DE102019200474A1 - Input system and method for ensuring uniform terminology in an application accessible to multiple users - Google Patents

Abstract

The invention relates to an input system (10) for ensuring uniform terminology conventions for terms in an application (20) that is accessible to several users (18). The input system (10) has a filter (12), comprising a database (14) and an algorithm (16). Based on a user input (22), the database (14) first generates an output list (24) for the user input (22) based on a predetermined search rule (26). The algorithm (16) weights and sorts the output list (24) according to a weighting rule (30) and according to user data stored in the user profile (28) for the respective user (18) and provides the sorted output list (24) to the user (18) Available who makes a selection from the output list (24) that he wants to use as a term in the application (20). Based on the selection, the algorithm (16) generates a learning data record (36), on the basis of which it adjusts the weighting rule (30). The algorithm (16) uses the learning data record (36) and / or the adapted weighting rule for all users (18) of a user group (38).

Description

The invention relates to an input system for ensuring uniform terminology in an application that is accessible to several users. Such an application can be, for example, a program for writing text, in particular a program for writing e-mails, and / or such an application can also be a program for writing specification sheets and / or work instructions and / or operating instructions, for example. The users can, for example, be employees of a company and / or a group, in particular a globally operating group. A respective conceptual convention prescribes how to name and / or abbreviate a certain object in a certain language. The terms used can be, for example, company-internal designations and / or common abbreviations. Furthermore, the invention relates to a method for ensuring uniform terms in an application accessible to several users using an input system according to the invention.

Various term management systems are known from the prior art. So the reveals DE 10 2011 012 068 A1 a method for managing technical terms. This procedure is intended to build up an archive of technical terms that can be used by several people. This is to avoid misunderstandings when using technical terms.

From the US 2011/0238584 A1 a term management database is known. Several users, for example employees of a company, can search for company-specific terms in this term management database. This is to ensure that, on the one hand, users understand company-specific terms better and, on the other hand, only use terms that have been approved by the company.

A disadvantage of the known prior art is that the term databases described are only available for in-house use. There is no interface to a broader knowledge pool, such as a global term database.

A problem with an application that is accessible to multiple users can consist in a non-uniform definition of a term, i.e. a certain abbreviation can have a different meaning for a subset of users than for another subgroup, without this being wrong because the abbreviation of technical terms from different subject areas can lead to the same abbreviation. An example of this is the abbreviation ESC, which stands for "Escape" in the computer area and "Electronic stability control" in the automotive area. Another problem arises with newly created abbreviations to newly introduced technical terms. It must be coordinated here that all affected users agree on the same abbreviation. The use of the correct abbreviation for a specific user group or a specific field is a form of the convention. Another form of the convention is the correct translation of a technical term into another language for a specific user group or a specific specialist area. For example, the translation of the technical term "electronic stability control" into German is "electronic stability program" and not "electronic stability control". In order to obtain an abbreviation and / or translation for a certain term in accordance with the terms of the convention, it is necessary to monitor in a complex manner which term convention has developed for a certain user group or a certain department.

The invention has for its object to support a user group in the input of uniform and correct for their specialist terms in an application. Here, a term is to be understood as an abbreviation and / or an advertisement for an abbreviation (full word) and / or a translation.

The object is solved by the subject matter of the independent claims. Advantageous developments of the invention are described by the dependent claims, the following description and the figures.

The invention provides an input system for ensuring uniform terminology in an application that is accessible to several users. Such an application can be, for example, a program for composing an email. Such an application can additionally or alternatively be a program for writing a specification and / or a work instruction. The input system has a filter that includes a database and an algorithm. The database is set up to receive user input for which the term convention valid for the term is to be ensured and which is intended for the application. In other words, a user can make an entry directed to the database. However, this input cannot be transferred directly to the application because it must first be ensured that the input does not contain a term that violates the term conventions intended for the application.

The database is set up to generate an output list based on the user input. The database follows a predetermined search rule. In other words, a user can, for example, write a text for a technical work instruction. To ensure that he uses unambiguous and / or uniform terms when writing this text, the user can enter a respective term in the database. On the basis of a stored search rule, the database can then propose to the term entered by the user one or more terms associated with the term entered by the predetermined search rule, which terms can be stored in one or more database entries. The search rule can, for example, search out all those database entries that match the user input to a predetermined minimum degree with a so-called database key of the entries, which can be determined, for example, by means of a so-called string matching method or a string comparison for a 100 percent match can. In its entries, the database preferably contains all those terms which correspond to the term convention, that is to say in particular a valid definition or correspondence for user input in relation to an abbreviation (for an entered full word) and / or a full word (for an entered abbreviation) and / or a translation (as a full word and / or abbreviation). If there is ambiguity, for example, a user input that represents an abbreviation, an output list with several database entries can of course also result from the database.

The algorithm is set up to carry out a weighting of each entry in the output list (list entry) on the basis of a user profile assigned to the user by means of a predetermined weighting rule. The user profile of the respective user can be formed from the user identity and / or from at least one indication of an area or departmental affiliation of the user within a company (generally an affiliation to a user group) and / or from a language assigned to the user and / or from naming the special interests of the user and / or from a previous search history and / or from entering keywords and / or from a contact network assigned to the user, i.e. taking into account the user's contacts, and / or a profession (profession and / or Education). A contact can be the user profile of a colleague of the user.

The algorithm is also designed to sort the output list according to the weighting and to make the sorted output list available to the user, to receive a selection from the user and, depending on the selection from the user input, to generate a revised user input and send it to the user Pass on application. In other words, the user can make a selection from the output list generated. The selection can concern the first hit in the output list or any other hit (which means that the sorting and thus the weighting was incorrect for the user). The algorithm can carry out a weighting of an entry in the output list in such a way that, after sorting, it is the first entry in the output list to indicate the one with the estimated greatest relevance or suitability for the user. The weighting therefore indicates the estimated relevance. This means that the algorithm provides each user with an output list sorted by relevance based on his or her user profile. However, the user can still select any hit from the output list, which is then passed on to the application by the algorithm. Only one of the terms proposed by the algorithm on the basis of the user input is therefore used in the application.

The invention is characterized in that the algorithm is set up to generate a learning data record based on a link between the user profile of the user and the user input of the user and the selection made by the user of the entry in the output list. In other words, the algorithm learns from the combination of user input, selection of the entry in the output list, and user profile of the user. A respective learning data record can be formed from a combination of user information and / or a previous search query (user input) and / or selection from the output list and / or a combination of user information with corrections or other inputs by the respective user.

The invention is further distinguished by the fact that the algorithm is set up to use the learning data record to adapt the weighting rule by means of a predetermined training method and to use the learning data record generated for the respective user of a user group and / or the adapted weighting rule for all users of the same user group. This means that all members of a user group benefit from the training that the algorithm undergoes based on the use of the input system described here by a respective user of the user group. As a training method e.g. a training for a neural network or generally a method of machine learning can be applied.

The invention results in the advantage that, by taking the user profile of a respective user into account by the algorithm, a custom-made output list of possible terms to be used or approved terms that are tailored to the needs of the respective user is generated. This significantly simplifies the selection of a particular term for the user. In addition, the invention has the advantage that by training the algorithm together for all users of a respective user group, the algorithm learns particularly effectively and quickly which conceptual convention applies within the user group. The algorithm is then adapted for all users in the user group in such a way that the weighting rule is reconfigured by the selection of the one user. Thus the conceptual convention is standardized within the user group.

The invention also includes embodiments which result in additional advantages.

One embodiment provides that the algorithm is set up to generate a new entry in the database on the basis of the user input and depending on the user profile of the user in accordance with a predetermined coding rule and to assign it to at least one category. A category is e.g. a level of confidentiality. In other words, the algorithm can be set up to provide a new entry in the database with a confidentiality level. A level of confidentiality can be "strictly confidential" or "confidential" or "public".

In other words, the database can be expanded by user input. The extension is not arbitrary, but is based on a predetermined coding rule. The coding rule takes into account the user's profile, e.g. the area or department affiliation of the user and / or a mention of special interests of the user and / or last search queries of the user and / or various user entries, such as keywords and / or contacts of the user. Furthermore, the coding rule can also take into account recent search queries from users who belong to the same user group as the user and / or who are in the user's contact network. The user can define at least one term that should belong to the new entry in the database. He can allow at least one term to be chosen flexibly. The assignment of the new entry to a category is also based on user input.

A preferred embodiment provides that the database comprises a list of abbreviations, wherein in the list of abbreviations at least one respective abbreviation and at least one respective full word (plain text) are assigned to one another as data record parts and form a respective data record. In other words, the database includes a list of abbreviations and full words that are associated with one another. Any number of full words can be assigned to an abbreviation. In other words, entering an abbreviation in the database can result in the output of an output list that includes several entries. Several abbreviations can also be assigned to a full word.

Another embodiment provides that a respective data record of the database is available in several languages. There is then such a data record several times, namely in different languages. In other words, there is a respective translation in at least one other language for a respective data record in the database. Each data record part of a data record in the database that is available in several languages is permanently linked to a respective language feature. The language feature indicates the language in which the respective data record part is written. Each data record part of a data record available in several languages comprises two pieces of information. The first piece of information is an abbreviation or a full word and the second piece of information is a language feature.

The search rule is set up to generate an output list when entering a first record part of a record in several languages in one language, which contains all the second record parts assigned to the entered first record part and all translations assigned to the record in accordance with an assignment rule in all languages in which they exist in the database and / or are selected by the user. The search rule is also set up to assign a respective translation to the first data record part assigned to the respective translation in its respective language. The language characteristic of the translation and the language characteristic of the first part of the data record assigned to the translation thus match. In other words, the translation and the abbreviation associated with the translation are in the same language.

According to this embodiment, it can therefore be provided that a user input comprises an abbreviation and a language feature that indicates the language in which the abbreviation was entered. Based on such user input, for example, an output list with two hits can be produced. The first hit can include a full word and a first language feature, which indicates which language the first hit is in. The second hit of the output list generated in this way can comprise a full word and a second language feature which indicates the language in which the second hit is available. If the second hit is in a different language than the first hit and the user input, the user is additionally offered the abbreviation associated with the second hit in combination with the second language feature.

For example, a user can write a user input that includes the abbreviation “ESP” and the language feature “German”. The result list then includes the hits "electronic stability program" linked with the language feature "German" in the first place and "electronic stability control" linked with the language feature "English" in the second place. The language characteristic of the second hit ("English") therefore does not match the language characteristic of the user input ("German"), nor does it match the language characteristic of the first hit (also "German"). The second hit is therefore automatically assigned the abbreviation “ESC” linked to the language feature “English”.

Another embodiment provides that the database is a global database that is set up to be fed by a local database. In other words, the database can be a publicly accessible database that can be stored on an Internet server, for example. A local database can be an internal company database, for example. If there is an entry in the company's internal local database in a category marked as publicly accessible, it can be fed into the global database. It can also be provided that an entry is made by a user in the local database and passed on to the global database, provided that no hit is found in the local database. This has the advantage that very large amounts of data can be accessed. At the same time, the restricted area of the local database is provided, in which confidential terms are subject to restricted access. An access restriction can, for example, be stored in a user's user profile.

A further embodiment provides that the input system comprises a user management module, each of the respective users being assigned to at least one user group of the user management module. A user is assigned to a user group based on the user profile assigned to the user. The user management module can recommend new areas of interest to the user based on his usage statistics and thus contribute to the individualization of the user profile. The user management module can be integrated into the database, or the algorithm can be accessed via an external interface.

A further embodiment provides that the input system comprises a quality management module which is set up to correct and / or delete a data record of the database if at least a predetermined number of users marks the data record as incorrect. In other words, a user can identify a bad record. If at least a predetermined number of users mark a data record as faulty, the data record can be corrected or deleted. This has the advantage that the database can be cleaned up.

A further embodiment provides that the input system comprises a graphical user interface, which is set up to record user input and to display an output list. In other words, it can be provided that the input system can be operated via a graphical user interface, for example that of a computer terminal. It can be provided that there is an input field and an output field. This has the advantage of particularly simple operation.

The invention also relates to a method for ensuring uniform terminology conventions for terms in an application accessible to several users using an input system, the input system having a filter that includes a database and an algorithm.

According to the method according to the invention, the database receives user input for which compliance with the conceptual convention is to be ensured and which is intended for the application. Based on the user input, the database generates an output list based on a predetermined search rule. On the basis of a user profile assigned to the user, the algorithm carries out a weighting of an entry in the output list using a predetermined weighting rule and sorts the output list according to the weighting. The algorithm provides the sorted output list to the user.

The algorithm receives a selection by the user and, depending on the selection from the user input, generates a revised user input which it passes on to the application. Based on the user's profile, the user's input, and the selection made by the user of the entry in the output list, the algorithm generates a learning data record, on the basis of which it adjusts the weighting rule in accordance with a predetermined training method.

The algorithm uses the learning data record generated for the respective user of a user group and / or the adapted weighting rule for all users of the user group.

The invention also includes further developments of the method according to the invention which have features as have already been described in connection with the further developments of the input system according to the invention. For this reason, the corresponding developments of the method according to the invention are not described again here.

The invention also includes combinations of the features of the described embodiments.

• 1 eine schematische Darstellung eines erfindungsgemäßen Eingabesystems;
• 2 eine schematische Darstellung der Zuordnung von in der Datenbank befindlichen Datensätzen zueinander gemäß einer Ausführungsform der Erfindung;
• 3 eine schematische Darstellung der Zuordnung von in der Datenbank gespeicherten Datensätzen zueinander gemäß einer weiteren Ausführungsform der Erfindung;
• 4 eine schematische Darstellung der Erzeugung eines neuen Eintrags in der Datenbank gemäß einer vorbestimmten Codierungsvorschrift;
• 5 eine schematische Darstellung verschiedener Ebenen von Interessensgruppen und deren Hierarchie; und
• 6 eine schematische Darstellung des erfindungsgemäßen Verfahrens gemäß einer Ausführungsform.

Exemplary embodiments of the invention are described below. This shows:

• 1 a schematic representation of an input system according to the invention;
• 2nd a schematic representation of the assignment of data records located in the database to each other according to an embodiment of the invention;
• 3rd a schematic representation of the assignment of data records stored in the database to each other according to a further embodiment of the invention;
• 4th a schematic representation of the creation of a new entry in the database according to a predetermined coding rule;
• 5 a schematic representation of different levels of interest groups and their hierarchy; and
• 6 is a schematic representation of the inventive method according to one embodiment.

The exemplary embodiments explained below are preferred embodiments of the invention. In the exemplary embodiments, the described components of the embodiments each represent individual features of the invention that are to be considered independently of one another, which further develop the invention independently of one another and are therefore also to be regarded individually or in a combination other than the one shown as part of the invention. Furthermore, the described embodiments can also be supplemented by further features of the invention that have already been described.

In the figures, the same reference numerals designate elements that have the same function.

1 shows an input system 10th which has a filter 12th includes. The filter 12th assigns a database 14 as well as an algorithm 16 on. Furthermore shows 1 a user 18th as well as an application 20 . The user 18th can be a user input 22 into the database 14 enter. Database 14 can the user input 22 received and based on user input 22 an output list 24th based on a predetermined search rule 26 (not in 1 shown). In the embodiment shown here, the database 14 also a user profile 28 include and this the algorithm 16 provide. The algorithm 16 can based on the user profile 28 by means of a predetermined weighting rule 30th (not in 1 shown) a weighting of an entry 32 on the output list 24th carry out and the output list ( 24th ) sort by. The algorithm 16 can the sorted output list 24th the user 18th provide.

The user 18th can hit the output list 24th choose. Based on this user's choice 18th can the algorithm 16 a revised user input 34 generate and apply this to the 20 pass on.

The algorithm 16 can the user profile 28 of the user 18th with user input 22 of the user 18th and with that by the user 18th selection of the entry 32 the results list 24th link and a learning record 36 based on this link. Based on the learning data record 36 can the algorithm 16 the weighting rule (not shown here) 30th to adjust. In addition, the algorithm 16 for the respective user 18th a user group 38 (not shown here) generated learning data record 36 and / or the adjusted weighting rule ( 30th ) to all users in the user group 38 apply.

2nd shows an exemplary and schematic representation of the in the database 14 linked entries. According to the embodiment described here, the database comprises 14 a list of abbreviations 40 (not in 2nd shown). In the list of abbreviations 40 is a combination of an abbreviation 42 and a language feature 44 a variety of combinations of a respective plain text 46 and a respective additional language feature 48 assigned. The individual plain texts 46 can be the same or different. The language features 48 can be the same or different. You can also do the same Language feature 44 be. But they can also be different from the language feature 44 be. If a language feature 48 from the language feature 44 is different, the relevant combination of plain text 46 and language feature 48 automatically a combination of abbreviation 42` and language feature 48 assigned.

The in 2nd The exemplary embodiment shown can be based on the following situation. One user 18th returns to the database 14 as an abbreviation 42 the term "ESP". As a language feature 44 he chooses "German". For example, two combinations are assigned to the input. The first combination can be from the plain text 46 "Electronic Stability Program" and the assigned language feature 48 "German" exist. The second combination can be from plain text 46 "Electronic stability control" and the other language feature 48 "English" exist. Because in the second combination the language feature 48 from the language feature 44 is different, the corresponding combination of abbreviation is automatically here 42` according to the language feature selected here 48 "English" selected. The abbreviation 42` is "ESC" in this case.

3rd shows how analogous to this as user input 22 from a respective user 18th a plain text 50 in combination with a language feature 44 can be chosen. As a result of such user input 22 are combinations of an abbreviation 52 and an associated further language feature 54 offered. Deviates the language feature 54 thereby from the language feature 44 a combination of the corresponding plain text is automatically created 50` with the appropriate language feature 54 generated.

The in 3rd described embodiment may be based on the following situation. One user 18th returns as user input 22 the plain text 50 "Electronic stability program" in combination with the language feature 44 "German". As an output list 24th For example, two entries are created. The first entry is a combination of the abbreviation 52 "ESP" with the language feature 54 "German". The second entry is a combination of the abbreviation 52 "ESC" in combination with the language feature 54 "English". In the second case, the language feature differs 54 , namely "English", from the language feature 44 , namely "German". Thereupon a combination of the English plain text is automatically created 50` "Electronic stability control" with the appropriate language feature 54 , namely "English".

4th shows an example of how based on a coding rule 52 (not shown here) new entries for the database can be created. One user 18th For example, choose two fixed terms 54 , which should definitely be part of the new entry. The two terms 54 can be different. The coding rule 52 now creates based on user input 22 placeholder 56 . Depending on a division or departmental affiliation and / or a nomination of interest and / or a last search and / or as a result of further user input and / or through last search queries of contacts in the contact network of the respective user and / or due to a general cluster of terms and / or the placeholders become based on a definition of a language characteristic 56 through the coding regulation 52 filled with corresponding entries. For example, a user 18th as fixed terms 54 choose a company name on the one hand and the word "system" on the other. Based on the coding regulation 52 can the algorithm 16 now based on the input factors mentioned above and the user profile 28 of the user 18th the placeholders 56 to fill. The algorithm can also generate a new abbreviation from the new entry created in this way.

Any user 18th is according to his user profile 28 at least one interest group 58 just like you in 5 shown are assigned. The respective interest groups 58 are further divided into sub-areas 60 . The subareas 60 can be divided into individual sections 62 . An interest group 58 can be "Media", for example. A subdivision of this interest group 58 in sub-areas 60 may happen. The subareas 60 in this case can be "Online", "TV", "Radio" and others. Such a subdivision of interest groups 58 enables a particularly precise assignment of a respective user input 22 and thus efficiently improves the training of the algorithm 16 .

• - S1: Die Datenbank 14 empfängt eine Benutzereingabe 22, für welche eine Einheitlichkeit der Begriffe sichergestellt werden soll und die für die Anwendung 20 bestimmt ist.
• - S2: Die Datenbank 14 erzeugt auf Basis der Benutzereingabe 22 eine Ausgabeliste 24 basierend auf einer vorbestimmten Suchvorschrift 26.
• - S3: Der Algorithmus führt auf Basis eines dem Benutzer 18 zugeordneten Benutzerprofils 28 mittels einer vorbestimmten Gewichtungsvorschrift 30 eine Gewichtung eines Eintrages 32 der Ausgabeliste 24 und eine Sortierung der Aufgabeliste 24 durch.
• - S4: Der Algorithmus 16 stellt die sortierte Ausgabeliste 24 dem Benutzer 18 zur Verfügung.
• - S5: Der Algorithmus 16 empfängt eine Auswahl durch den Benutzer 18 und erzeugt in Abhängigkeit von der Auswahl aus der Benutzereingabe 22 eine überarbeitete Benutzereingabe 34.
• - S6: Der Algorithmus 16 gibt die überarbeitete Benutzereingabe 34 an die Anwendung 20 weiter.
• - S7: Der Algorithmus 16 erzeugt basierend auf dem Benutzerprofil 28 des Benutzers 18, der Benutzereingabe 22 des Benutzers 18 und der durch den Benutzer 18 getroffenen Auswahl des Eintrags 32 der Ausgabeliste 24 einen Lerndatensatz 36.
• - S8: Der Algorithmus 16 passt anhand des Lerndatensatzes 36 die Gewichtungsvorschrift 30 an.
• - S9: Der Algorithmus 16 verwendet den für den jeweiligen Benutzer 18 einer Benutzergruppe 38 erzeugten Lerndatensatz 36 und/oder die angepasste Gewichtungsvorschrift für alle Benutzer der Benutzergruppe 38.

This in 6 The method shown includes the steps:

• - S1: The database 14 receives user input 22 , for which a uniformity of the terms should be ensured and for the application 20 is determined.
• - S2: The database 14 created based on user input 22 an output list 24th based on a predetermined search rule 26 .
• - S3: The algorithm runs on the basis of a user 18th assigned user profile 28 by means of a predetermined weighting rule 30th a weighting of an entry 32 the output list 24th and a sorting of the task list 24th by.
• - S4: the algorithm 16 provides the sorted output list 24th the user 18th to disposal.
• - S5: The algorithm 16 receives a selection from the user 18th and generated depending on the selection from the user input 22 a revised user input 34 .
• - S6: the algorithm 16 returns the revised user input 34 to the application 20 further.
• - S7: the algorithm 16 created based on the user profile 28 of the user 18th , user input 22 of the user 18th and that by the user 18th selection of the entry 32 the output list 24th a learning record 36 .
• - S8: the algorithm 16 fits based on the learning data record 36 the weighting rule 30th on.
• - S9: the algorithm 16 uses that for the respective user 18th a user group 38 generated learning data record 36 and / or the adjusted weighting rule for all users of the user group 38 .

In a specific embodiment, the actual directory (that is, the database) is based on a list of abbreviations and full words assigned to one another. This list is filled with fixed abbreviation assignments (abbreviation to full word combination and vice versa) by users or an administration.

The abbreviations are divided into categories that will later serve as interest groups. These are enriched by meta information from user searches (i.e. with information that is stored in a user's respective user profile).

Abbreviations in various languages are stored in the list (i.e. in the database) and the respective assignments are linked (e.g. ESP (DE) == ESC (EN) and Electronic Stability Program (DE) == Electronic Stability Control (EN)).

Using a search function (i.e. based on a predetermined search rule), the list (i.e. the database) can map both abbreviations to plain texts and plain texts to abbreviations (i.e. generate an output list). All theoretically possible results can be displayed.

There is a local directory (ie a local database), which is an information sink, provided that the information is marked as "internal / confidential" (i.e. categorized), so that the data in the local directory is reserved for the respective user group.

There is a global directory (i.e. a global database) that can be a source for all queries that come to no or inadequate results in the local directory or require more information requested by the user. The global directory is a sink of information, provided that information is marked as "public" (i.e. categorized).

The functionalities are made available to the users by means of a graphical (user) interface. For administration, access is also implemented using a graphical (user) interface. Each user (i.e. each user) can add abbreviations (i.e. entries) via an input mask, which are transferred to the local or global directory depending on the data classification (i.e. categorization). For this purpose, a date can be assigned when information (abbreviation + full word) (i.e. a data record) automatically changes its classification from confidential → internal → public. The local and global directories are hard-linked. This means that all information on a data record is transferred from one directory to the other directory if there is a corresponding release for the respective data record.

In a specific embodiment, user management (ie a user management module) is provided. A user can be any user from a global list. Each user receives a user name or authentication and is thus assigned to a specific user group. This user system (i.e. the user management module) is maintained within the abbreviation directory (i.e. in the database) or is based on an external interface. For large organizations, this can be done using the departmental and departmental organization and standard authentication for IT systems. There is an interface to IT user management systems for the transfer of assignments and rights. The assignment to an area (i.e. a user group) can be done by the user (i.e. the user) himself.

Possible areas (i.e. user groups) are software development, process, systems engineer, hardware development, human resources or others.

Multiple mentioning is possible when assigning to interest groups. Interest groups include development, IT, social affairs, controlling, procurement and others.

The user management (i.e. the user management module) recommends the user based on usage statistics and content Evaluation of new interest groups and more detailed individualization of the profile. In order to achieve a high quality of the recommendations, the user can provide inputs that are automatically scanned for keywords / interests (emails, cookies, browser history, etc.) that establish a relationship with other users (contact and environment analysis).

The search queries can be released anonymously to other users of the area, the interest group, etc. - the interest group itself is enriched with metadata.

In a specific embodiment, a quality management (ie a quality management module) is provided. Every user (i.e. every user) has the opportunity to correct incorrect or inadequate information (in the database) or to request a deletion. If information (i.e. a data record in the database) is marked as incorrect by many users (threshold must be defined and fit relative to the use of the information), it is deleted. The predetermined number of users who must mark a data record in the database as incorrect so that it is corrected and / or deleted can depend on how relevant the data record proves to be over a predetermined period of time, i.e. how often it is used, for example is called by a user.

In a specific embodiment, it can be provided to classify information (that is, data records of the database) (that is, to assign it to categories). Information for the list of abbreviations (i.e. data records for the database) must be classified. Possible classifications (i.e. categories) are: confidential (abbreviations and plain text remain within the specified user group (strong restrictions)), internal (abbreviations and plain text remain within the specified user group), public (data are made publicly available and feed the global database).

Existing abbreviations can be decoded on the basis of data analysis. The search for an abbreviation can result in an ambiguous plain text representation + language as well as a reference to the abbreviation in the respective language. It is considered a special case if there is a clear mapping from abbreviation to plain text + language. The decoding is then completed in this case.

The search for a plain text can result in an ambiguous abbreviation + language and a reference to the plain text in the respective language. It is also a special case here if there is a clear representation of an abbreviation for plain text + language. The decoding is then completed in this case.

The evaluation of the recommendation (i.e. the weighting of an entry in the output list and a sorting of the output list) is based on the respective user profile (i.e. on the user profile of the respective user) and prioritizes or weights the results according to relevance for the user (i.e. according to a predetermined one Weighting rule). The weighting (i.e. the weighting rule) takes the following aspects into account: belonging to a department or department, mentioning interests, last searches, result of user inputs (keywords, contacts, ...) for evaluation.

If an internal analysis (i.e. a search in the local database) only provides inadequate results or if the user (i.e. the user) requests further results, the global directory (i.e., based on the user information (i.e. based on the user input)) access to the global database).

Abbreviations can be encoded on the basis of data analysis (that is, a new entry in the database can be created in accordance with a predetermined encoding rule) (“abbreviation generator”). If a clear text description is required for an abbreviation, it can be generated using a generator. If an extension (e.g. technical filler word) is required for a full word combination (i.e. for fixed terms), the form is determined on the basis of the plain text specifications (i.e. the fixed terms) and placeholders; the placeholders are filled with plain text suggestions. The algorithm can also generate a new abbreviation from the new entry created in this way. The coding regulation takes the following aspects of the user profile of each user into account: area or department affiliation, interest, last searches, result of user inputs (keywords, contacts, ...) for evaluation, searching the contact network, general accumulation of terms and / or systems and / or technologies, etc. in the global directory (i.e. in the global database). The required languages (i.e. language characteristics) are defined as basic information and defined whether the same abbreviation should be used in all languages or whether deviations in automatic translation (e.g. ESP == ESC) are permitted.

In a specific embodiment, the learning component of the system comprises the following basics for the self-learning component (that is, to form a learning data record): user information mapped to search queries and selections, selection and / or corrections and / or inputs, etc. of the users, general distribution of terms for coding component ("abbreviation generator").

Information for adapting the weights is also taken into account in order to form the learning data record. For example, there is a change in the weighting of the results based on changing user behavior and internal and / or external user settings (such as a change of area, etc.).

The results (ie the entries in the output list) are displayed in accordance with the predetermined search rule and the predetermined weighting rule. The results of the coding and decoding are based on the relevance for the respective user. The relevance of the results is weighted by an algorithm based on the following factors: interest groups, degree of agreement (in the case of plain text search), frequency of the search query, information from user profile (frequency of keyword analysis, contact network, etc.).

Definition of categories and / or interest groups: If an interest group exceeds its defined upper limit of information, the group is divided into sub-areas and / or sub-areas based on the main topics. Interest groups can thus be chosen more broadly or more specifically, depending on the needs of the user.

Overall, the examples show how the invention can be used to support a group of users in entering uniform terms that are correct for their specialist area into an application.

QUOTES INCLUDE IN THE DESCRIPTION

This list of documents listed by the applicant has been generated automatically and is only included for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 102011012068 A1 [0002]
• US 2011/0238584 A1 [0003]

Claims (9)

Input system (10) for ensuring uniform terminology conventions for terms in an application (20) accessible to several users (18), the input system (10) having a filter (12) which comprises a database (14) and an algorithm (16) The database (14) is set up to receive a user input (22) for which the uniformity of the terms is to be ensured and which is intended for the application (20), and an output list () based on the user input (22). 24) based on a predetermined search rule (26), and wherein the algorithm (16) is set up to use a predetermined weighting rule (30) to weight each entry (32) based on a user profile (28) assigned to the user (18) ) perform the output list (24), sort the output list (24) according to the weighting, make the sorted output list (24) available to the user (18), a selection by ch to receive the user (18) and, depending on the selection from the user input (22), to generate a revised user input (34) and to pass this on to the application (20), the algorithm (16) also being set up based on generate a learning data record (36) on the user profile (28) of the user (18), the user input (22) of the user (18) and the selection (32) made by the user (18) of the entry (32) of the output list (24), on the basis of which he adjusts the weighting rule (30) by means of a predetermined training method, and the learning data record (36) generated for the respective user (18) of a user group (38) and / or the adjusted weighting rule (30) for all users (18) of the user group (38) to be used. Input system (10) after Claim 1 The algorithm (16) is set up on the basis of the user input (22) and depending on the user profile (28) of the user (18) according to a predetermined coding rule (52) a new entry (32) in the database (14) to generate and assign at least one category and / or user group. Input system (10) according to one of the preceding claims, wherein the database (14) comprises a list of abbreviations (40), wherein in the list of abbreviations (40) at least one respective abbreviation (42) and at least one respective full word (46) are assigned to one another as data record parts and form a respective data record. Input system (10) after Claim 3 , wherein a respective data record of the database (14) is available in several languages, each data record part of a data record available in several languages being permanently linked to a respective language feature (44, 48, 54) indicating the respective language, the search rule (26) is set up to generate, in response to the user input of a first data record part of a data record in several languages in combination with a first language feature (44), an output list (24) which contains all further data record parts of the same data record assigned to the entered first data record part and all according to one The assignment rule comprises translations assigned to the data record in combination with their respective language features (48, 54), the search rule also being set up to give the respective translation the first part of the data record assigned to the respective translation in combination with the respective language feature (48, 54) of the translation output assigned. Input system (10) according to one of the preceding claims, wherein the database (14) is a global database which is set up to be fed by a local database. Input system (10) according to one of the preceding claims, comprising a user management module with user groups (38), each of the respective users (18) being assigned to at least one user group (38) of the user management module. Input system (10) according to one of the preceding claims, comprising a quality management module which is set up to correct and / or delete a data record of the database (14) if at least a predetermined number of users (18) marks the data record as faulty. Input system (10) according to one of the preceding claims, comprising a graphical user interface which is set up to record a user input (22) and to display an output list (24). Method for ensuring uniform conceptual conventions for terms in an application (20) accessible to several users (18) using an input system (10), the input system (10) having a filter (12) which has a database (14) and an algorithm (16), wherein the database (14) receives (S1) a user input (22) for which the uniformity of the terms is to be ensured and which is intended for the application (20) and on the basis of the user input (22) Output list (24) based on a predetermined search rule (26) generates (S2), and wherein the algorithm (16) based on a user profile (28) assigned to the user (18) by means of a predetermined Weighting rule (30) carries out a weighting of an entry (32) of the output list (24) and sorts the output list (24) according to the weighting (S3), the sorted output list (24) provides the user (18) (S4), one Selection by the user (18) receives and depending on the selection from the user input (22) generates a revised user input (34) (S5) and passes it on to the application (20) (S6), the algorithm (16) continuing based on the user profile (28) of the user (18), the user input (22) of the user (18) and the selection (32) made by the user (18) of the output list, a learning data record (36) is generated (S7), on the basis of which he adjusts the weighting rule (30) according to a predetermined training method (S8), and the learning data record (36) generated for the respective user (18) of a user group (38) and / or the adjusted weighting rule (30) for all users (18 ) the user group (38) uses (S9).

Images