crossroads of information retrieval (IR), machine learning (ML), and computational linguistics (CL), and consists in the realization of text classifiers, i.e. software systems capable of assigning texts to one or more categories, or classes, from a predefined set. Applications range from the automated indexing of scientific articles, to e-mail routing, spam filtering, authorship attribution, and automated survey coding. This article will focus on the ML approach to ATC, whereby a software system (called the learner) automatically builds a classifier for the categories of interest by generalizing from a "training" set of pre-classified texts.

Type

a

0.0035117732 = product of:
  0.010535319 = sum of:
    0.010535319 = weight(_text_:a in 3389) [ClassicSimilarity], result of:
      0.010535319 = score(doc=3389,freq=14.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.20223314 = fieldWeight in 3389, product of:
          3.7416575 = tf(freq=14.0), with freq of:
            14.0 = termFreq=14.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=3389)
  0.33333334 = coord(1/3)

Abstract

The automated categorization (or classification) of texts into predefined categories has witnessed a booming interest in the last 10 years, due to the increased availability of documents in digital form and the ensuing need to organize them. In the research community the dominant approach to this problem is based an machine learning techniques: a general inductive process automatically builds a classifier by learning, from a set of preclassified documents, the characteristics of the categories. The advantages of this approach over the knowledge engineering approach (consisting in the manual definition of a classifier by domain experts) are a very good effectiveness, considerable savings in terms of expert labor power, and straightforward portability to different domains. This survey discusses the main approaches to text categorization that fall within the machine learning paradigm. We will discuss in detail issues pertaining to three different problems, namely, document representation, classifier construction, and classifier evaluation.

Type

a

0.0033183135 = product of:
  0.0099549405 = sum of:
    0.0099549405 = weight(_text_:a in 5172) [ClassicSimilarity], result of:
      0.0099549405 = score(doc=5172,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.19109234 = fieldWeight in 5172, product of:
          4.2426405 = tf(freq=18.0), with freq of:
            18.0 = termFreq=18.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0390625 = fieldNorm(doc=5172)
  0.33333334 = coord(1/3)

Abstract

In this issue Giorgetti, and Sebastiani suggest that answers to open ended questions in survey instruments can be coded automatically by creating classifiers which learn from training sets of manually coded answers. The manual effort required is only that of classifying a representative set of documents, not creating a dictionary of words that trigger an assignment. They use a naive Bayesian probabilistic learner from Mc Callum's RAINBOW package and the multi-class support vector machine learner from Hsu and Lin's BSVM package, both examples of text categorization techniques. Data from the 1996 General Social Survey by the U.S. National Opinion Research Center provided a set of answers to three questions (previously tested by Viechnicki using a dictionary approach), their associated manually assigned category codes, and a complete set of predefined category codes. The learners were run on three random disjoint subsets of the answer sets to create the classifiers and a remaining set was used as a test set. The dictionary approach is out preformed by 18% for RAINBOW and by 17% for BSVM, while the standard deviation of the results is reduced by 28% and 34% respectively over the dictionary approach.

Type

a

0.0029264777 = product of:
  0.008779433 = sum of:
    0.008779433 = weight(_text_:a in 1140) [ClassicSimilarity], result of:
      0.008779433 = score(doc=1140,freq=14.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.1685276 = fieldWeight in 1140, product of:
          3.7416575 = tf(freq=14.0), with freq of:
            14.0 = termFreq=14.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0390625 = fieldNorm(doc=1140)
  0.33333334 = coord(1/3)

Abstract

The logical approach to information retrieval has recently been the object of active research. It is our contention that researchers have put a lot of effort in trying to address some difficult problems of IR within this framework, but little effort in checking that the resulting models satisfy those well-formedness criteria that, in the field of mathematical logic, are considered essential and conductive to effective modelling of a real-world phenomenon. The main motivation of this paper is not to propose a new logical model of IR, but to discuss some central issues in the application of logic to IR. The first issue we touch upon is the logical relationship we might want to enforce between formulae d, representing a document and n, representing an information need; we analyse the different implications of models based on truth, validity or logical consequentiality. The relationship between this issue and the issue of partiality vs. totality of information is subsequently analysed, in the context of a broader discussion of the role of denotational semantics in IR modelling. Finally, the relationship between the paradoxes of material implication and the (in)adequacy of classical logic for IR modelling purposes is discusses

Series

Trends in ... a critical review

Type

a

0.0027093915 = product of:
  0.008128175 = sum of:
    0.008128175 = weight(_text_:a in 4101) [ClassicSimilarity], result of:
      0.008128175 = score(doc=4101,freq=12.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.15602624 = fieldWeight in 4101, product of:
          3.4641016 = tf(freq=12.0), with freq of:
            12.0 = termFreq=12.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0390625 = fieldNorm(doc=4101)
  0.33333334 = coord(1/3)

Abstract

Hierarchical text classification (HTC) approaches have recently attracted a lot of interest on the part of researchers in human language technology and machine learning, since they have been shown to bring about equal, if not better, classification accuracy with respect to their "flat" counterparts while allowing exponential time savings at both learning and classification time. A typical component of HTC methods is a "local" policy for selecting negative examples: Given a category c, its negative training examples are by default identified with the training examples that are negative for c and positive for the categories which are siblings of c in the hierarchy. However, this policy has always been taken for granted and never been subjected to careful scrutiny since first proposed 15 years ago. This article proposes a thorough experimental comparison between this policy and three other policies for the selection of negative examples in HTC contexts, one of which (BEST LOCAL (k)) is being proposed for the first time in this article. We compare these policies on the hierarchical versions of three supervised learning algorithms (boosting, support vector machines, and naïve Bayes) by performing experiments on two standard TC datasets, REUTERS-21578 and RCV1-V2.

Type

a

0.002654651 = product of:
  0.007963953 = sum of:
    0.007963953 = weight(_text_:a in 846) [ClassicSimilarity], result of:
      0.007963953 = score(doc=846,freq=8.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.15287387 = fieldWeight in 846, product of:
          2.828427 = tf(freq=8.0), with freq of:
            8.0 = termFreq=8.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=846)
  0.33333334 = coord(1/3)

Abstract

It is well known that, within the Latin production of written text, peculiar metric schemes were followed not only in poetic compositions, but also in many prose works. Such metric patterns were based on so-called syllabic quantity, that is, on the length of the involved syllables, and there is substantial evidence suggesting that certain authors had a preference for certain metric patterns over others. In this research we investigate the possibility to employ syllabic quantity as a base for deriving rhythmic features for the task of computational authorship attribution of Latin prose texts. We test the impact of these features on the authorship attribution task when combined with other topic-agnostic features. Our experiments, carried out on three different datasets using support vector machines (SVMs) show that rhythmic features based on syllabic quantity are beneficial in discriminating among Latin prose authors.

Type

a

0.002473325 = product of:
  0.0074199745 = sum of:
    0.0074199745 = weight(_text_:a in 3456) [ClassicSimilarity], result of:
      0.0074199745 = score(doc=3456,freq=10.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.14243183 = fieldWeight in 3456, product of:
          3.1622777 = tf(freq=10.0), with freq of:
            10.0 = termFreq=10.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0390625 = fieldNorm(doc=3456)
  0.33333334 = coord(1/3)

Abstract

The existence, public availability, and widespread acceptance of a standard benchmark for a given information retrieval (IR) task are beneficial to research an this task, because they allow different researchers to experimentally compare their own systems by comparing the results they have obtained an this benchmark. The Reuters-21578 test collection, together with its earlier variants, has been such a standard benchmark for the text categorization (TC) task throughout the last 10 years. However, the benefits that this has brought about have somehow been limited by the fact that different researchers have "carved" different subsets out of this collection and tested their systems an one of these subsets only; systems that have been tested an different Reuters-21578 subsets are thus not readily comparable. In this article, we present a systematic, comparative experimental study of the three subsets of Reuters-21578 that have been most popular among TC researchers. The results we obtain allow us to determine the relative hardness of these subsets, thus establishing an indirect means for comparing TC systems that have, or will be, tested an these different subsets.

Type

a