IMPLEMENTATION OF THE APPROACH TO ONTOLOGICAL MODELING OF THE MUSICAL SUPERGENRE “METAL”

Background. The relevance of the article is due to the development of modern ontological methods of structuring information and the need to systematize data in many new specific subject areas. Such subject areas include the musical art of the "metal" variety, which is quite common today, but insufficiently studied within the terminology. The subject of the article are ontological models and tools for creating ontological knowledge bases. Objective. The purpose of the paper is to increase the correctness of the semantic search in the knowledge base of the musical supergenre "metal". The scientific problem is the need to improve the terminology in this subject area and build an ontological knowledge model that increases the accuracy of information retrieval for the target audience, compared to the existing relational model implemented on one of the known web resources. Methods. Classification method, generalization method, software optimization methods, analytical method. The way to solve the problem: selection based on the comparative characteristics of the best web resource of the subject area and identifying the shortcomings of its model of knowledge representation, designing an ontological knowledge model and testing its effectiveness. Results. The average SUM for all users is 83.85%, which is a good indicator for ontological knowledge bases. At the same time, a similar method of checking the database of the supergenre "metal" on the basis of the site "Encyclopedia Metallum", which used the classical relational model of database organization, showed much lower results. Thus, the average SUM for 10 users was 75.32%, respectively. Conclusions. The scientific novelty of the obtained results is as follows: For the first time an ontological model (ontology) of the subject area was created: musical supergenre "metal", which showed much higher efficiency of semantic search than the best relational model of this subject area, implemented as a web resource. The developed structure can be used to create ontologies of related musical supergenres with similar terminology. Future research also plans to integrate this ontological knowledge model with applied web-based and desktop applications.


Introduction
Today, the substantial humanity`s challenge is the significant growth of daily information. Therefore, experts are no longer faced with the question of separating important information from the supposedly secondary one, but the problem of formalization, conceptualization and classification of individual subject areas. The task is complicated by the emergence of new, specific areas, especially those related to science and culture. Ontological knowledge bases are ideal for the development of such a high-quality terminology system, as well as its modelling and implementation in the modern information environment. After all, they are more effective than traditional models of knowledge representation on the criteria of quality of graphic representation, data structuring and the possibility of advanced information retrieval by means of query creation. This approach is especially justified for the latest insufficiently researched subject areas, knowledge of which should be properly systematized. On the example of such a specific newest subject area, namely the subject area of the musical supergenre "metal", it was decided to conduct a study of the method of ontological design to obtain specific knowledge.

A.
Establishing the relevance of the study The relevance of the study can be proved from two sides: from the growing interest of the scientific community to ontological knowledge bases and the study of their capabilities in general and from the chosen subject area of the musical supergenre "metal", which brought together many fans around the world and became a kind of cultural phenomenon of the XX and XXI centuries.

Review of analogues in terms of ontological knowledge bases
It is extremely important to review the latest literature on the creation of ontologies.
A significant problem in the development of ontologies is the presentation of knowledge graphs in a user-friendly form with the ability to integrate them into the interface of a particular API. It is proposed to solve this challenge by means of setting a number of restrictions and specifications, as well as by high-level SPARQL-queries. These queries are proposed to be integrated with modern web-oriented programming languages such as JavaScript [1].
Protégé, as one of the most powerful tools for developing ontologies, has its own limitations. Such limitations often give rise to a debate as to whether Protégé can meet the needs of a wide variety of developers for different research or design purposes. Therefore, one of the scientific studies compared the work with this tool on a standard tutorial, in a research laboratory and during development for practical purposes [2]. It was found that despite the slight differences in patterns and approaches to work, it is possible to develop a general set of recommendations that avoids significant problems in the end use.
A controversial issue in the development of ontologies of different levels is the classification of entities in domain ontologies under the upper classes of ontology [3]. Integrating top ontologies in this way is difficult and, despite new automated methods, remains largely a manual task. During the mass collective expertise conducted by the authors of this study, an important conclusion was made about the improvement of the methodological base around the manual integration of domain and top ontologies.
The advantages and disadvantages of all the above ontological knowledge bases are shown in Table 1. The sample of experiments is too small, which can lead to doubts about the validity of the conclusions.

III
The possibility of creating a universal terminology system for related subject areas by certain features is shown.
No attention is paid to subject areas, which are difficult to allocate a place in the modern classification.

C. Review of analogues in terms of knowledge bases of the supergenre "metal"
The goal of the research is to increase the correctness of the semantic search in the knowledge base of the musical supergenre «metal». Relevance of the research caused sufficient prevalence of kind of music named «metal», that has both many genres and a basis for philosophical currents of modernity that found many fans throughout the world. Nowadays, we can identify several well-known resources, covering not only the events or news of the metal underground, but also describing the existing rigid classification of certain objects in this subject area. This is an important indicator for the metal community as a counterculture that has existed for half a century and keeps increasingly expanding its horizons through the rapid development of new, more modern subgenres and integration with many other philosophical currents and cultural phenomena.
Encyclopaedia Metallum [4] -this resource systematizes different genres of metal, groups, artists, albums and their varieties, record labels. The site has detailed information even about beginner bands, it is also possible to track the rating of the band by written reviews. Available information about band members. The site has information retrieval functions based on relational cross and multi-parameter queries.
DARK ALBUM [5] offers the ability to search for interesting information on the parameters of the genre, country and year of foundation, also on the site there are reviews and lyrics of famous tracks, subculture news.
Altwall [6] is a resource focused on news and concerts, but it also has a metal genealogy tree. The site positions itself as a researcher of contemporary music, so there are other, smaller, supergenres.
Analysing the above software products, we can identify a number of disadvantages and advantages for each of them. The comparative characteristics are given in Table 2.

2.
Implementation of the musical supergenre "metal" ontological model

A. Initialization of the Terminological
System of the Supergenre "Metal" The subject area of the musical supergenre "metal" is very voluminous both in terms of the dynamics of changes in information and in terms of the general structure of basic concepts. The key basic concepts of the subject area from which to start are the concepts of "genre" and "band". Further, the terminology expands and takes into account a large number of concepts, the main of which are: "country of origin", "status", "year of foundation", "years of activity", "lyrical themes", "label" and others.
After identifying key concepts, it is advisable to build a hierarchy of the subject area, at the root of which will be the terminology, and at the nodes -the nomenclature.
The semantic network of frame signs is used as a model of nomenclature knowledge representation. The S semantic network is constructed as a combination of frame characters Фі: S=⋃� Фі.
It can be assumed that in general a set of terms. This means that the power of the set Ф={Фі} is greater than the power of the set Term. A graphical representation of a fragment of the semantic network S for constructing an ontology of metal music can be seen in Fig. 1. It should be noted that the created graph is basic, so it can be modified and expanded if necessary, and the main purpose of its creation is the need for graphical implementation of the terminology.

B. Ontology Coding
Ontology coding involves first of all the definition of tools, so the development environment for creating a powerful knowledge base. To do this, it is necessary to compare several modern tools for designing ontological knowledge bases and determine the most effective for this development. The comparative characteristics are given in Table 3.

Ontolingua
The presence of a large archive of ontologies, which allows them to be corrected and created using existing ones, the ability to import and graphical modelling.

Protégé
Convenience of converting the hierarchy of the subject area into program code, high quality visualization, the presence of a number of plugins, constant improvement and free access.

OntoEdit
The convenience of use, development of an ontology under the guidance of a methodology and through an inference process, development of axioms, extensible structure through plugins.
Closed source code.
Thus, Protégé is the most convenient and powerful tool for designing the ontology of the musical supergenre "metal" due to open source code and high-quality visualization.
The following are the main stages of ontology coding.

1.
Selection of classes and subclasses, establishing a clear hierarchy between them.

3.
Set data types or create new ones as needed.

4.
Establishing the possibility or impossibility of crossing certain classes.

5.
Constructing expressions and organizing information.

6.
Granting properties to individuals.

7.
Implementation of a quick search for information by means of SPARQL.
An important result of the work is the basic graph of the ontology built automatically by the environment, which will be further expanded.
This graph is shown in Fig. 2. This metric is determined by the formula in the context of ontology development, the Completion indicator is a Boolean value and indicates the achievement of the goal of obtaining information by the user, Satisfaction shows the level of satisfaction with structuring information in this ontology, Errors shows the occurrence of certain inconsistencies or other errors in obtaining information and Times shows the average number of requests.
To determine the correctness of the search for information in the implemented ontological knowledge base, it was provided to 10 users, each of whom implemented 10 queries on it.
The average SUM metric for all users will be 83.85%, which is quite a good indicator for ontological knowledge bases. At the same time, a similar method of testing the database of the supergenre "metal" of the site "Encyclopaedia Metallum", which used the classical relational model of database organization, showed much lower results. Thus, the average SUM metric for 10 users was 75.32%, respectively. Fig. 3 illustrates the graph of the average scores of ten experts on SUM metrics according to two models of the knowledge base of the music supergenre "metal" -a relational base, implemented on the most famous web resource in this subject area called "Encyclopaedia Metallum", and ontological model created in this research. The graph clearly shows that despite some differences due to individual experiments of experts, the developed ontological model is ahead of relational in search efficiency, and in some cases the user found the necessary information with 100% efficiency. The study proposes a new ontological model of knowledge base (ontology) of the subject area: musical supergenre "metal".

2.
By means of the Protégé, the ontology of the subject area of the musical supergenre "metal" is realized in the form of a knowledge base.

3.
Testing of the implemented knowledge base and evaluation of the correctness of the semantic search according to the SUM metric was conducted.

4.
The purpose of the study -to increase the correctness of the semantic search of information is achieved through the use in the implementation of the knowledge base of a new ontological model (ontology) of the musical supergenre «metal». In comparison with the classical, relational model of the organization in databases application of new ontology allows to raise correctness of search by 8%.

5.
It is revealed that by means of Protege, it is possible to design various bases of knowledge even of informal subject areas. However, it allows you to create your own data types, create new relationships, make a quality formalization of information -in other words, there are ample opportunities to expand functionality.

6.
The scientific problem is the need to improve the terminology in this subject area and build an ontological knowledge model that increases the accuracy of information retrieval for the target audience, compared to the existing relational model implemented on one of the known web resources.

7.
The following methods were used: classification method, generalization method, software optimization methods, analytical method.

8.
The way to solve the problem: selection based on the comparative characteristics of the best web resource of the subject area and identifying the shortcomings of its model of knowledge representation, designing an ontological knowledge model and testing its effectiveness.

9.
The developed structure can be used when creating ontologies of related musical supergenres with similar terminology.
10. Future research also plans to integrate this ontological knowledge model with applied webbased and desktop applications.