RU2107942C1 - Method for detection of object position in storage repository using search subject criterion - Google Patents

Abstract

FIELD: computer engineering, in particular, information retrieval, databases. SUBSTANCE: method involves building a limited list of keywords for each object, setting object address, filling corresponding fields in information carrier with elements of this list using object address, information is read by corresponding reading unit. When subject search is performed, method involves generation of search subject query, filling information carrier, which is analogous to information carrier with field of object addresses, with this query represented as subject search instruction, reading search subject instruction from information carrier, comparison of words of search subject instruction with field titles which are filled in field of information carriers which are read by corresponding reading unit, and detection of object address by means of these comparisons. Preliminary Ontology is developed and universal information retrieval language is designed. Lexical items of this language are categories of Ontology. Each meaning of each category in Ontology is assigned universal unique code, array of descriptors of universal information retrieval language is generated. Information words list about each object is enriched with maximal list of information words which describe this object according to which corresponding region of Ontology is assigned to given object. Keywords which are detected for this object and implied by corresponding region of Ontology are encoded by means of arrays of descriptors. Then subject-sensitive code description of object with address as its search subject description is entered on corresponding information carrier as whole record. When subject search instruction is entered, all keywords of subject search request are encoded by means of descriptors of universal information retrieval language. EFFECT: increased recall of subject search. 3 cl, 2 dwg

Description

 The invention is intended to find objects of location, storage, stay, birth, etc. for queries that do not contain exact names, under which objects appear in inventory books, file cabinets or catalogs and for which storage workers find these objects using search engines containing the addresses of objects.

 The need to find objects by search thematic features arises in all areas of human activity, in particular when searching for sources of information (documents) for queries with a specific topic (thematic queries) when planning and conducting research, experimental design work, when analyzing the results of these activities when writing applications for inventions in patenting, in the examination of applications for peaceful novelty, etc.

 The number of thematic queries all over the world annually amounts to tens of millions, by which searches are carried out for objects of all kinds: documents, goods, things, materials, fossil organisms, people, works of art, etc.

 Thematic queries take the form of individual words, groups of words or phrases (less often - texts from several sentences), which involve incomplete search thematic features of objects.

 Since it is possible to make thematic search descriptions (i.e., prepared for thematic searches) of descriptions for any objects, the patented method in this application will be described by the example of searches of subject collections of documents. However, as search thematic signs of objects, not only symbolic signs (letters, numbers, etc.) can be used, but also signs such as body weight, color, hardness, etc. (i.e., physicochemical and other properties) and by which sorting, production and other similar actions are performed, the effectiveness of which can be improved by using the patented method.

 Incomplete search thematic features of documents are inherent, as a rule, to groups of documents. Therefore, information retrieval systems designed to satisfy thematic queries usually issue groups of documents called subject collections of documents. Such systems can be called subject search systems.

 Thematic search systems include themed searchable descriptions of documents with storage addresses that make up the base of thematic searches - items and systematic card or computer catalogs of libraries, museums, archives, warehouses, exhibitions, etc. In empirical computer systems of thematic searches, databases of thematic searches are fields of keywords, descriptors, formats, etc. This is a subject (in the library sense) or unitary organization of these bases.

 Using the thematic search databases in the thematic search systems using incomplete search thematic features, the addresses (storage locations or locations) of documents in the repository (storage codes, record numbers on machine-readable media of document databases in thematic search systems, geographical coordinates, etc.) are established , by which the documents themselves are then found in document collections, documentary databases, etc. Therefore, some subject search systems may also have a separate document search system ntov for full featured search - bibliographic data. In the medium of the bibliographic description of the document, as well as in the thematic search description, the storage address of the document is indicated - its location in the document collection, document database, etc.

 Until recently, the theory of many foundations of subject search systems (philosophy, information and classification theory, etc.) was absent [1 - 3 (26)]. Here, in parentheses enclosed in squares, the total number of publications found on this topic is given using the Informatron thematic search system, which includes 2000 documents on the theory and practice of informatronics. Therefore, thematic search systems were developed empirically and, as a rule, these thematic search systems cannot give out what users want. This has led the information service industry on thematic crisis requests [4–6 (15)].

 An analysis of the results of thematic searches using empirical systems of thematic searches, conducted on the basis of search informology [7 - 9 (13)], which is the main patented method, showed that the reasons for the unsatisfactory operation of empirical systems of thematic searches are, in particular, the following.

 1. A unitary organization of thematic search databases, which allows you to use an average of one hundred lexicans in each database to represent thousands of search-related thematic descriptions and a significantly larger number of search-related thematic prescriptions, operating with millions of informative layers - names of objects, phenomena, laws, sciences and practices. The reality of these figures is confirmed, for example, by the fact that more than eleven million are known to chemical compounds alone.

 2. Low-vocabulary of information retrieval languages, which forces one to replace the natural informative words of documents and thematic queries with, as a rule, the lexicans of the information retrieval language of this base of thematic search, approximately corresponding to their meanings, which is a direct distortion of the meanings of documents and queries and results in automatic search the issuance of irrelevant, inaccurate, inappropriate documents request subject. Automatically obtained information has to be sorted and replenished for a long time by visual-manual method using dialogs, interactive procedures and paging (browsing), which significantly lengthens the search process and makes it more expensive.

 On average, each source of information is represented in the database of thematic searches by three lexicants, while a full search thematic description should, according to the uncertainty principle, include an average of one hundred written and implied informative lexicon words.

 Implied words are those words that logically (logonymously) follow from the written ones. For example, the article says the words “brain,” “heart,” and “liver”; the words "organs" and "anatomy" are not always written, but are necessarily implied.

 Thus, almost all documents presented in the above descriptions are incomplete and distorted.

 Evidence that empirical search-related topic descriptions includes an average of three words each is that for four or more word-specific search engine descriptions, none of the more than one hundred empirical search engine search engines automatically generated a thematic selection of documents.

 3. Low representativeness. Representativeness - the number of potential thematic search terms in each thematic search description.

The number of search thematic features of documents, which are individual words of the search thematic description and their potential combinations (samples), is calculated by the Cardano formula
X = 2 ⁿ -1,
Where
n is the number of words by which the document appears in its thematic search description.

With n equal to 100, X is approximately equal to 10 ³⁰ , and with n equal to 3, X is equal to 10 ¹ . If we accept that 100 words are all informative words of a document and all of them contain a document in its thematic search description, then this is 100% representativeness of the search for thematic features of the document in the thematic searches database, which provides the ability to search for a document by any of its 10 ³⁰ one -, two-, ... and a verbatim search-related thematic features of the document. These characteristics are potential samples of 100 words presented in the search thematic description. Samples are implemented upon receipt of the relevant search requirements. In other words, this verbatim thematic search thematic description of the document allows you to find a document for 10 ³⁰ different types (topics) of search prescriptions.

If we accept that 100% representativeness is 10 ³⁰ potential search thematic features of a complete full informational description, then 10 ^{1 of} these signs of three-word empirical writings is 10 ^-25 % representativeness.

 Such an insignificant indicator of the representativeness of empirical descriptions is one of the reasons for the unsatisfactory operation of unitermic systems of thematic searches and their networks, their unprofitability, etc.

 4. The use of layers of natural languages as search thematic features of documents, which in recent years has become almost the only way to visually solve the problem of relevance of subject collections of documents [10 (7)] automatically issued by empirical systems of thematic searches. However, this did not solve the problem with automatic searches.

 The polysemy of natural language layers during automatic searches still leads to low relevance indicators and, in many cases, to 100% of documents in thematic collections not matching to query topics, and word synonyms are the reason for the local (for this thematic searches system) incomplete thematic collections of documents. Local completeness is all documents available in a specific database and including in texts and implying informative words indicated in the subject of the request.

 When forming search prescriptions based on thematic requests, employees of empirical systems of thematic searches have to solve the problem of polysemy, homonymy and synonymy as many times as there are thematic requests, i.e. worldwide - millions of times annually. The solution to this problem in most cases is unsatisfactory, because most of those who formulate search prescriptions are not professionals in natural linguistics and / or in the subject matter of queries. Therefore, in order to obtain a sufficiently complete and relevant thematic selection of documents on a single query topic, dozens and more search requirements are generated and searches are performed in almost all databases and files of the thematic search system during visual control of search results.

 5. The use of phrases as lexicants. To imagine how large the distortions of documents and queries are in the databases of thematic searches, one should imagine, for example, the “Captain's Daughter” by A.S. Pushkin translated into English only with the help of a hundred phrases of the Russian-English phrasebook.

 A search in this way on one subject lasts an average of about two hours, although thematic collections of documents can be searched for in tens of seconds by a single search prescription of a computer.

 These and other reasons are the basis of the global crisis of thematic informatics [4 - 6 (15)], which causes damage to many billions of rubles annually in all spheres of human activity throughout Russia. It manifests itself in the fact that 80% of scientific publications, patent applications, design developments, and now databases do not contain new information and duplicate, as a rule, not the best of the known [11, 12 (15)]; many institutions and their units duplicate each other, causing confusion, about 50% of computers are inactive, and only 5% of the computer's computer time in the country is used in general. Due to the lack of until recently the theory of thematic search systems [1 - 3 (26)] and under the influence of imposing propaganda, accusing computer owners of incompetence and timeliness, computers are acquired more for prestige than for thematic searches; An exhaustive informational study of the programs planned, implemented and proposed for implementation often costs more than the cost of their implementation in materials, etc. The proposed method is aimed at solving all these and many other problems in the field of intellectual and creative activity and its practical application.

 It is a well-known way to search by bibliographic, requisite and other full queries - full search features of documents. This method has the following properties that do not allow searches by topic queries.

 Bibliographic search systems include complete descriptions of search bibliographic features of documents with storage addresses. These descriptions make up the bibliographic search database of the bibliographic search system. Each description refers to only one document and one search bibliographic item of a document, and only one specific document is issued per bibliographic request, if it is in the fund or database. Bibliographic search systems can issue specific documents, as a rule, only from complete bibliographic descriptions.

 Upon receipt of a bibliographic request using a search bibliographic database (alphabetical author-title card index or computer catalog), a storage cipher or record number in document databases is established. A document is found or displayed on the display screen using the set storage cipher or record number.

 The presence of the requested document in the bibliographic search system is determined using the search bibliographic database: if the searched document is in the bibliographic search system, then the search bibliographic database has a corresponding description and there is no search problem. The bibliographic search system operates on the principle of "Get what you requested." However, the ratio of the number of bibliographic and thematic queries is 1 to 1000, respectively.

 As mentioned above, bibliographic searches are a mandatory step in the work and thematic search systems. Bibliographic search systems are older than thematic search systems. Therefore, according to the law of conservation of properties (phenomena, laws, morphology) of elements and parts (older forms of matter) in the objects formed by them (lower forms of matter) and the law of one-directionality of universality [7 - 9 (13)] searches for thematic queries using thematic search systems should also provide a bibliographic search, and a thematic search should be characterized by the same certainty as searches using bibliographic search systems. The proposed method to a large extent provides this.

 Closest to the claimed one is a method of establishing the address of an object by a search thematic feature, which consists in the fact that they pre-compose a limited set of informative words for each object of this store and determine the address of the object, apply the elements of this set in the form of the address of the object to the corresponding fields of the storage medium for removal using an appropriate reading medium, thereby forming thematic collections of storage addresses in the fields of the information carrier, at t For a mathematical search, a thematic search query is compiled and applied in the form of a search thematic prescription onto a storage medium similar to the storage medium with address fields of objects, the search thematic prescription is read from the storage medium, words according to predefined criteria are compared with the names of the readable search thematic prescription with the names printed on the storage medium fields read by the corresponding reading means, and determine the address of the object according to the results of these comparisons [13]. The disadvantages of this prototype are indicated above. Another drawback: as the number of fields (i.e., the number of lexicans of the information retrieval language) increases, the search time exponentially increases and the number of documents entered into the system decreases.

 The task, therefore, is to develop such a way of representing the maximum number of undistorted search thematic features in search thematic document descriptions, which, based on the database of thematic searches on the subject of queries without changing these queries, would ensure the issuance of thematic collections of documents with maximum local completeness and relevance if only documents are available in the subject search system. In addition, it is desirable that this method already at the stage of forming search prescriptions identify the topics on which the given topic search system can automatically provide information, as well as the probability of issuing and the volume of a thematic selection of documents.

 Employees of bibliographic search systems would not have thought of changing the bibliographic request received by them. The employees of the thematic search system, due to the low vocabulary of information retrieval languages, are forced to change thematic queries, referring, in particular, to the fact that users do not know how to make a query, do not know what they want, etc. The root cause of the non-identity of search prescriptions and thematic queries is the imperfection of empirical systems of thematic search and a very limited number of lexicons in each information retrieval language. Based on the proposed method, the thematic search system can implement “Get what you asked for” (or before the search: “There is no information in your thematic search system that is of interest to you or it’s not automatically available”).

 To solve the problem, i.e. to automatically search for the most complete and accurate thematic collections of documents (objects), a method has been developed to establish the address of the object by the search thematic feature, which consists in the fact that they pre-compose a limited set of informative layers for each object and determine the address of the object, apply the elements of this set in the form of an address object into the corresponding fields of the storage medium adapted for reading using the appropriate reading means, thereby forming in the fields of the storage medium In the thematic search, they compose a thematic search query and put it in the form of a search thematic prescription on a storage medium similar to the storage medium with address fields of objects, read the search thematic prescription from the storage medium, compare the words of the read search thematic prescription with those printed on information carriers by the names of the fields read by the corresponding reading means and determine the address of the object based on the results of these comparisons, in this case, according to the present invention, they compose a Universal Classification in advance based on the morphological and genesis relationships of the forms of matter and its attributes, the names of which are used as verbal thematic features of the desired objects, on the basis of the Universal Classification a universal information retrieval language is formed in advance , which use the classifiers of the Universal Classification, assign to each meaning of each classifier of the Universal Classification a universal code, form an array of descons of a universal information retrieval language, for each object add a limited set of informative layers to the most complete set of informative layers that characterize it, according to which they associate a given section of the Universal Classification of this object, combine identified for this object and imply corresponding to it section of the Universal Classification informative words, encode them using the aforementioned array of descons, after four Go and apply the received thematic code description of the object with its address as its search thematic description to the corresponding information carrier in the form of an integral record, and with the aforementioned compilation of the search thematic presentation, encode all informative words of its thematic query using the aforementioned array of descons of the universal information-search language .

 In FIG. 1 and 2 shows a diagram of the universal classification used in the proposed method.

 The invention will be described in detail using the subject search of documents as an example.

The method of establishing the address of an object by a search subject topic according to this invention consists of the following operations:
- pre-compile the Universal Classification of informative words, based on the morphological and genesis relationships of all forms of matter and its attributes, proven by modern theory and practice and used as verbal thematic search features of the desired objects;
- pre-form a universal information retrieval language, the lexicants of which use the classifiers of the Universal Classification;
- when turning the classifier into a lexicant, each code of each classifier of the Universal Classification is assigned a unique code;
- form an array of descons of a universal information retrieval language;
- for each object of this repository, a thematic search description is compiled in advance from the maximum number of informative words and the address of the object, while informative words characterizing this object are identified, according to which the corresponding section of the Universal Classification is associated with this object;
- combine the informative words identified for this object and implied for it by the corresponding section of the Universal Classification;
- encode these words using the mentioned array of descons;
- apply the received full thematic code description of each object with its address as its code search thematic description to the corresponding information carrier in the form of an integral record;
- when searching for an object, make up a thematic search instruction by encoding all informative words of a thematic query using the aforementioned array of descons of a universal information retrieval language;
- put the received code search thematic order on the storage medium similar to the storage medium with a search thematic description;
- read the search topic order from the information carrier;
- compare the vocabulary of a read search thematic prescription with the vocabulary of search-related themed descriptions printed on an information carrier read by an appropriate reading medium;
- determine the address of the object according to the results of these comparisons.

 The Universal Classification is shown in FIG. 1 and 2. It is based on those morphological-genesis relationships that are inherent in all forms of matter and its attributes and which are known to modern science, i.e. proved by modern theory and confirmed by practice. Universal Classification was created on the basis of the doctrine of universal classification - classifications [7 - 9 (13)], which is as objective as possible.

 In FIG. 1 and 2 arrows indicate the direction of progress (evolution, development) of the forms of matter and its attributes. In these same directions, fundamentality, universality and seniority are decreasing, and the specificity and complexity of objects, phenomena, laws, sciences and practices are increasing.

 Reading the scheme of the Universal Classification in the direction of the arrows allows us to trace the genesis of the forms of matter and its attributes. The analysis of the scheme in the opposite direction reveals the essence of the regression (degradation, revolution) of all forms of matter and its attributes.

 The step (beginning) of the OBJECTIVE WORLD scheme is enclosed in a frame, which includes the most generalizing and unique category, which can cover all the denotations of all forms of matter and its attributes.

 Part of the classification branch, going from the Subject step, is also framed to highlight objects, phenomena, laws, sciences and practices related to the theory of subjective information (informology) and understanding information only as a neuro-physiological process (phenomenon), which not characteristic of inanimate forms of matter; non-living systems operate on the basis of signals. Signals are the basis of information.

 The names of objects (material formations) are printed in FIG. 1 and 2 in capital letters without spaces between them (OBJECTIVE WORLD), phenomena - in lower case letters (movement), laws - in the form of words starting with a capital letter and connected by a hyphen, and a dot is placed after them (Law-Indestructible-Matter.), science - in large letters in discharge (MORPHOLOGY) and practice - in words in discharge with the first capital letter (Metallurgy).

 A dash over the middle of the word or on the side means that further (up or to the side) there is a branch that is identical to the branch of the adjacent word.

 A colon means an enumeration of classifiers, which is carried out up, left or right (in the direction of the arrows).

 Two points horizontally or vertically indicate that, accordingly, the list of classifiers is not complete or they are omitted, as in the ATOMA branch.

 Three horizontal dots mean that the phenomena of objects of a given level include all the phenomena of all underlying (older) objects.

 Universal Classification is a universal organization of knowledge objects or a model of Unified Knowledge that has absorbed the intelligence of modern science.

 A universal information retrieval language is formed on the basis of the Universal Classification. The verbal parts of lexicants, i.e. the "words" of this language are the classifiers of the Universal Classification, those words that are included in the corresponding levels of this classification (see Figs. 1 and 2). Moreover, each meaning of each natural word is assigned a unique code. For example, the word "key" has such meanings: a water key, a key for encryption, a key for opening locks. Therefore, each of these senses is assigned its own key: A, B, C, respectively. Then the synonyms of each meaning will receive the same code - say, “spring” and “stream” will have the same code A as “water key”. Each natural word paired with its semantic code forms a lexicant - a descon (descriptor code) of the form "Key-A", "Spring-A", "Key-B", "Skeleton Key-B", etc. All descons in alphabetical order form an array of universal information retrieval language, which is used in the future to compile search thematic descriptions and thematic prescriptions. It can also be noted here that the reverse lexicants, i.e. A-Key, B-Key, A-Stream, etc. are called codecs (code descriptor) and constitute another set of lexicans of the information retrieval language, which is used, in particular, to create an alphabetical list (array) of descons. Informative words of new knowledge become new classifiers, classifiers become codecs, and codecs become descons.

 Then, for each document of this repository, the maximum number of informative words of this document is revealed. This is the principle of the totality of theming, which is due to the concept of uncertainty associated with the fact that the developers of thematic search systems cannot foresee in advance by which thematic search terms the documents will be searched. According to these informative words, the section of the Universal Classification is determined to which this document corresponds. In fact, we are talking about the allocation of the classification domain of the document.

 The information retrieval language for documents of a given level of the Universal Classification (subject-oriented information retrieval language) must be complete in order to ensure the relevance of the search results to the query (the principle of completeness of the subject-oriented information retrieval language). This means that the number of lexical units (lexicans) of the information retrieval language of a given subject-oriented topic search system should include not only words specific to a given field of knowledge, which is covered by a specific documentary database, but also the entire terminology of all older ones, i.e. e. located to the left and below on the Universal Classification (see drawing) in comparison with this field of knowledge, for which a subject-oriented information retrieval language and a thematic search system are being developed. The principle of completeness of a subject-oriented information retrieval language is determined by the laws of conservation of elements with progress and the unidirectionality of universality. In practice, the completeness of a subject-oriented information retrieval language provides the possibility of introducing into the subject search system those documents with new knowledge that increasingly reveal the essence of this field of knowledge; however, such an information retrieval language, thematic search descriptions of documents and the thematic search databases are not redone and can be useful for any length of time. = Further, using the Universal Classification, all those words that are not found in this document, but are implied on the basis of the identified section of the Universal Classification (as in the above example, the words "organs" and "anatomy" for the words "heart", "brain", " liver").

 All identified and implied words in the document are combined and encoded using the previously described predefined array of descons. At the same time, the meanings of the words are encoded, about which it can be said that one word can have several meanings (the above example with the word "key"), or one meaning can be expressed in a few words (water, moisture, liquid, drink).

 When creating an information retrieval language only from the words of documents used in the process of generating their thematic search descriptions, the completeness principle is maintained by creating free codes that are assigned to new descriptors (word meanings) identified during the formation of search thematic descriptions. It also allows you to develop an information retrieval language and a database of subject searches without alteration.

 But the most limitless and effective possibilities will have thematic search systems, the creation of which will use the full Universal Classification and universal information retrieval language for the domain organization of the thematic search databases: these systems can become universal without changes if the corresponding documents are introduced into them. Universality allows avoiding the formation of many special systems of thematic searches.

 The presentation of documents in the thematic search descriptions by codes of exactly the meanings of words dictated by the context most fully reflects the meaning of this document, and searches by codes of meanings of words then automatically automatically provide the most complete and relevant thematic selection of documents. The resulting set of codes is a thematic code search description of the document.

 This code-based thematic search description is applied to an appropriate storage medium capable of being read using any reader or at least the possibility of mechanical sorting. Moreover, this description with the address of this document is applied in the form of an integral record.

 If it is necessary to find a document (some documents) by certain thematic attributes in this thematic search system, users compose thematic queries, which, using the same array of descons, encode and turn them into code-based search thematic prescriptions in the same way this is discussed above for code-based search engineered document descriptions. If all informative words of a thematic query can be transformed into a code search order without changing the meanings of its words and the corresponding lexicants have usage numbers, then there will always be a certain probability of an effective search.

 The resulting code search topic order is applied to the same storage medium as described above for code search themed document descriptions.

 Next, this medium is read with a code search thematic prescription using the same reader for which such a storage medium is suitable.

 The read code order is compared with the code descriptions available in the database for documents of this repository, which can be done, for example, using a computer. It is important to note that in doing so, the entire base of thematic searches is viewed, and not some part of it, as is the case in databases organized according to a unitary and not a domain principle. As a result, for one viewing, the user is given a complete thematic selection of the addresses of documents available in a particular database on a question of interest to him. The time to receive a complete thematic selection of document addresses is equal to the time during which the reader compares the search order with all the descriptions of this database of thematic searches. Using the database of bibliographic searches and other devices at found addresses, the thematic search system issues a thematic selection of documents or their copies or images on the screen.

 At the stage of compiling search-related thematic descriptions of documents, each case of using one or another deskon can be specially noted. As a result, when compiling a code search thematic prescription, even before it is read from the information medium and even before it is applied to this medium, we can conclude that there are documents of interest to the user in a specific database, since there is no descendant (the number of its occurrences or uses is equal to zero in the list of descons of a specific database) from the number contained in the search subject prescription makes automatic search in this specific database useless x If the number of uses of any deskon whose code is included in the search subject prescription is nonzero for a specific database, then this number characterizes a certain probability of obtaining document addresses using this thematic search system for this search subject prescription, as well as the maximum possible the number of these addresses. This probability is the higher, the shorter the prescription and the greater the number of uses of its descons.

The considered method ensures the achievement of the following technical results:
1. The information content of large thematic search systems is increased by many orders of magnitude - from 10 ¹⁰ for empirical thematic search systems to 10 ¹⁰⁰ and higher for informological systems (due to the principle of the totality of theming and the maximum completeness of search-related thematic descriptions of documents or representativeness).

 2. An increase by hundreds of times of the completeness and tens of times of relevance of the thematic collections of documents, since the problem of polysemy, homonymy and synonymy is eliminated at the stage of creating the information retrieval language, as well as due to the principle of totality of theming.

 3. The implementation of the thematic search in the automatic mode of computer operation and a one-time search on one query topic due to the combination of principles when creating an information retrieval language and search thematic descriptions of documents.

 4. Simplification of the machine language due to the domain organization of databases of thematic searches, reducing the search time on one topic due to the dialoguelessness and automatic search, and therefore cheaper the search for thematic collections of documents.

 5. Unification of the vocabulary of search thematic databases through the use of ready-made full Universal Classifications and the universality of the information retrieval language, which allows you to create a network of thematic search systems without additional costs for the development of translator languages, required when each thematic search database has its own information retrieval tongue.

 6. The increase in the number of types and types of computers used as the computer basis of the thematic search system due to the simplification of the machine language of searches in the domain organization of databases of thematic searches.

 7. Dialog-free (single, automatic) allows the machine to process all databases of thematic searches (at a time) to obtain thematic collections of documents on many requested topics through the use of computers with parallel processors. The parallelism of searches speeds up and reduces their cost price even more.

 8. The domain-based organization of thematic search databases makes it possible to use the entire capacity of the carrier with benefit, which is unattainable with the unitary organization of the thematic search databases, when about half of the storage capacity “walks” free for various reasons.

 9. In the case of domain-based organization of thematic search databases, there will be practically no duplication of addresses that are recorded once and which, when the unitary organization of thematic search databases are recorded, as many times as the keywords of this document are presented in the corresponding fields of the medium.

 10. In the formation of a universal information-search language by coding the meanings of natural words, the problem of polysemy, homonymy and synonymy is solved once, which is solved with the help of information-search languages with natural words in the role of lexicants each time the thematic queries are converted into search thematic prescriptions - dozens million times annually worldwide.

 11. Within one thematic search system with a domain organization of thematic search databases, it will be enough to have one document base, one Universal Classification, one universal information retrieval language and one thematic search database, while empirical thematic search systems, as a rule, have Dozens and more of documentary databases, classifications, information retrieval languages and databases of thematic searches (files).

 12. Due to the average of stock-code search-related thematic descriptions of documents, it becomes possible to conduct thematic searches according to search prescriptions (thematic queries) theoretically from one-lexicant to stolexicant (on average to ten-lexicant) search thematic prescriptions, to change the number of lexicants in the prescriptions within these limits, many times more precisely, automatically control the volume of thematic collections of documents and the nature of the information contained in them: the more lexicans in the prescription, the less e volume and more specific information in thematic collections of documents, and vice versa. Unitary databases of thematic searches allow achieving regulation ranging from one, on average, to only three-lexicant search thematic prescriptions, which is why thematic collections of documents, as a rule, are very large, include a lot of unused information (information noise) and require visual-manual dialogue or interactive or browsing (automated) sorting and replenishment, which takes hours, or even days and weeks of work.

 13. Informological systems of thematic searches can serve as long as you like.

 14. In the process of theming, the numbers of the use of the lexicans of the universal information retrieval language are put down, which makes this language real, that is, suitable for thematic searches only in this thematic search system. The reality of a ready-made universal information retrieval language allows you to calculate and know in advance the probability of obtaining and the maximum possible volume of a future thematic collection of documents before searching by the numbers of lexicants whose codes are included in the search thematic prescription, by changing this number before searching by changing the number of codes in the prescription. This knowledge allows, without conducting searches, to select the system of thematic searches on the Internet such as the Internet, which is more likely to give the necessary thematic selection of documents. In other words, the reality of information retrieval languages allows you to search for optimal thematic search systems, for example, using a computer database of information retrieval languages in a network of topic search systems. More precisely, it creates the opportunity to precede the search for a thematic selection of documents by searching for the optimal system of thematic searches.

 15. The integrity of the record of the search for thematic descriptions allows you to make it arbitrarily large, and the domain database of thematic searches on an infinite number of carriers.

The considered method is implemented in several thematic search systems, in particular, in the Biomed system [7 - 9 (13)], which, based on 15,000 abstracts, is theoretically capable of delivering thematic collections of documents on approximately 10 ¹² topics (information indicator). This system is implemented on maps with edge perforation having 17 fields. The average description includes 12 lexicans. Each field has 141 silent codes. The information-search language of the Biomed thematic search system includes only real lexical units of 10,000 lexicants used in the compilation of search-related thematic descriptions of documents and has about 5 million free codes. The total number of potential thematic collections is 3 ¹⁷ - 1 ≈ 10 ¹² . Local completeness for several thousand searches performed is on average 30%, relevance - 50%. Thematic collections of documents are made only on real thematic requests.

For comparison, the real information content of all systematic cards of the Russian State Library is equal to the number of collections of bibliographic cards, which is about 10 ¹⁰ . The completeness and relevance of subject collections of documents issued by empirical subject search systems is, on average, equal to 1% and 5%. Of the 10 ¹⁰ ready-made thematic collections of cards of this library, only a few percent are relevant to real requests just like ready-made thematic collections of storage addresses in computer databases of thematic searches with a unitary organization.

 The illustrated illustrative example of the implementation of the method of the present invention for searching documents is in no way limiting, since this method is suitable for searching for any objects whose addresses are known. The scope of patent claims is determined only by the attached claims.

Claims (3)

 1. The way to establish the address of an object by a search thematic feature, which consists in preliminarily compiling a limited set of informative words for each given object and determining the address of the object, applying the elements of this set in the form of the address of the object to the corresponding fields of the storage medium adapted for reading with corresponding reading means, thereby forming thematic collections of addresses of objects in the fields of the information carrier, when performing a thematic search, make up a thematic search request and apply it in the form of a search thematic prescription to a storage medium similar to a storage medium with object address fields, read the search thematic prescription from the storage medium, compare the words of the read search thematic prescription with the field names printed on the storage medium read by the corresponding reading means, and determine the address of the object according to the results of these comparisons, characterized in that they pre-compile a Universal Classification based on sea of the genetic-genesis relations of the forms of matter and its attributes, the names of which are used as verbal thematic features of the desired objects, on the basis of the Universal Classification, a universal information-search language is formed in advance, the lexicants of which use the classifiers of the Universal Classification, assign each meaning to each classifier of the Universal Classification unique code, form an array of descons of a universal information retrieval language, for each object additional they form a limited set of informative words to the maximum complete set of informative words characterizing it, according to which the corresponding section of the Universal Classification is associated with a given object, the informative words identified for this object and implied for it by the corresponding section of the Universal Classification are combined, encoded using the mentioned array of descons, after which is why the obtained thematic code description of the object with the address as its search theming is applied nnogo describing the corresponding media in the form of whole record, and in said drawing thematic search instructions encoded word corresponding to all informative theme request by said array DESCON universal information retrieval language.  2. The method according to claim 1, characterized in that as the descons of the universal information retrieval language of a particular subject search system are used to compile code-based thematic descriptions, the number of uses for each descon is indicated.  3. The method according to claim 1 or 2, characterized in that after compiling a code search topic order before searching using the numbers of descons whose codes are included in the said order, they determine the probability of obtaining the addresses of objects using this topic search system for this search topic order and the maximum possible number of these addresses.

Images