KNOWLEDGE BASE VERIFICATION AND CORRECTION SYSTEM AND METHOD THEREOF

Hereinafter, with reference to the drawing objects of the present invention embodiment for example detailed as follows. Examples of the present invention embodiment to the present invention average per industry knowledge with entire surface account for which ball number are disclosed. The present invention refers to various modification may have bar can apply in various forms, a specific embodiment and examples detailed drawing example broadcast receiver. However, the present invention is defined with respect to a particular disclosure form be but is, all changing range of idea and techniques of the present invention, including the water to replacement should understood to evenly. Each drawing while similar references in a similar components are described used for substrate. In the attached drawing, the dimensions of the structures of the present invention can be higher than the virtualization clarity shown is reduced by reducing the thread number or are disclosed.

The term used in application specific embodiment is only used to account for example, defining the present invention intending to be is endured. It is apparent that a single representation of the differently in order not providing language translators, comprising a plurality of representation. In the application, the term "comprising" or "having disclosed" specification of articles feature, number, step, operation, component, piece or specify a combination not present included, another aspect of one or more moveable number, step, operation, component, piece or a combination of pre-times the number should not understood to presence or additional possibility.

Not defined differently, scientific or technical terms so that all terms in the present invention thus is provided to the person with skill in the art is a device generally by the same meanings. Generally dictionary used for providing language translators such as defined terms have the meanings associated technology must be consistent semantics and having interprets, the application manifest in defined present for, or overly formal sense interpreted not ideal.

Hereinafter drawing and description, one block can be hardware blocks or software block display or described components. For example, each of the components may be an independent hardware block as receiving signals between each other, or one processor software running in a block may be filled.

Figure 1 shows a knowledge base of the present invention also in the embodiment according to exemplary calibration system (100) block are disclosed.

The collected data extracted from data knowledge based on established the knowledge base, a biological evaluation content on the other hand, incomplete knowledge data, e.g. formulating can be knowledge data contains an error. For example, constructing knowledge base of data to be based, database (e.g., wick producer Oh (Wikipedia)), such as CSV file may be shaping data, news, blog, SNS, document file data such as non-disapproval. Shaping data when an unverified knowledge base, knowledge base schema mapping rules created by the expert understand by the knowledge base can be built relatively for hereinafter. On the other hand, when non-knowledge base from an unverified, morphological analysis, including using natural language processing techniques like parsed to extract the non-structured knowledge from data is required for the job, the limits non-natural language processing techniques such as the reliability of data can be an error due to the knowledge base.

Internet data collected through evaluation amount of non-constructed from incomplete knowledge data included in the knowledge base in order to correct, a few designated expert knowledge data to detect and correct the incomplete knowledge base included in a mobile communication can be impossible. Incomplete knowledge data included in the knowledge base of the calibration may be continued as a knowledge base for heat emission efficiency of the requisite tasks, detecting incomplete knowledge data included in the present invention refers to a knowledge base, the system and method can correct for hereinafter grow the wood sourcing through number under public affairs substrate.

The reference also 1, a knowledge base calibration system (100) is a knowledge base (200) and web service system (300) and communicate with, user terminals (500) web service system (300) can be in communication with. A knowledge base calibration system (100), a knowledge base (200), web service system (300) and user terminals (500) LAN (local area network) is connected to the network (wide area network) and WAN may communicate with each other such as by, dedicated channel communicating through one-to-one communication between quantization of disapproval. As shown in fig. 1, knowledge base calibration system (100) is a knowledge base interface (102), user interface (103), query generation unit (110), question selecting unit (120), a user information storage unit (130), candidate knowledge generation unit (140), candidate knowledge store (150), candidate knowledge section (160), user level analysis unit (170) can be a.

A knowledge base interface (102) includes a knowledge base calibration system (100) to other components contained other knowledge base (200) can be under public affairs interface to access number. For example, as which carry, query generation unit (110) and candidate knowledge section (160) includes a knowledge base interface (102) through a knowledge base (200) by which the user accesses, a knowledge base (200) may be stored in knowledge data receiving, knowledge data corrected knowledge base (200) to send the disapproval.

Similarly, user interface (103) is knowledge base calibration system (100) to other components included in the other web service system (300) can be under public affairs accesses for interface number. For example, as which carry, question selecting portion (120) and candidate knowledge generation unit (140) includes a user interface (103) through web service system (300) by the user accesses the network, may be transmitting question, receiving an answer to the questions may be filled.

Query generation unit (110) includes a knowledge base (200) from incomplete knowledge data, knowledge data and knowledge data string including errors can be detected, the correct answer can be produce incomplete knowledge data interrogation. For example, query generation unit (110) is to make the increment can be knowledge or missing knowledge detecting, for correcting the same content can answer a query generating. Some of the plurality of users asking questions generated 1308. ball number. Query generation unit (110) also carry 2 and 3 for details of the new reference also will be disclosed.

A user information storage unit (120) comprises a plurality of users can be storing level. The user's answer to the question on a user level (or user level) diagram (e.g., the number of answer, such as period) and answer the reliability of user evaluations according to corresponding scores can be. As which carry, user level is query generation unit (110) to determine whether any user asking questions generated by each have been assigned based on can be used. In addition, a user information storage part (120) can be stores identification information of the user, user identification information of the web service system (300) external system (e.g., reward system of Figure 5 (400)) such as synchronizing user can be used.

Question selecting unit (130) includes a user level of information, i.e. user level information can be assigned questions and question determines the number of users. For example, question selecting unit (130) includes a high user level relative to a user asking questions can be a larger number of units, high difficulty can be allocating a question (e.g., question click thereon). The, without considering the level of a user to enter responses to user number generation public box door number points, e.g. answer reliability of price, such as answer of delay can be resolved. Question selecting unit (130) includes a user interface (103) through the outside of the at least one knowledge base calibration system of predetermined questions can be, at this time a query is allocated can be output with user information.

Candidate knowledge generation unit (140) is question selecting portion (130) including an answer to the question and its questions selected by at least one candidate knowledge data can be generated. In one in the embodiment, candidate knowledge generation unit (140) is question selecting part (130) can be predetermined from upon receiving a, user interface (103) can be predetermined through receiving an answer to the query. In other in the embodiment, candidate knowledge generation unit (140) includes a user interface (103) to receive both an answer to the question and through and may be filled. Candidate knowledge data knowledge base (200) is comprised of knowledge data be stored, which carry as, candidate knowledge section (160) when upon successful verification by knowledge base (200) knowledge of calibration data can be used.

Candidate knowledge store (150) includes a candidate knowledge generation unit (140) can be generated by storing candidate knowledge data. As aforementioned, candidate knowledge generation unit (140) is capable of producing answer question and corresponding candidate knowledge data and including, candidate knowledge store (150) comprises a plurality of questions and answer to a candidate knowledge generation unit (140) can be generated by storing candidate knowledge data. As shown in fig. 1, candidate knowledge store (150) includes a candidate knowledge generation unit (140) and capable of receiving a candidate knowledge data, candidate knowledge section (160) and user level analysis unit (170) under public affairs knowledge data number can be a candidate.

Candidate knowledge section (160) have the same query corresponding to the at least one candidate knowledge data can be verifying candidate based on knowledge data. Candidate knowledge section (160) is a candidate upon successful verification based on knowledge data knowledge base interface (102) through knowledge data stored in a knowledge base can be correct. With reference to the 4 which also carry as, candidate knowledge section (160) have the same number from a plurality of user questions can be verifying knowledge data by evaluating candidate presents answers. Candidate knowledge section (160) for details of the new also refers to carry 4 will be described.

User level analysis unit (170) and candidate candidate knowledge is knowledge data section (160) based on the result of verification on the candidate knowledge data, user information storage unit for storing (120) updating user level information can be stored. For example, user level analysis unit (170) is candidate knowledge section (160) knowledge data based on the result of verification on the candidate, upon successful verification answer user level performance of a user information storage unit for storing a number adjusted upwards (120) updating user level information can be stored. In addition, user level analysis unit (170) is of constant period (e.g., 1 main, such as 1 month) of a user level information can be updating. I.e., the performance of a user in the predetermined cycle number based on the number of answer, the number of answer can be less than a predetermined reference value when the user level LCP alignment, the user level can be adjusted upwards or other.

In one example of Figure 1, a user information storage unit (120) and candidate knowledge store (150) includes a knowledge base calibration system (100) being accessed by components is is also shown. In one in the embodiment, a user information storage unit (120) and candidate knowledge store (150) and the center of an access processing adapters may each include, a knowledge base calibration system (100) a user information storage unit for storing components of the adapters (120) and candidate knowledge store (150) or storing data thereof can read data stored. In other in the embodiment, a user information storage unit (120) and candidate knowledge store (150) to access components are a user information storage unit (120) and candidate knowledge store (150) may each include a adapters to access disapproval.

Although also Figure 1 shows a knowledge base calibration system (100) the knowledge base (200) accesses the shown example without intending to be limited, in the embodiment of the present invention according to an exemplary knowledge base calibration system (100) is a knowledge base (200) may include, in this case a knowledge base calibration system (100) the knowledge base function, i.e. number of knowledge data can be co service. In addition, also Figure 1 shows a knowledge base calibration system (100) web service system (300) through user terminals (500) is shown in communication with a real example, in the embodiment of the present invention according to an exemplary knowledge base calibration system (100) comprises a user interface (103) through user terminals (500) can be communicate directly with. In addition, knowledge base calibration system (100) includes a plurality of knowledge base and a plurality of web service system that communicates with the disapproval.

In the embodiment of Figure 1 also according of the present invention Figure 2 shows a exemplary question generation unit (110) example (110') representing block degrees and, Figure 3 representing the incomplete knowledge data example are disclosed. With reference to 1 also as aforementioned, query generation unit (110') of Figure 1 includes a knowledge base (200) can be incomplete knowledge data from detecting, answer incomplete knowledge data retrieval in the interrogation can be generate.

The reference 2 also, query generation unit (110') is knowledge data (10) can be receiving. For example, query generation unit (110') is a knowledge base (200) for detecting incomplete knowledge data from the knowledge base (200) can be stored in knowledge data scan, knowledge data (10) can be receiving. Knowledge base of Figure 1 (200) knowledge data stored in the ontology data, such as RDF (Resource Description Framework) format such as the thin film may be, can be generated by RDF schema comprising one instance. The reference also 3, a human classification as instances' will have a ' may have a plurality of attributes. For example, as shown in fig. 3, 'will have a' person 'born' is provided, 'nationality', 'gametes', 'sex', 'belongs' may have attributes such as, instance of' will have a ' values (e.g., 1994, such as a compensation) includes for each attributes may have disclosed.

The reference 2 also, query generation unit (110') is attribute manager (112), the error detecting section (114) and question output (116) can be a. Attribute manager (112) is classification (e.g., human, tissue, geographic, event or) constituting knowledge data according stores information relating to the attributes of the can, the error detecting section (114) can be attributed under public affairs number information. Also with reference to the example 3, attribute manager (112) is' born ',' nationality ',' gametes ',' sex ',' belongs' for a person can be stored as properties, the error detecting section (114) can be under public affairs to number. In addition, attribute manager (112) is number under public affairs surface of the format attribute information about attributes, such as number, URL, full text information with respect to the can.

An error detecting section (114) is incomplete knowledge data, for example knowledge data can be detecting knowledge data including error string. I.e., attribute manager (112) under public affairs number is based on the attribute information, knowledge data (10) can be whether the knowledge data is incomplete. For example, the error detecting section (114) comprises a plurality of values including detecting an attribute instance having knowledge data can be written. As shown in fig. 3, a knowledge base of Figure 1 (200) as knowledge stored in instance of 'will have a' the attribute value of the at least one of the values ('UN', 'foreign cylindrical part') 'belongs' 2 can be writing, the error detecting section (114) is' will have a - - UN belongs, diplomatic cylindrical part ' including a knowledge data can be detected. As another example, the error detecting section (114) is included in the value of the attribute (e.g., 'born') knowledge data type (e.g., 4 digit number) having different properties (e.g., a keyboard) when having type value, including error determines disapproval.

In addition, the error detecting section (114) is insufficient knowledge data, i.e. value including an attribute instance can be detecting missing having knowledge data. As shown in fig. 3, a knowledge base of Figure 1 (200) as knowledge stored in instance of 'will have a' missing 'gametes' can be a value of the attribute, the error detecting section (114) is' - will have a female gametes - empty ' including detecting a knowledge data can be insufficient.

Question output (116) is an error detecting section (114) transfers the query to knowledge data detected by means (20) can be generate. For example, question output (116) comprises a plurality of values in the case of knowledge data written attribute includes instances including, instances, a plurality of values is written attribute, including generating a plurality of values can be thoughts or questions. In addition, query output (116) including in the case of knowledge data having an attribute instance value is missing, instances, create one click thereon including an attribute value missing question may be filled. The reference also 3, question output (116) is' will have a - - UN belongs, foreign cylindrical part ' including a knowledge data, as shown in fig. 4, " will have a door belongs is? (1) (2) (3) other cylindrical part foreign UN " including generating a question can be. In addition, query output (116) is '- will have a female gametes - empty' including a knowledge data, " the gametes will have a door?" Including generating a question can be. With reference to 1 also as aforementioned, question output (116) to generated by the interrogation, question selecting portion of Figure 1 (120) by a user can be assigned disclosed.

In the embodiment of the present invention also Figure 4 shows a section of Figure 1 according to an exemplary candidate knowledge (160) to explain the operation of the are disclosed. Specifically, Figure 4 candidate knowledge section (160) candidate knowledge store (150) exhibits example candidate knowledge data received from. With reference to 1 also as aforementioned, candidate knowledge section (160) includes a question and answer can verify the candidate including knowledge data and, based on the successful candidate knowledge data verifying knowledge base (200) for correcting the knowledge data can be stored.

The reference also 3 and 4 also, 'will have a - - UN belongs, foreign cylindrical part' including a knowledge data in order to correct, " will have a door belongs is?" Of Figure 1 that the challenge question generating unit (110) is produced by, question selecting portion (120) by 5 name assigned to users (lighter, a, bottle, positive, free) and, 5 users name (lighter, a, bottle, positive, free) by receiving the answer from candidate knowledge generation unit (140) also by 4 such as shown in candidate knowledge data can be generated. In one example shown also 4, candidate knowledge data can be user and user level information.

Candidate knowledge section (160) have the same number the same answer survey questions based on the number of candidate knowledge data can be verifying the performance of a user. I.e., candidate knowledge section (160) have the same answer number the performance of a number of users is the predetermined ratio or more, including such answer can be determining candidate knowledge data upon successful verification. For example, in one example shown also 4, candidate knowledge section (160) is '(1)' the reply to users (lighter, a, bottle, positive) the proportion of the number of and answer overall user (4/5) predetermined ratio (2/3) the capacitor from, 'will have a - - UN attribute, diplomatic cylindrical part' including a knowledge data ' will have a - attribute - UN'can be correct to include.

In addition, candidate knowledge section (160) have the same number of users as well as the performance of a number of users answer level information for verifying an identity further based upon candidate knowledge data may be filled. I.e., the user's answer on high level by adding weight to improve the reliability of the verification can be candidate knowledge data. For example, in one example of Figure 4, candidate knowledge section (160) is '(1)' sum and the level of the users (lighter, a, bottle, positive) the reply '(2)' (free) can be calculated sum level of a user answer, at the highest level there is a predetermined ratio if it exceeds a sum ratio, answer candidate knowledge data corresponding to such level sum including determining upon successful verification can.

Candidate knowledge section (160) is there is verification of candidate knowledge data, e.g. the same answer number is a predetermined ratio or foreseeable sum of the performance of a number of users is less than the highest level or when there is a predetermined ratio ratio less than, verification of candidate knowledge data fails can be Walsh. Candidate knowledge section (160) failed verification information about a question by candidate knowledge data selecting unit (130) to 1308. ball number. Question selecting unit (130) includes a verification query time failed candidate knowledge data corresponding to predetermined user with different users can be assigned that determines. The, failed programming verification data based on predetermined user presents a number from different user with candidate knowledge time answer candidate knowledge section (160) can be keyholder verification by.

Figure 5 shows a knowledge base calibration system of the present invention also in the embodiment according to exemplary (100a) block are disclosed. Knowledge base calibration system of Figure 1 (100) as compared to, knowledge base calibration system of Figure 5 (100a) is reward system (400) for communicating with reward interface (104a) can be further comprises. The description hereinafter of Figure 5, described with reference to and overlaps a content related to the first 1 also are disclosed.

Reward interface (104a) is knowledge base verification system (100a) external reward system (400) the user information storage unit for storing (120a) can be under public affairs interface to access number. For example, as which carry, user information storage unit for storing (120a) level information stored in the user reward system (400) can be utilized by, reward interface (104a) is reward system (400) including a plurality of reward systems interface number under public affairsunder public affairs number can be user level information.

Reward system (400) is a knowledge base verification system (100a) (or user information storage unit for storing (120)) of user level information can be under public affairs number a reward to a user based. For example, reward system (400) answer the performance of a number corresponding to the number of the user for user reward may be under public affairs, user level periodically by checking user periodically to a user corresponding to number under public affairs reward can be. Reward system (400) is beneficial for use under public affairs reward the user number comprised of, a non-number is by way of example, coupon, cash, member stores a point or the like can be used.

A knowledge base verification system (100a) presents a number answer questions number may be obtained from the performance of a user and reward, the user can be a pair of syringes for the answer on among. With reference to 1 also as aforementioned, a user information storage unit (120a) stored user level information is user level analysis unit (170a) can be updated by, user level information so that the user answer by the calculation based on reliability of the state diagram and answer, higher efficiency and reliability under public affairs number knowledge base (200) can be realized for verifying.

In the embodiment according to Figure 6 shows a knowledge base of the present invention representing an exemplary method for calibrating order also are disclosed. Specifically, Figure 6 shows a knowledge base (200) number to the user for conditions indicative of Figure 1 also order generating method which ball question are disclosed. Figure 6 shows a 1 and 2 also refers to be described hereinafter in also are disclosed. As shown in fig. 6, a knowledge base (200) (S12, S14, S16, S18) calibration method comprising the multiple stages can be.

In step S12, a knowledge base (200) from detecting incomplete knowledge data can be performed. For example, question generation unit of Figure 2 (110') included in an error detecting section (114) is attribute manager (112) is number under public affairs knowledge data based on the attribute information (10) is incomplete knowledge data, e.g. whether the knowledge data knowledge data including error string can be, can be detecting incomplete knowledge data.

In step S14, incomplete knowledge data generating calibration of a query can be performed. For example, question generation unit of Figure 2 (110') included in question output (116) is an error detecting section (114) can be detected by means generates questions from incomplete knowledge data. As aforementioned, including a plurality of selectable answers questions are generated the interrogation may be thoughts, questions are click thereon disapproval.

In step S16, the query to level for selecting and question number of users can be performed. For example, question of Figure 1 selecting unit (120) includes a query generation unit (110) of a plurality of queries generated by against, a user information storage unit (130) based on information stored in the user level, the number of questions assigned user can be assigned a user and determining questions. Preferably similar, question selecting unit (120) includes a user information storage unit for storing (130) based on information stored in the user level, query generation unit (110) by means of a query to a allocating multiple user measured may be filled.

In step S18, transmitting predetermined questions can be performed. For example, question selecting portion of Figure 1 (120) by predetermined questions are user information with user interface (103) through web service system (300) can be transmitted to, web service system (300) user information for a selected user terminals (500) transmitting to at least one of query can be received.

Figure 7 shows a calibration method of the present invention also in the embodiment according to exemplary knowledge base indicating order disclosed. Specifically, Figure 7 shows a knowledge base by evaluating the answers received from the user's telephone (200) are independent each other and indicating order correcting method, method of Figure 6 corresponding to order can be subsequently performed. In hereinafter, be described with reference to Figure 7 shows a 1 also are disclosed. As shown in fig. 7, a knowledge base (200) (S21, S23, S25, S27, S29) calibration method comprising the multiple stages can be.

In step S21, receiving an answer to the questions can be performed. 6 also 1 and also with reference to as aforementioned, in order to correct incomplete knowledge data generated queries (e.g., web service system (300) through) can be applied to user number ball, an answer to the questions can be received from a user. For example, web service system of Figure 1 (300) user terminals (500) and capable of receiving answers from, a knowledge base calibration system (100) can be received answers from transmitting. A knowledge base calibration system (100) candidate knowledge generation unit (140) includes a user interface (103) can be received through the answers.

In step S23, generating candidate knowledge data can be performed. For example, of Figure 1 candidate knowledge generation unit (140) is question, answer and knowledge data including user information of the user with candidate and can be generate. In addition, in one in the embodiment, candidate knowledge generation unit (140) is additionally answer of the user level information knowledge data further including generating candidate disapproval. Candidate knowledge generation unit (140) candidate knowledge data generated by candidate knowledge store (150) can be stored.

In step S25, validating candidate knowledge data can be performed. For example, candidate knowledge of Figure 1 section (160) includes a candidate knowledge data storage candidate knowledge store (150) by releasing the same question for a plurality of answers including candidate knowledge data and capable of receiving, candidate knowledge data can be verifying. 4 also refers to as aforementioned, candidate knowledge section (160) have the same answer number may be based on the performance of a number of users to a candidate knowledge data verifying, a variable number of users the same answer for verifying the performance of a candidate knowledge data may be filled.

If upon successful verification of candidate knowledge data, knowledge data in step S27 correcting can be performed. For example, candidate knowledge of Figure 1 section (160) is at least one of values is one property of 2 including knowledge data are written in error, verified answer on knowledge data to include corresponding values can be correct. In addition, missing attribute by and having insufficient knowledge data, corresponding attribute value corresponding verified to have correct answer on knowledge data may be filled.

There is verification of candidate knowledge data, allocating a predetermined user in step S29 time different user to enter responses can be performed. For example, question of Figure 1 selecting unit (120) includes a candidate knowledge section (160) fails verification from a candidate knowledge data receiving (or corresponding question) can be, question selecting portion (120) is not previously assigned questions are questions can be allocating a plurality of users. New user questions are allocated to number ball can be switched on, their knowledge data can be received from the answer are used re-sword candidate of liver fibrosis.

Drawing and specification are exemplary in the embodiment in diode been disclosure. The specification describes in the embodiment using a specified terms should, this technical idea of the present invention only to explain the meaning of the present invention claim is used in limiting a number range for valve timing of the previously used are not correct. The art therefrom if various deformation and equally in the embodiment is enabling other person with skill in the art will understand. The, scope of protection of the present invention is defined by attached claim descriptive true generated by the technical idea of will.