DE10007477A1 - Computer network system for experts database management in organization, sends queries stating problem repeatedly to group of receivers who recommend experts - Google Patents

Abstract

A server (S) receives query stating the problem from a user and transmits the query to receivers (E1). The server analyzes the response from the receivers who recommend experts for the problem and transmits modified query to the recommended experts. The server receives the experts responses and stores it as competence profile in experts database. Independent claims are also included for the following: (a) Experts database management method; (b) Experts database management program

Description

The present invention relates to a computer network system stem, a process for operating an expert database on a computer network system with a central unit and a majority of those in the Zen network traleinheit arranged workstations and a Compu Program for performing such a method.

A generic expert database that also includes skill Database called is used for structured and retrieval deposit of information and data People with special knowledge and skills. Derar Current databases are preferably found in larger organizations tion such as companies, authorities and the like.

The organizations addressed reach today in many cases a size (and specialization) at which it the special knowledge of individual employees becomes difficult to locate and identify. Especially at spatially distributed, decentralized companies there are often special skills within the organization in other parts of the organization need to be. However, an employee has no knowledge about the skills of his colleagues, they can do it themselves are not localized and used if necessary, so that potential achievements have not been achieved or innovations have been neglected  or services are started on the market buys that were also available in their own company would be. For these reasons, an efficient expert Database essential.

As already mentioned, these expert databases are used for Storage of knowledge and skills of employees an organization to find experts facilitate. The storage of personal data is based on so-called know-how profiles (hereinafter referred to as competence profiles), which the other Mit be made available to workers. From this da an employee can use this database according to the ver available, in-house skills ("Skills") search and a colleague with the required Find skills. The competency profiles must, however previously created, maintained and maintained by the employees themselves to be edited. The known systems are simpler Nature, d. H. a simple, relational database or else provide a document database that can be researched in full text for this purpose in itself sufficient technical solution gene.

In practice, however, these systems only fulfill exceptionally do their purpose. It is particularly problematic that the maintenance and consequently the timeliness of the data often is not guaranteed, since it is in the care of the com competence profiles for the individual employee by only one difficult to mediate activity, since this of the Usually disclose his personal knowledge has no personal advantage (or at least not in it recognizes an advantage). The efficiency of the database fails thus on the question of motivation. Add to that the difficulty  that simply stating a particular skill uses little, if not can also be specified to what degree the ability is mastered.

In addition to these difficulties, the (efficient) use make the database seem very unlikely, there is also the problem that the introduction of one of the like system is relatively expensive. Because if everybody apply his knowledge and skills in free form for then there is a rediscoverability of the competence profiles not given. On the other hand, fixed terms being able to give is a thorough analysis of the in the Un used terminology and areas of expertise mandatory. This task poses in larger sub however make a significant investment.

In contrast, according to the invention, a computer network stem according to claim 1, a method with the features of Claim 3 and a computer program or Computerpro Gram product with the features of claim 8 proposed The invention is based on a cooperative, decade central approach based on the existing motivation of the users who starts.

According to the invention, a request from a user (An questioning) for an expert with a certain knowledge descriptive terms comprehensive request profile of the central unit automatically sends a competence request to ei or more in connection with the knowledge requested nissen recorded in the database and / or one or more re randomly selected recipients. In your Answering the competence request to the central unit can the recipient yourself and / or another  Name users as experts. Depending on the received The central unit provides answers to experts identified the inquirer with and / or sends further competence inquiries (recommended in answers received) as experts other recipients and / or to one or more arbitrary selected other recipients. Get the data Experts are fil saved in a storage unit of the central unit chert. This procedure, that is, the request from the central office by sending out competence inquiries, is communicated via a predetermined time period and / or repeated as many times up to at least one expert reliably determined and at the request could be communicated.

The invention thus comprises two states, on the one hand the zu stated that a sought - after competence profile is known, and on the other hand the state that the competence profile sought is unknown, so the system is in a "learning phase" located. The state mentioned first (searched compe tenzprofil is known), however, is hereinafter referred to as Designated "productive phase".

In an embodiment of the invention, the recipients in ih Responding to the central unit suggestions for a Kate gorization of the descriptions describing the knowledge sought Make handles and there is a categorized rejection Save the expert data in the Zen storage unit tral unit.

In a further embodiment of the invention, terms that describe certain knowledge and skills, Be train indicators assigned.  

In a particularly advantageous embodiment of the invention Terms that are mentioned in context are similar terms of weighting similarity indicators assigned to the gate.

According to the invention, the inquirer is advantageously several experts identified by the central unit in one Ranking according to the frequency of their mentions given.

The invention takes advantage of the fact that users Although it is difficult to motivate, your own knowledge to be stored in a system because of personal benefit this activity is estimated to be low, but for the documentation of the knowledge of other employees there is quite a high level of motivation if the user himself seeks such knowledge and skills. The second we The substantial knowledge on which the invention is based is The fact that the informal relations between the Col usually, the users of the computer network system represent the key knowledge that the employees need for their daily work. An employee knows Not a contact person for every technical question, but knows after some time of service, which one Colleague may have a suitably competent contact partner knows.

The invention comprises a computer network system with a Central unit, with a plurality of in the network the central unit arranged work stations and one Expert database in which a variety of competency pro filing one with reference to the data or several key words representing an expert  can be stored, a method for operating an Ex perten database on such a computer network system and a computer program or a computer program pro product, the appropriate computer code means of a corresponding Programming language for performing such a procedure rens when running the computer program or computer pro gram product on a computer.

Further advantages and refinements of the invention result itself from the description and the accompanying drawing.

It is understood that the above and the Features to be explained below not only in the combination given in each case, but also in others Combinations or alone can be used without to leave the scope of the present invention.

The invention is based on an embodiment in the Drawing is shown schematically and is un below ter described in detail with reference to the drawing.

Fig. 1 illustrates a schematic representation of the receipt of a request to the central unit and its competence request to a plurality of Emp recipients (learning phase).

Figure 2 illustrates the categorization of descriptive terms.

Fig. 3 illustrates the receipt of positive feedback conditions at the central unit.

Fig. 4 illustrates the assignment of competency profi len, which are stored in the central unit personal.

Fig. 5 illustrates the operation of the inventive method in the productive phase.

Fig. 6 shows a flowchart of the sequence of the inventive method when receiving a new request email M1.

Fig. 7 shows a flowchart which illustrates the sequence of the inventive method when responding to a search mail.

Fig. 8 shows a flowchart which illustrates the sequence of the method according to the invention when accessing the database via a dialog interface.

A user A of a computer network according to the invention stems is looking for a special problem with which he is facing, in that shown in the figures Example an expert in the field of glass coating. Around To obtain information according to the invention, the user writes an electronic message M1 (hereinafter "request email" in which he describes the specific problem, and sends this request email M1 as requestor A to one Central unit of the computer network system, the server S. The server S reallocates a unique request to the request mer, which for the entire time of processing the Fal les stays the same. He also takes over the request Letter from the requester A.  

In the learning phase, server S does not yet have In information about the target group and their knowledge se, d. H. there are no competency profiles yet.

For this reason, the server S automatically generates an email M2 to randomly (arbitrarily) selected recipients (recipient group E1 of FIG. 1). The recipient group E1 is selected from a list of names and addresses stored in the system ("mailing addresses"). The number of recipients E1 of such a competence request can be specified or set in the system according to the invention.

The mail M2 is drafted by the server S so that it is the Pro specified by the request email M1 of the requestor A. reproduces visuals. Preferably, that of the request A in question M1 formulated in his request email from System taken over and (as competence request M2) sends. The system relies on the original request Lich added a questionnaire, which the arbitrarily selected recipients E1 of mail M2 must be completed. With the In this questionnaire, recipients E1 are asked whether they are even as an expert on the one reflected in the request Recognize the problem if you can recommend someone who can help to solve this problem (although this be both a possible expert but also a person can, who in turn knows or can know an expert te), and what terms in the opinion of the Mail M2 answering recipient E1 formulated in the request can be assigned to the specific problem.  

If none of the recipients E1 is a sought-after expert, then part Share this with the server S. If a recipient (in within a given time) does not depend on the competence ask, there are several ways how the system points to it responds: Either the system immediately assumes that the recipient in question is not an expert and also not an expert potential experts, or the system is like that set that it is assumed that the recipient ger (due to business trip, vacation, illness or son absence) to answer the competence question was hindered within the given time, and resends the competence request to the recipient in question or a corresponding reminder.

In the response mail to the server S of those recipients who are not themselves experts on the topic sought, they share potential experts and / or categorization terms for the problems described in the competence profile, if possible. The latter is illustrated by way of example in FIG. 2. In the responses of the recipients E1 to the server S in FIG. 2, the keywords “coating technology” and “process engineering” are mentioned as classification terms for the query problem “glass coating”.

On the basis of the information received, the server S sends competence inquiries in the form of a new mail M2 to the now named potential experts (recipient E2), and this procedure is repeated until at least one expert has been found who is explicitly opposed to the server S. as an expert on the topic you're looking for. If the server S is not informed of a potential expert, it sends new competence inquiries to a number of arbitrarily selected recipients, so that the inquiry cycle begins again at the first level (recipient group 1 ).

Fig. 3, the inventive method can be executed at the competence requests for the snowball system illustrated. The illustration of FIG. 3 for simplicity sake is carried out by means of a branch or tree structure Zvi rule the receivers of each layer. In fact, however, communication does not take place directly between the individual recipients, but basically via server S.

At each level, designated by a recipient group E1 to E5, competence inquiries are sent out in the form of Mail M2. Even if there is no sought-after expert in a addressed recipient group, the query chain continues through the recommendation of potential experts, so that further recipient groups can be written to by the server S, the system being set / programmed so that no one receives the same query sent several times. In the example shown in FIG. 3, a recipient of the recipient group E3 calls himself an expert X1 by sending a mail MX1 back to the server S, in which he announces that he is an "expert in process engineering". Since the desired competence profile, i.e. the request originating from the inquirer A, but formulated in more detail, namely concerns the problem of glass coating, the request system according to the invention continues until a recipient from the recipient group E5 calls himself an expert X2 by calling in a mail MX2 to the server S that he is "an expert in process engineering and glass coating". Simultaneously with their self-designation as experts, users X1 and X2 assign terms in their reply mails MX1 and MX2, respectively, under which the request or problem of the requester A is to be categorized.

In addition, the competence profile of the two experts X1 and X2, which is now known to the server S, is categorized as a competence profile P1 and P2 in a storage unit (not shown in more detail) and assigned to the experts X1 and X2 in a personal manner (see FIG. 4).

Finally, the requestor A gets the result, that is, the Names and contact details of the identified experts X1 and X2, communicated.

During the learning phase, those in the database are also saved stored competence profiles of the experts with the selected terms added so that already spewed competency profiles always updated and new ones most of the knowledge. To do this, the individual terms in itself from the database technology be known with reference indicators.

Rankings (so-called scoring lists) are also advantageously created by the system according to the invention:

• 1. Be related to the problem requested whose categorization several terms in the in the Server-managed collection of terms entered or removed selected from the list of terms, so these terms referred to as related or similar. Be handles often called together, so these pairs have a stronger similarity than pairs of terms that rarely  to be named. For this purpose the terms provided with so-called similarity indicators.
• 2. A first recipient became more frequent than a second Recipient to a specific term as (potential ler) expert reported, the first recipient is in in order of precedence before the second recipient.

These rankings can be used individually or together the inquirer with the transmission of the expert data to be shared.

So far, the system and method according to the invention has been described on the basis of the learning phase, that is, on the assumption that no expert data are yet available in the system or are at least not yet available for the problem requested. As soon as data is sufficiently available in the database, an inquiry according to the invention can be processed and answered in the so-called productive phase, which is described below with reference to FIG. 5. In practice, there is often a mixed operation of the learning phase and the productive phase. In particular, the learning phase flows smoothly into the productive phase.

The productive phase differs from the learning phase there through that after sending the mail M1 of the requestor A the server S in advance based on that in the database system existing and filed competence profiles or several experts X1 to X5 propose, whose competency profiles P1 to P5 meet the request. In addition, and / or similarity statements experts suggested become.  

These experts are also sent via email M2 wrote and can now expressly become an expert in name this problem. The further procedure is now as described in the learning phase (which the system as sol The accepted expert can now confirm himself as an expert gene and / or another person as a potential expert name, etc.).

The list of terms mentioned above is used to form egg a categorization of knowledge and skills related inquiries. In addition, the Ka tegorization with the terms also relatives drawing (or similarity) of the individual terms to use it for further inquiries.

The stored competence profile according to the invention contains among other things those terms that the query problem categorize and be the specialist areas of the person write. In addition, the expert data can also be used or other data are maintained by the person himself.

The ranking for experts supports the Nen of experts through the system. It takes advantage of that an employee "reluctantly" calls himself an expert. The Ranking for relatives / similarities the terms are used when finding an expert, if a term has been assigned that is related to or similar to another term.

If no expert is found after a specified time or no employee has named himself as an expert, so the system randomly selects additional employees to help them  to send an email M2, with which the request zy klus starts again. The number of repetitions of the Cycle can also be specified, or the cycle is repeated until at least one expert is found has been.

Fig. 6 shows a flow chart of the sequence of the inventive method when receiving a new request mail M1.

At 10 , the request email M1 is received by the system, ie in the server S. Experts are then searched for in the left branch of the flowchart at 20 on the basis of keywords contained in the request email in the existing database. The extraction of the keywords on the basis of which the search is carried out from the text of the request is advantageously carried out using the so-called stop word file. If experts are found in the database on the requested problem, these experts will be sent an inquiry form (questionnaire), which is generated at 22.

Regardless of this, according to the right branch of the flowchart at 30, a new request is generated based on the request email, which is sent to randomly selected recipients E1 (at 32). The inquiries are sent by email to both the experts identified at 20, 22 and the recipients selected at 32 at 12. If no experts are identified within a predetermined or adjustable period 14 , the requesting party A is informed ( at 16) and recipients are randomly selected and written again.

The system can, for example, be set such that Inquirer A should return the result of the reported experts are not satisfied, can cause that a new request is generated, whereby from this poss possibly only a limited number of times len use can be made.

Fig. 7 shows a flowchart which illustrates the sequence of the inventive method when responding to a search email.

At 40 , the recipient of the search mail M2 replies to the form (questionnaire) sent with this mail that contains the three answer fields "bin expert", "know experts" and "category". The answers to these three points are treated separately by the system. If the responding recipient has assigned one or more terms / keywords to the query problem, then these are included in 42 in the network of similar terms (synonym network). If the answering recipient names a potential expert in his reply, then the expert's request for competence is sent to the expert at 44.

If the answering recipient calls himself an expert, so he is included in the expert list at 46. The Ab The experts found in this way are saved under Assignment to that of the expert in his answer himself keywords given regarding the requested pro stupid. Inquirer A receives a message at 48 with which communicated to him the name of the expert found becomes.  

Fig. 8 shows a flow chart which illustrates the sequence of the inventive method when accessing the database via a dialog interface.

The system according to the invention can be set up in such a way that the request of A requester A is not just via email is possible, but that the requesting party also has a slide log-oriented interface directly to the internal data basis can be accessed by selecting a term or enters a keyword for which the system provides information about the available experts. The display takes place for example by specifying a list of similar items handles and all experts who agree on these terms stores are. With direct access to the database in contrast to the E-mail request no new data in the database testifies, but only queried existing data.

When accessing via a dialog interface, the user (inquirer) enters a term at 50. He can either type this term himself at his workstation or select it from a list of available terms. At 52 , the system according to the invention shows the user the terms similar to the entered term and any experts. The expert data is displayed on onymized so that the user does not yet know who the displayed expert is. If the user is not satisfied with the selection result proposed to him at the time, the expert can either choose another term ( 54 ) or he can send a freely formulated query to the system, which is then sent as a competence query M2 to randomly selected recipients is sent ( 56 ). If there are one or more experts in the list displayed at 52 for the keyword of interest to the inquirer, the system according to the invention, after confirmation by the inquirer, generates an inquiry ( 58 ) and sends this to the selected expert (the one for the Requester not yet identified at this point).

All terms in the learning phase each new Request entered will already be in the above mentioned synonym network inserted. For this, the terms that are assigned to a request without weighting with connected to each other. Advantageously has an additional Administrator the opportunity to expand the network add or no longer required terms / edges / nodes to remove from the network. After the learning phase, the Weights of the edges connecting the individual terms further modified to verbes the results of requests ser. An increase in the edge weighting can ins special based on sharing of terms by different users, by common Mentioned by various experts (in the learning phase) and / or by subsequent assignment by different Experts (in the usage phase) are done. A weakening the weighting takes place particularly through aging.

The described invention is of course not based on limited use in a (large) company. Also in others organizations such as authorities, law firms etc. or between individual organizations in cooperation with each other or in all types of intra- and internetworks the invention can be used.

Claims (9)

1. Computer network system with a central unit (S), a plurality of in the network with the Zentralein unit (S) arranged work stations and an expert Database in which a variety of competency profiles under Allocation to the data related to one or more Save keywords as a person representing an expert are bar, the central unit (S) on a Workstation input request (M1) from a user (A) for an expert with a certain knowledge writing terms comprehensive competence profile automa table a competence request (M2) to one or more in Relation to the requested knowledge in the data bank and / or one or more arbitrarily selected recipient (E1) sends out in their reply to the central unit (S) itself and / or another Name users as experts and the central unit (S) depending on the replies received from the recipients (E1) identified experts (X1, X2) on the inquirer (A) shares and / or further competence inquiries to as experts recommended additional recipients (E2, E3) and / or one or several arbitrarily selected additional recipients (E2, E3, . . .) sends out and the data of determined experts together menhang with their competence profile (P1, P2) in one room saves the central unit (S).   2. Computer network system according to claim 1, in which a Repetition of requests (M2) by the central unit (S) over a predetermined period and / or until at least an expert (X1, X2) determined and the inquirer (A) with could be shared. 3. Method for operating an expert database on a computer network system with a central unit (S) and a plurality of workstations arranged in the network with the central unit (S), with the following steps:

• - on a request (M1) from a user (A) (requests from) for an expert with a certain descriptive terms comprehensive competence profile automatically sending a competence request (M2) of the central unit to one or more in connection with the requested knowledge and / or one or more arbitrarily selected recipients (E1) recorded in the database, the recipients (E1) being able to name themselves and / or another user as experts in their response to the central unit (S),
• - Depending on the replies received from the recipients (E1) to the central unit, notifying experts (X1, X2) to the inquirer (A) and / or sending further competence inquiries to other recipients recommended as experts (E2, E3) and / or one or several arbitrarily selected additional recipients (E2, E3...),
• - Saving the data of the experts determined in connection with their competence profile (P1, P2) in a storage unit of the central unit (S),
• - Repetition of the inquiries (M2) by the central unit (S) over a predetermined period of time and / or until at least one expert (X1, X2) has been determined and communicated to the inquirer (A).

4. The method of claim 3, wherein the recipient in your response to the central unit suggestions for a Ka tegorization of the descriptive knowledge Can make terms and a categorized saving the expert data in the storage unit of the central processing unit he follows. 5. The method of claim 3 or 4, in which certain Knowledge descriptive terms reference indicators be classified. 6. The method of claim 5, in the terms that be mentioned several times in context, the the similarity of the Terms of weighting similarity indicator will be shown. 7. The method according to any one of claims 3 to 6, in which several experts identified in a ranking accordingly the frequency of their mentions from the central unit are given. 8. Computer program with computer program code means which are suitable for a method according to one of the claims 3 to 7 perform when the program on a computer expires. 9. Computer program according to claim 8, which on a compu readable medium is deposited.