-This paper presents an ontological model of a the the text document as an electronic archive resource. The article also presents an ontology-based algorithm for the classification technical documents.
INTRODUCTION
The task of categorizing text documents to simplify the search for information in an electronic archive of an organization is more relevant than ever. In most cases, archiving is structured
manually by archive specialists.
Specialists should have knowledge in the subject area and take into account the specifics of the stored documentation.
Automation of categorization of the archive of electronic text documents should be carried out taking into account the semantics of information in the documents. Otherwise, the experience of highly qualified specialists developing this documentation will be difficult to extract from unstructured resources for further use.
Currently, researchers offer various ways to solve this problem. In [1], the ant colony classification algorithm is used to classify data and is used to quickly search for large amounts of data from intelligent archives.
Characteristics of a text document are taken into account during its analysis and processing and are included in the document model.
The extended Boolean model of the document does not represent terms with values of 0, 1, but with weighting coefficients using the theory of fuzzy sets [2-5]. In this case, the value of the weight coefficient is determined from the interval [0, 1], thus we obtain that ∈ [0, 1] .
The vector model formally presents text documents as a matrix of terms and documents [6]:
= | |×| |, where = { 1, . . . , ,. . . , }; = { 1,. . . , ,. . . , }, is a vector in the -dimensional space .
In [7] [8], the authors present an ontology designed to
model archival descriptions
of collections of historical
documents. In [9], the authors present aspects of the current
activities of the digital library related to the Semantic Web
and present their functionality. They show examples ranging
from general architectural descriptions to the detailed use of
specific ontologies. In [
One of the solutions to this problem is the use of intelligent algorithms for the analysis of text documents with the division of the archive into classes in accordance with the semantics of the subject area. The semantics of the subject area
will be concluded in the subject ontology, formed through the analysis of textual documentation.
Domestic and foreign researchers (Gavrilova T.A. [
II.
MODELS OF APPLIED ONTOLOGY OF TEXT DOCUMENTS
The construction of an ontology in the classification of
documents in electronic archives is necessary to take into
account the characteristics of the subject area and to increase
the speed of document search. Ontology defines a semantic
scale that defines whether a document belongs to one class
[
Thus, the formal model of applied ontology of the electronic archive of project documentation is: = ⟨ , , , ⟩, where is the set of terms of design documentation for an electronic archive;
is a set of terms of a problem area;
is a set of ontology relationships. Many relationships include the following: = { , , }, where is the hierarchy relation; whole relationship; is an association relation.
is a part-to
Formally the set of terms of design documentation for an electronic archive is: = ( 1 ∪ 2 ∪ … ∪ ) ∪ is a set of terms of the problem area extracted from the documents of the electronic archive.
Formally, the functions of interpretation of the subject where , = 1, area; ontology are: where : { = { ,
}, } → { } is an interpretation function that defines the correspondence between the terms of the problem area and the terms of the design documentation of the electronic archive; : { } → { } is an interpretation function that defines the correspondence between the terms of the problem domain extracted from electronic archive documents and the terms of the problem domain.
The main in the ontology of the electronic archive is the relation "associate_with".
subject area to which the project document of the electronic archive belongs and determines the subject of the document. Copyright © 2020 for this paper by its authors.
The characteristic of the weight of the term in a text document is the frequency of the i-th term in the document. Hence, the following patterns are relevant: . document d; n is a measure of the power of the text input of high-frequency terms in a document are system-wide; terms with a low frequency in a particular document do not provide an improvement in the quality of search for documents in the archive.
The most indicative are terms that have an average frequency of occurrence in a document, but most fully characterize a document in a problem area [16, 17].
If the frequency of occurrence of one term is significantly higher in the document than the frequency of its occurrence in all analyzed documents of the electronic archive, then this term is semantically significant. Formally, this rule is = ⋅ ( ) , where is an indicator of the semantic significance of the term in this document; is the total number of all documents in the electronic archive; is the value of the index of the normalized frequency of the term ; the total number of documents containing the term . ( ) is
Thus, the ontological model of an electronic archive document is: = ⟨ , ⟩, where
, is the set of terms of the problem area of the j-th document of the electronic archive.
_
ℎ( , ) = 1.
This equality assumes that the document is mapped into the space of terms , 2, … , .. If the document , then the set of terms of the document can IV. is the -th term of be represented as follows:
= { 1 , 2 , … , }, where n is the total number of terms in the document d.
III.
THE ONTOLOGICAL INDEX MODEL
The ontological indexing algorithm for text documents of the electronic archive is shown in Figure 1.
The degree of semantic significance of an electronic ontology term is the value of coincidence of the term context environment with the set of terms of the electronic archive document. Contextual environment is composed of terms that are semantically close to the analyzed concept of a problem area [19].
Hence the formally semantic index of the i-th document is: {( 1 , 1), ( 2 , 2), … , ( , ), … , ( , )}, where l is the total number of terms in the i-th document of the electronic archive after text preprocessing.
The degree of expression of the concept in the i-th document d will be calculated as follows: μ( ) = 1 − 1 ∑| − |, where , are indicators of the frequency of the term in the description of the k-th term of the ontology in the , ℎ = [0.7 … 1.0], = [0.5 … 0.7), = [0 … 0.5), where is the n-th class of documents (classification basis); is the k-th concept of ontology; m is the linguistic label.
At the second step, the degree of belonging of the document d to each class is calculated using the following expression: =1
0 ( ) = − ∑(1 − ), where k is the number of parameters of the class ; is a sign of a document matching the d -th property of the class , which is calculated using the following expression: = { 1, ∈ , μ( ) ∈ .
Thus, the document d corresponds to the characteristic if it contains concepts characterizing the given attribute and its degree of expression is included in the interval of the linguistic label. conducted to assess the quality of the classification of documents in the electronic archive of the Federal Scientific
Mars. JSC
Mars is an organization engaged in the design, development and maintenance of automated systems, software and hardware for the Russian
For the experiments the following sets of documents Navy. were selected: technical specifications; patent research reports; specifications; testing programs and techniques; programmer, manuals.
user, system administrator, 1037 design documents were selected for experiments. Figure 2 shows the signs of expert dividing documents into classes according to certain criteria. etc. the term. of: 1) 2) 3) 4) conducted to select the preferred method for defining the
The test set consisted of a public data set containing about 10,000 tweets. Each tweet contains a text describing either a disaster or some other information. Each tweet has a label that determines whether the tweet belongs to the “ disaster” or “other” class.
The first step is data preprocessing. Pretreatment consists
Tokenization according to words - splitting the analyzed text into separate words.
At the second step, many indices of the analyzed poorly structured information resources were erased, which were based on the following
models of information resource representation: a word bag (Fig. 3), a statistical model based on
TF-IDF (Fig. 4), and a linguistic model based on
The following assessments of the classification quality were obtained: “Bag of words” - 62.23%, “TF-IDF” 74.78%, “Word2Vec” - 81.7%.
Thus, the linguistic model will be the best method for defining the terms of a semi-structured information resource. However, it is recommended to use a statistical model if low computational complexity of the algorithm and high speed of operation are required.
An ontological set of document indixes and classic indexes, which include the term-frequency values, were built. The classification quality assessment model from [20] was used as an evaluation function. The results of the experiments are presented in figures 6 and 7.
As can be seen from the results of the experiments, the classification of ontological representations is faster (up to 27 times) relative to the classification time of classical indices.
The quality of classification of ontological representations in comparison with the results of classification of classical indices is slightly worse only when divided by the class of documentation. When dividing a multitude of documents by type of documentation, section of documentation and subject of work, the quality of classification of ontological representations is higher than that of classical indices.
Thus, an ontological model of a text document as an electronic archive resource and an ontologically oriented algorithm for the classification of technical documents is proposed. As can be seen from the results of the experiments, the formation of the ontological presentation of each individual document in the archive can significantly increase the speed of automatic classification of documents (up to 27 times) while maintaining or slightly improving the quality of classification.
In future works, it is planned to introduce fuzzy elements in the ontological representation of project documents.
This paper has been approved within the framework of the federal target project “R&D for Priority Areas of the Russian Science-and-Technology Complex Development for 2014-2020”, government contract No 05.604.21.0252 on the subject “The development and research of models, methods and algorithms for classifying large semistructured data based on hybridization of semantic-ontological analysis and machine learning”. [1] [2] [3] [4] [5] [6] [7] [8] [9]
Addison-Wesley