Knowledge query method and system

Abstract

The present application discloses a knowledge query method and system, which can be applied to the financial field or other fields. The first server receives a query statement; if the first server does not search for a first knowledge point with a correlation higher than the first threshold with the query statement, and searches for a second knowledge point with a correlation higher than the second threshold and less than or equal to the first threshold with the query statement, and the first server is not a head office server, the query statement and the identifier of the second knowledge point are sent to the second server; the Mth server obtains the query range corresponding to the identifier, and M takes values ​​from 2 in sequence; if the Mth server does not search for the first knowledge point from the query range, and the Mth server is not a head office server, the query statement and the identifier are sent to the M+1th server, M=M+1 is set, and the step of returning to the Mth server to obtain the query range is returned until the first knowledge point is queried or the Mth server is a head office server. Thereby, the search results presented by the client are highly likely to be the search results required by the user.

Description

Technical Field

[0001] The present application relates to the field of database technology, and more specifically, to a knowledge query method and system. Background Art

[0002] The user can input a query statement through the client, and the server can search its corresponding database for knowledge points whose relevance to the query statement is higher than a preset threshold, and feed back to the client.

[0003] If no knowledge point with a relevance higher than the preset threshold to the query statement is found in the database corresponding to the server, information indicating that no relevant knowledge point has been found is fed back to the client, or knowledge points with a relevance lower than the preset threshold are fed back to the client, so that the search results presented to the client are most likely not the search results required by the user. Summary of the invention

[0004] In view of this, the present application provides a knowledge query method and system.

[0005] To achieve the above objectives, this application provides the following technical solutions:

[0006] According to a first aspect of an embodiment of the present disclosure, a knowledge point query method is provided, comprising:

[0007] A first server receives a query statement sent by a client, where the first server is a server closest to the client;

[0008] If the first server does not search for a first knowledge point whose relevance to the query statement is higher than a first threshold value from the first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than a second threshold value and less than or equal to the first threshold value is searched from the first database, and the first server is not a head office server, the query statement and the identification ID of the second knowledge point are sent to a second server, the second server includes a server of a superior branch of the first server, and / or a server at the same level as the first server;

[0009] The Mth server obtains a query range from the Mth database corresponding to the Mth server, the query range including multiple knowledge points belonging to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, the associated knowledge points and the second knowledge point do not belong to the same source document, and M is a positive integer starting from 2;

[0010] If the Mth server searches for the first knowledge point from the query range, the first knowledge point is sent to the M-1th server, so that the M-1th server sends the first knowledge point to the client;

[0011] If the Mth server does not search the first knowledge point from the query scope, and the Mth server is not a head office server, the query statement and the identifier of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server is returned to step M. The Mth server obtains the query scope from the second database corresponding to the Mth server, and the Mth server includes the server of the upper branch of the M-1th server, and / or the server at the same level as the M-1th server;

[0012] If the Mth server does not search for the first knowledge point from the query scope, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented.

[0013] According to a second aspect of an embodiment of the present disclosure, a knowledge query system is provided, including:

[0014] The first server is used to:

[0015] Receive query statements sent by the client;

[0016] If the first server does not search for a first knowledge point whose relevance to the query statement is higher than a first threshold value from the first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than a second threshold value and less than or equal to the first threshold value is searched from the first database, the query statement and the identification ID of the second knowledge point are sent to a second server, the second server includes a server of an upper branch of the first server, and / or a server at the same level as the first server;

[0017] The Mth server is used for:

[0018] Obtaining a query range from an Mth database corresponding to the Mth server, the query range including a plurality of knowledge points belonging to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, the associated knowledge points not belonging to the same source document as the second knowledge point, and M being a positive integer starting from 2;

[0019] If the first knowledge point is found in the query range, sending the first knowledge point to the M-1th server, so that the M-1th server sends the first knowledge point to the client;

[0020] If the first knowledge point is not found in the query scope, and the Mth server is not a head office server, the query statement and the identifier of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server is returned to step M. The Mth server obtains the query scope from the second database corresponding to the Mth server, and the Mth server includes the server of the upper branch of the M-1th server, and / or the server at the same level as the M-1th server;

[0021] If the first knowledge point is not found in the query scope, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented.

[0022] Through the above technical solution, it can be known that in the knowledge point query method provided by the present application, the first server receives the query statement sent by the client, and the first server is the server closest to the client; if the first server does not search for the first knowledge point whose relevance to the query statement is higher than the first threshold from the first database corresponding to the first server, and the second knowledge point whose relevance to the query statement is higher than the second threshold and less than or equal to the first threshold is searched from the first database, and the first server is not the head office server, then the search range is expanded, that is, the query statement and the identification ID of the second knowledge point are sent to the second server, the second server includes the server of the superior branch of the first server, and / or the server at the same level as the first server; the Mth server obtains the query range from the Mth database corresponding to the Mth server, and the query range includes multiple knowledge points belonging to the same source document as the second knowledge point and related knowledge points whose relevance to the second knowledge point is greater than or equal to the third threshold. point, the associated knowledge point and the second knowledge point do not belong to the same source document, and M is a positive integer starting from 2; if the Mth server searches for the first knowledge point from the query range, the first knowledge point is sent to the M-1th server, so that the M-1th server sends the first knowledge point to the client; if the Mth server does not search for the first knowledge point from the query range, and the Mth server is not a head office server, the search range is expanded, the query statement and the identification ID of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server returns to step M to obtain the query range from the second database corresponding to the Mth server, the Mth server includes the server of the superior branch of the M-1 server, and / or the server at the same level as the M-1 server; if the Mth server does not search for the first knowledge point from the query range, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented. During the search process, if the first knowledge point is not found in a certain database, the search scope will be expanded until the first knowledge point is found, thereby improving the retrieval accuracy and making it very likely that the retrieval results presented by the client are the retrieval results required by the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0024] Figure 1 A hierarchical network architecture diagram in the banking field provided by an embodiment of the present application;

[0025] Figure 2 A flowchart of a method for querying knowledge points provided in an embodiment of the present application;

[0026] Figure 3 A structural diagram of an implementation method of the knowledge query system provided in an embodiment of the present application;

[0027] Figure 4 A structural diagram of a knowledge query system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0029] The embodiments of the present application provide a knowledge point query method and system. Before introducing the technical solution provided by the embodiments of the present application, the application scenarios involved in the embodiments of the present application are first described.

[0030] It should be noted that the knowledge point query method and system provided by the present invention can be used in the financial field or other fields, for example, can be used in retrieval application scenarios in the financial field. Other fields are any fields other than the financial field, for example, retrieval application scenarios in the electric power field. The above is only an example and does not limit the application field of the knowledge point query method and system provided by the present invention.

[0031] The following describes the application scenarios of the embodiments of the present application by taking the banking field in the financial field as an example.

[0032] like Figure 1 As shown, it is a hierarchical network architecture diagram in the banking field provided by an embodiment of the present application.

[0033] Figure 1 For example only, Figure 1 The hierarchical network architecture of the bank shown in the figure includes three layers, namely, the head office server and head office database (collectively referred to as equipment) located in the head office, the branch server and branch database (collectively referred to as equipment) located in the first-level branch, and the branch server and branch database (collectively referred to as equipment) located in the branch in the city where the first-level branch is located. The equipment located in the branch communicates with the equipment located in the first-level branch through a communication link; the equipment located in the first-level branch communicates with the equipment located in the head office through a communication link.

[0034] Exemplarily, the hierarchical network architecture may include four or more layers. For example, the four-layer network architecture may include: equipment located at the head office, equipment located at the first-level branch, equipment located at the second-level branch, and equipment located at the sub-branch. The equipment located at the sub-branch communicates with the equipment located at the second-level branch through a communication link; the equipment located at the second-level branch communicates with the equipment located at the first-level branch through a communication link; the equipment located at the first-level branch communicates with the equipment located at the head office through a communication link.

[0035] In the embodiment of the present application, the server located at the head office is called the head office server, the database located at the head office is called the head office database, the server located at the branch (first-level branch or second-level branch) is called the branch server, the database located at the branch is called the branch database, the server located at the branch is called the branch server, and the database located at the branch is called the branch database.

[0036] Exemplarily, each server corresponds to one or more databases. Figure 1 In this example, each server corresponds to one database. Figure 1 In the figure, the server and its corresponding database are connected with a dotted line to indicate the corresponding relationship between the two.

[0037] Illustratively, there may be one or more servers at the same level. Figure 1 The example in which each level includes at least two servers is used for explanation. For example, there are two servers at the same level of branch, two servers at the same level of head office, and two servers at the same level of branch.

[0038] It is understandable that different servers at the same level may be located in different or the same regions, for example, a branch server and branch database located in Henan, and a branch server and branch database located in Hebei.

[0039] For example, the knowledge points stored in the databases corresponding to servers in different regions at the same level may be different. For example, the knowledge points stored in the database of a branch in Henan may be different from those stored in the database of a branch in Hebei. This is because the policies in different regions may be different. The following is an example of ETC (Electronic Toll Collection).

[0040] It is understandable that different regions have different ETC policies. For example, the ECT recharge policies are different. For example, Henan has a preferential policy of giving 50 yuan for a 50 yuan ETC recharge, so the knowledge point with the knowledge title "ECT Recharge Policy" contains the knowledge content "ETC recharge 500 yuan and give 50 yuan", while Hebei does not have a preferential policy of giving away gifts for recharge, so the knowledge point with the knowledge title "ECT Recharge Policy" contains the knowledge content "No preferential policy".

[0041] Exemplarily, in summary, the knowledge points with the same knowledge title stored in the databases corresponding to the servers in different regions at the same level may contain different knowledge contents.

[0042] Exemplarily, knowledge points with the same knowledge title stored in databases corresponding to servers in different regions at the same level may contain the same knowledge content.

[0043] Exemplarily, the head office database stores knowledge points in databases corresponding to servers at various levels and in various regions.

[0044] Exemplarily, the branch database located in the branch stores knowledge points related to the region where the branch database is located. For example, the branch database located in Henan stores knowledge points related to Henan, but does not store knowledge points related to Hebei. Exemplarily, the branch database located in the sub-branch stores knowledge points related to the region where the branch database is located. For example, the branch database located in Henan stores knowledge points related to Henan, but does not store knowledge points related to Hebei.

[0045] The structure of the knowledge points is explained below.

[0046] Exemplarily, the knowledge points are obtained by splitting the source document. The source documents corresponding to different knowledge points may be the same or different.

[0047] For example, the source document includes 10 paragraphs. Exemplarily, the source text is divided into 10 knowledge points, that is, each paragraph corresponds to one knowledge point; Exemplarily, the source document is divided into 5 knowledge points, and one knowledge point includes one or more paragraphs in the source document; Exemplarily, the source document is divided into 20 knowledge points, and one knowledge point includes one or more sentences in a paragraph.

[0048] Exemplarily, knowledge points are obtained based on other knowledge points.

[0049] Exemplarily, there are many ways to represent knowledge points, and the present application embodiment provides but is not limited to: any one of a linked list, an array, a structure, and a table. The structure of a knowledge point is described below using a table as an example.

[0050] Exemplarily, the structure of a knowledge point may be as shown in Table 1.

[0051] Table 1 Structure of knowledge points

[0052]

[0053]

[0054] Exemplarily, the body of the knowledge includes any one of the knowledge title and the knowledge content; Exemplarily, the knowledge content corresponds to the knowledge title. For example, the knowledge content corresponding to the knowledge title "Deposit Period of Bank of China Fortune Personal Notice Deposits" may be: Regardless of the actual deposit period, personal notice deposits are divided into two types: 1-day notice deposits and 7-day notice deposits according to the length of the depositor's advance notice. 1-day notice deposits must notify the agreed withdrawal of deposits 1 day in advance, while 7-day notice deposits must notify the agreed withdrawal of deposits 7 days in advance. For 1-day notice deposits and 7-day notice deposits, customers must come to the counter 1 day or 7 days in advance to make an appointment for withdrawal registration. The so-called transfer means that the system can automatically transfer the principal and interest to the customer's current account on the due date, but the customer is required to come to the counter every cycle (7 days) to handle the agreed transfer business.

[0055] Exemplarily, the "business directory" field included in the knowledge point may include one or more levels of directories. Table 1 is described by taking the example that the business directory includes two levels of directories.

[0056] Exemplarily, the business directory included in the knowledge point is used to indicate the storage path of the knowledge text included in the knowledge point. Exemplarily, the business directory included in the knowledge point is the name of the storage device storing the knowledge point and / or the name of the folder.

[0057] Exemplarily, the knowledge title can be obtained from the knowledge content text based on natural language processing technology; exemplary, the knowledge title can be obtained from the source document.

[0058] Exemplarily, the "graph label" field contained in the knowledge point refers to the attribute information of the knowledge text. Exemplarily, the graph label includes: product (the product described by the content of the knowledge text), bank (which branch or which head office the content of the knowledge text is aimed at), customer type. Exemplarily, customer types include but are not limited to: individual customers, wealth management customers, ordinary customers, private banking customers, BOC wealth management customers, etc.

[0059] Exemplarily, the "personal label" field contained in the knowledge point is added by a manual agent. The manual agent can mark the knowledge point based on his or her understanding of the knowledge point. For example, the manual agent with the manual agent identifier A in Table 1 marked "Fuden's special time deposit". Next time, the manual agent with the manual agent identifier A can accurately search for the knowledge points shown in Table 1 based on the query statement "Fuden's special time deposit".

[0060] It should be noted that sometimes when a human agent searches for a required knowledge point, but the human agent has some understanding of the knowledge point or has his own habit of naming the knowledge point, then the human agent can add his own understanding to the personalized label of the knowledge point. In this way, the thinking habits of different human agents can be taken into account, and the annotation of knowledge can be enriched, thereby speeding up the efficiency of knowledge query and improving the accuracy of knowledge query.

[0061] For example, for the same knowledge point, the personalized tags marked by different manual agents may be different or the same; because the manual agent needs to log in before searching for knowledge points, the query statement of the manual agent includes the identification of the manual agent. Therefore, in the process of retrieving knowledge points by personalized tags, it will not be affected by the personalized tags marked by other manual agents.

[0062] Exemplarily, the “management attribute” field included in the knowledge point refers to the information of the administrator who manages the knowledge point. For example, the content of the “management attribute” field includes the department to which the administrator belongs and the user group to which the administrator belongs.

[0063] Exemplarily, the knowledge point includes a "source document context index" field and a "position in source document" field. Exemplarily, the "source document context index" field may include: a first identifier and a second identifier, wherein the first identifier is the position of the previous paragraph of the knowledge point in the source document, and the second identifier is the position of the next paragraph of the knowledge point in the source document. There are many ways to represent the first identifier and the second identifier, for example, using one or more of letters, numbers, and special symbols. Table 1 uses the first identifier and the second identifier represented by numbers as an example for explanation.

[0064] Exemplarily, the identifier of a knowledge point may be randomly assigned, and different knowledge points may have different identifiers; exemplary, the identifier of a knowledge point is related to the position of the knowledge text contained in the knowledge point in the source document.

[0065] For example, the 3rd, 4th and 5th paragraphs contained in the source document correspond to a knowledge point respectively, and the knowledge point shown in Table 1 corresponds to the 4th paragraph contained in the source document, so the "position in the source document" is 4; exemplarily, the source document context index includes a first identifier and a second identifier, as shown in Table 1, the first identifier is 3 and the second identifier is 5.

[0066] Exemplarily, the knowledge point also includes: at least one of an identification ID of a source document and a name of the source document.

[0067] Exemplarily, the knowledge point further includes: an association relationship identification ID, where the association relationship identification ID is the ID of a knowledge point whose correlation with the knowledge point is higher than or equal to a third threshold.

[0068] In summary, the knowledge points provided in the embodiments of the present application include the source document context index (i.e., the first identifier and the second identifier), so that multiple knowledge points from the same source document are associated. The knowledge points provided in the embodiments of the present application include the association relationship identifier ID, so that multiple knowledge points from different source documents are associated.

[0069] Exemplarily, the number of association relationship identification IDs included in a knowledge point may be one or more.

[0070] The structure of the knowledge points in Table 1 is only an example and does not limit the structure of the knowledge points. For example, the knowledge points may include: business catalog, knowledge text, graph label, personal label, one or more fields in management attributes.

[0071] Exemplarily, the knowledge point also includes a receiving group, and the receiving group includes identifiers of users who can query and obtain the knowledge point.

[0072] The keywords in the query statements involved in the embodiments of the present application may include keywords belonging to one or more fields in the business catalog, knowledge text, graph label, personal label, and management attribute. In the process of retrieving the first knowledge point whose relevance to the query statement is greater than or equal to the first threshold from the knowledge base, the relevance between the query statement and one or more of the business catalog, knowledge text, graph label, personal label, and management attribute contained in the knowledge point can be obtained to obtain the first knowledge point whose relevance to the query statement is greater than or equal to the first threshold, so that the client can display the link to the first knowledge point whose relevance to the query statement is greater than or equal to the first threshold.

[0073] Illustratively, the number of the first knowledge points may be one or more.

[0074] The following describes the knowledge point query method provided by the embodiment of the present application in combination with the above application scenario. Figure 2 As shown, it is a flowchart of the knowledge point query method provided in an embodiment of the present application, and the method includes the following steps S21 to S26.

[0075] Step S21: The first server receives a query statement sent by the client.

[0076] Step S22: If the first server does not search for a first knowledge point whose relevance to the query statement is higher than the first threshold from the first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than the second threshold and less than or equal to the first threshold is searched from the first database, and the first server is not the head office server, send the query statement and the identification ID of the second knowledge point to the second server.

[0077] The first server is a server that is closest to the client.

[0078] It can be understood that if the first server is closest to the client, the transmission path between the first server and the client is the shortest, that is, if the first knowledge point can be searched from the first database, the speed of searching for the first knowledge point is the fastest; if the first knowledge point cannot be searched from the first database, the speed of searching for the second knowledge point from the first database is the fastest.

[0079] Exemplarily, if the first knowledge point can be searched from the first database, the first knowledge point is fed back to the client.

[0080] Exemplarily, the first server may be a branch server, a branch server or a head office server. If the first server is a branch server of a second-level branch, the server of the parent branch of the first server is a branch server of a first-level branch, and the server at the same level as the first server is a branch server of a second-level branch. If the first server is a branch server of a first-level branch, the server of the parent branch of the first server is a head office server of the head office, and the server at the same level as the first server is a branch server of a first-level branch.

[0081] It can be understood that if the first server does not search for a first knowledge point whose relevance to the query statement is higher than the first threshold from the first database corresponding to the first server, it means that the first knowledge point is not stored in the first database. If the first server is the head office server, it means that the entire bank lacks the first knowledge point.

[0082] If the first server does not search for a first knowledge point with a relevance to the query statement higher than a first threshold from the first database corresponding to the first server, and the first server is not a head office server, there may be the following three situations.

[0083] The first situation: the head office server does not synchronize the first knowledge point stored in the head office database to any branch bank database or sub-branch bank database.

[0084] The second situation: the head office server synchronizes the first knowledge point stored in the head office database to the branch database of the superior branch of the first database, or synchronizes it to the database at the same level as the first database, but does not synchronize it to the first database.

[0085] The third situation: the first knowledge point is not stored in the head office database corresponding to the head office server, that is, the first knowledge point is missing.

[0086] In summary, the query statement and the identification ID of the second knowledge point can be sent to the second server. The second server includes a server of a superior branch of the first server. Exemplarily, the second server includes a server of a superior branch in the same region as the first server.

[0087] Exemplarily, the second server includes a server at the same level as the first server. Exemplarily, the second server includes a server at the same level and in the same region as the first server.

[0088] Exemplarily, the number of the second servers is one or more. If the number of the second servers is multiple, the second servers can be accessed in order from near to far based on the distance from the client.

[0089] Exemplarily, the difference between the first threshold and the second threshold can be determined based on actual conditions. Exemplarily, the difference between the first threshold and the second threshold is less than or equal to threshold A, and threshold A makes the first knowledge point and the second knowledge point belong to the same application scenario, for example, the ETC application scenario.

[0090] Exemplarily, the difference between the first threshold and the second threshold is less than or equal to 0.05.

[0091] Exemplarily, the relevance between the knowledge point and the query statement is calculated based at least on the knowledge title and knowledge content contained in the knowledge point. The reason is as follows: the first database may store knowledge points whose knowledge titles are exactly the same as the query statement, but whose "knowledge content" field does not describe the content corresponding to the knowledge title.

[0092] For example, the query statement is: ETC withholding preferential policy, the knowledge title of the first knowledge point is ETC withholding preferential policy, and the knowledge title of the second knowledge point is ETC withholding preferential policy. However, the specific content of the "knowledge content" field contained in the second knowledge point is: follow the policy of the head office.

[0093] Exemplarily, the first database may store knowledge points whose knowledge titles are not completely the same as the query statement, but are very similar.

[0094] For example, the query statement is: ETC deduction preferential policy, the knowledge title of the first knowledge point is ETC deduction preferential policy, and the knowledge title of the second knowledge point is ETC recharge preferential policy.

[0095] Step S23: The Mth server obtains the query scope from the Mth database corresponding to the Mth server.

[0096] The query scope includes multiple knowledge points that belong to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, and the associated knowledge points do not belong to the same source document as the second knowledge point.

[0097] M is a positive integer starting from 2. That is, M = 2, 3, 4, ...

[0098] Exemplarily, the associated knowledge point can be obtained through the "association relationship identification ID" field contained in the second knowledge point, and multiple knowledge points belonging to the same source document as the second knowledge point can be obtained through the "source document name" field contained in the second knowledge point, thereby obtaining the query scope.

[0099] Since the Mth server only searches for the first knowledge point within the query scope, the query time is shortened. Since the second knowledge point has a high relevance to the query statement, the second knowledge point and the first knowledge point belong to the same application scenario, for example, the ETC application scenario. Therefore, the knowledge points included in the query scope have already covered the knowledge points with a high relevance to the query statement. Therefore, searching for the first knowledge point within the query scope not only shortens the query time, but also improves the query accuracy.

[0100] Step S24: If the Mth server searches for the first knowledge point from the query range, the first knowledge point is sent to the M-1th server, so that the M-1th server sends the first knowledge point to the client.

[0101] Exemplarily, the Mth server needs to send the first knowledge point to the M-1th server, and the M-1th server stores the first knowledge point in the M-1th database; the M-1th server needs to send the first knowledge point to the M-2th server, and the M-2th server stores the first knowledge point in the M-2th database; the M-2th server needs to send the first knowledge point to the M-3th server, and the M-3th server stores the first knowledge point in the M-3th database;…, until it is sent to the first server, the first server stores the first knowledge point in the first database, and feeds back to the client.

[0102] Step S25: If the Mth server does not search the first knowledge point from the query scope, and the Mth server is not a head office server, send the query statement and the identifier of the second knowledge point to the M+1th server, set M=M+1, and return to step S23.

[0103] The Mth server includes a server of an upper-level branch of the M-1th server and / or a server at the same level as the M-1th server.

[0104] In order for those skilled in the art to better understand the knowledge query system provided by the embodiment of the present application, a specific example is given below for illustration.

[0105] like Figure 3 , which is a structural diagram of an implementation method of the knowledge query system provided in an embodiment of the present application.

[0106] Figure 3 The hierarchical network architecture of the bank corresponding to the knowledge query system includes three layers, namely, the head office server 31 and the head office database 32 located in the head office, the branch server 33 and the branch database 34 located in the first-level branch, and the branch server 35 and the branch database 36 located in the city where the first-level branch is located.

[0107] The user inputs a query statement on the client, and the branch server 35 (corresponding to the first server) is closest to the client. If the branch server 35 does not search the branch database 36 for a first knowledge point whose relevance to the query statement is higher than the first threshold, and searches the branch database 36 for a second knowledge point whose relevance to the query statement is higher than the second threshold and less than or equal to the first threshold, the query statement and the identification ID of the second knowledge point are sent to the branch server 33 (corresponding to the second server).

[0108] The branch server 33 obtains the query scope from the branch database 34. If the branch server 33 fails to find the first knowledge point in the query scope, it sends the query statement and the identifier of the second knowledge point to the head office server 31 (corresponding to the third server).

[0109] The head office server 31 obtains the query scope from the head office database 32 . If the head office server 31 does not search for the first knowledge point in the query scope, it determines that the first knowledge point is a knowledge point to be supplemented.

[0110] If the head office server 31 searches for the first knowledge point in the query range, it sends the first knowledge point to the branch server 33, and the branch server 33 stores the first knowledge point in the branch database 34, and sends the first knowledge point to the sub-branch server 35. The sub-branch server 35 stores the first knowledge point in the sub-branch database 36, and sends the first knowledge point to the client.

[0111] Step S26: If the Mth server does not search for the first knowledge point from the query scope, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented.

[0112] In an optional implementation, the above devices are sorted from near to far according to the distance from the client, and the sorting result is: first server, second server, third server, ..., Mth server, .... That is, the embodiment of the present application queries each server in turn according to the distance from the client from near to far.

[0113] In an optional implementation, the distances between the second server, the third server, ..., the Mth server and the client are not limited.

[0114] In the knowledge point query method provided in the embodiment of the present application, a first server receives a query statement sent by a client, and the first server is the server closest to the client; if the first server does not search for a first knowledge point whose relevance to the query statement is higher than a first threshold from a first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than a second threshold and less than or equal to the first threshold is searched from the first database, and the first server is not a head office server, then the search range is expanded, that is, the query statement and the identification ID of the second knowledge point are sent to the second server, and the second server includes the server of the superior branch of the first server, and / or the server at the same level as the first server; the Mth server obtains the query range from the Mth database corresponding to the Mth server, and the query range includes multiple knowledge points belonging to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, the The associated knowledge point and the second knowledge point do not belong to the same source document, and M is a positive integer starting from 2; if the Mth server searches for the first knowledge point from the query range, the first knowledge point is sent to the M-1th server, so that the M-1th server sends the first knowledge point to the client; if the Mth server does not search for the first knowledge point from the query range, and the Mth server is not a head office server, the search range is expanded, the query statement and the identifier of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server returns to step M to obtain the query range from the second database corresponding to the Mth server, the Mth server includes the server of the superior branch of the M-1 server, and / or the server at the same level as the M-1 server; if the Mth server does not search for the first knowledge point from the query range, and the Mth server is a head office server, the first knowledge point corresponding to the query statement is determined to be a knowledge point to be supplemented. During the search process, if the first knowledge point is not found in a certain database, the search scope will be expanded until the first knowledge point is found, thereby improving the retrieval accuracy and making it very likely that the retrieval results presented by the client are the retrieval results required by the user.

[0115] In an optional implementation, the following method for supplementing knowledge points is also included, and the method includes the following steps A11 to A16.

[0116] Step A11: If the first knowledge point corresponding to the query statement is a knowledge point to be supplemented, the head office server calculates the relevance between the query statement and the preset business tags of each server.

[0117] Exemplarily, each server includes a branch server or a sub-branch server. The preset business tag of each server is pre-set and is related to the main business of the sub-branch or branch to which the server belongs.

[0118] For example, if a branch is mainly responsible for ETC business, the preset business label corresponding to the branch is ETC.

[0119] Step A12: the head office server sends the query statement to a target server of a branch bank whose corresponding preset business tag has a correlation with the query statement that is greater than or equal to a fourth threshold, and the supplementary knowledge task includes the query statement.

[0120] Exemplarily, the target server may be a branch server or a sub-branch server.

[0121] Step A13: The target server determines the knowledge type of the supplementary knowledge constituting the first knowledge point based on the query statement.

[0122] Exemplarily, cluster analysis is performed on the keywords included in the query statement to obtain the knowledge types constituting the first knowledge point.

[0123] Exemplarily, the cluster analysis method includes: converting each keyword into a corresponding word vector. This step can be implemented using a pre-built word vector model.

[0124] Among them, the word vector model (word2vec) is an existing mathematical model. After training with a large amount of corpus, the word vector model can convert each keyword into a corresponding word vector. Specifically in this embodiment, the word vector model can be used to convert each keyword into a corresponding word vector.

[0125] The keywords are clustered according to the similarity between the word vectors corresponding to each two keywords. For each two keywords, if the similarity between the word vectors of the two keywords is greater than or equal to the preset threshold, it is determined that the two keywords correspond to the same knowledge type. Conversely, if the similarity between the word vectors of the two keywords is less than the preset threshold, it is determined that the two keywords correspond to different knowledge types. After clustering is completed, the knowledge type is determined.

[0126] For example, the query statement includes: ETC business, materials, and payment. Then, through clustering, it can be determined that the knowledge types include: ETC business processing materials and ETC business payment rules.

[0127] Step A14: the target server sends the supplementary knowledge task to the editing terminal corresponding to the knowledge type included in the supplementary knowledge task, and the supplementary knowledge task includes: the knowledge type of the supplementary knowledge and the supplementary direction information of the supplementary knowledge.

[0128] Exemplarily, the editing terminal may be a work terminal of any employee of the bank.

[0129] The following example illustrates the supplementary direction information. For example, the supplementary direction information may include: the need to provide as detailed an explanation and description as possible for the knowledge type of the supplementary knowledge, for example, the need to provide a directory of the corresponding knowledge type to be supplemented.

[0130] Step A15: the target server receives the supplementary knowledge uploaded by each of the editing terminals, and combines the supplementary knowledge uploaded by each of the editing terminals into the first knowledge point.

[0131] Exemplarily, if there are multiple editing terminals, the supplementary knowledge uploaded by the multiple editing terminals can be merged to obtain the first knowledge point.

[0132] Exemplarily, the process of merging the supplementary knowledge uploaded by multiple editing terminals to obtain the first knowledge point may include the following steps B11 to B13.

[0133] Step B11: According to the similarity between each two pieces of supplementary knowledge, duplicate supplementary knowledge can be identified from the received multiple pieces of supplementary knowledge, and the duplicate supplementary knowledge can be deleted.

[0134] Step B12: For each item of supplementary knowledge, it can be detected whether the supplementary knowledge is related to the corresponding knowledge type. If the supplementary knowledge is not related to the corresponding knowledge type, the supplementary knowledge is deleted.

[0135] Step B13: Merge the remaining supplementary knowledge.

[0136] Step A16: The target server synchronizes the first knowledge point to a database corresponding to the server that has searched for the first knowledge point.

[0137] In an optional implementation, if the target server detects that the first knowledge point has changed, it obtains change information; the target server sends the change information to the server that has searched for the first knowledge point.

[0138] Exemplarily, the server that has searched the first knowledge point may include one or more servers from the first server to the Mth server.

[0139] Exemplarily, the server that has searched for the first knowledge point may agree to change the first knowledge point stored in its corresponding database based on the change information, so that the first knowledge point stored in the database corresponding to the server is the same as the first knowledge point stored in the target database corresponding to the target server.

[0140] For example, it is understandable that policies in different regions are different, so there may be gaps in knowledge points. Therefore, the server that has searched for the first knowledge point may not agree to change the first knowledge point stored in its corresponding database based on the change information, so that the first knowledge point stored in the database corresponding to the server is different from the first knowledge point stored in the target database corresponding to the target server.

[0141] The method is described in detail in the embodiments disclosed in the above-mentioned application. The method of the application can be implemented by various forms of devices. Therefore, the application also discloses a device, and a specific embodiment is given below for detailed description.

[0142] like Figure 4 FIG. 1 is a structural diagram of a knowledge query system provided in an embodiment of the present application, and the system includes:

[0143] The first server 41 is used for:

[0144] Receive query statements sent by the client;

[0145] If the first server does not search for a first knowledge point whose relevance to the query statement is higher than a first threshold value from the first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than a second threshold value and less than or equal to the first threshold value is searched from the first database, the query statement and the identification ID of the second knowledge point are sent to a second server, the second server includes a server of an upper branch of the first server, and / or a server at the same level as the first server;

[0146] The Mth server 42 is used for:

[0147] Obtaining a query range from a second database corresponding to the Mth server, the query range including multiple knowledge points belonging to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, the associated knowledge points not belonging to the same source document as the second knowledge point, and M being a positive integer starting from 2;

[0148] If the first knowledge point is found in the query range, sending the first knowledge point to the M-1th server, so that the M-1th server sends the first knowledge point to the client;

[0149] If the first knowledge point is not found in the query scope, and the Mth server is not a head office server, the query statement and the identifier of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server is returned to step M. The Mth server obtains the query scope from the second database corresponding to the Mth server, and the Mth server includes the server of the upper branch of the M-1th server, and / or the server at the same level as the M-1th server;

[0150] If the first knowledge point is not found in the query scope, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented.

[0151] In an optional implementation, the head office server is further used for:

[0152] If the first knowledge point corresponding to the query statement is a knowledge point to be supplemented, the head office server calculates the relevance between the query statement and the preset business tags of each branch;

[0153] Sending the query statement to a target server of a branch whose corresponding preset business tag has a correlation with the query statement that is greater than or equal to a fourth threshold, wherein the supplementary knowledge task includes the query statement;

[0154] The target server is used to:

[0155] Determining the knowledge type of the supplementary knowledge constituting the first knowledge point based on the query statement;

[0156] Sending a supplementary knowledge task to an editing terminal corresponding to a knowledge type included in the supplementary knowledge task, wherein the supplementary knowledge task includes: a knowledge type of the supplementary knowledge and supplementary direction information of the supplementary knowledge;

[0157] receiving the supplementary knowledge uploaded by each of the editing terminals, and combining the supplementary knowledge uploaded by each of the editing terminals into the first knowledge point;

[0158] The first knowledge point is synchronized to a database corresponding to the server that has searched for the first knowledge point.

[0159] In an optional implementation, the target server is further used for:

[0160] If a change is detected in the first knowledge point, obtaining change information;

[0161] The change information is sent to the server that has searched the first knowledge point.

[0162] It should be noted that the features described in the various embodiments in this specification can be replaced or combined with each other. For the device or system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0163] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0164] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0165] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A knowledge point query method, characterized in that: include: A first server receives a query statement sent by a client, where the first server is a server closest to the client; If the first server does not search for a first knowledge point whose relevance to the query statement is higher than a first threshold value from the first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than a second threshold value and less than or equal to the first threshold value is searched from the first database, and the first server is not a head office server, the query statement and the identification ID of the second knowledge point are sent to a second server, the second server includes a server of a superior branch of the first server, and / or a server at the same level as the first server; The Mth server obtains a query range from the Mth database corresponding to the Mth server, the query range including multiple knowledge points belonging to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, the associated knowledge points and the second knowledge point do not belong to the same source document, and M is a positive integer starting from 2; If the Mth server searches for the first knowledge point from the query range, the first knowledge point is sent to the M-1th server, so that the M-1th server sends the first knowledge point to the client; If the Mth server does not search the first knowledge point from the query scope, and the Mth server is not a head office server, the query statement and the identifier of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server is returned to step M. The Mth server obtains the query scope from the second database corresponding to the Mth server, and the Mth server includes the server of the upper branch of the M-1th server, and / or the server at the same level as the M-1th server; If the Mth server does not search for the first knowledge point from the query scope, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented.

2. The method for querying knowledge points according to claim 1, characterized in that: Also includes: If the first knowledge point corresponding to the query statement is a knowledge point to be supplemented, the head office server calculates the relevance between the query statement and the preset business tags of each server; The head office server sends the query statement to a target server of a branch whose corresponding preset business tag has a correlation with the query statement that is greater than or equal to a fourth threshold, and the supplementary knowledge task includes the query statement; The target server determines the knowledge type of the supplementary knowledge constituting the first knowledge point based on the query statement; The target server sends the supplementary knowledge task to the editing terminal corresponding to the knowledge type included in the supplementary knowledge task, wherein the supplementary knowledge task includes: the knowledge type of the supplementary knowledge and the supplementary direction information of the supplementary knowledge; The target server receives the supplementary knowledge uploaded by each editing terminal, and combines the supplementary knowledge uploaded by each editing terminal into the first knowledge point; The target server synchronizes the first knowledge point to a database corresponding to the server that has searched for the first knowledge point.

3. The method for querying knowledge points according to claim 2, characterized in that: Also includes: If the target server detects that the first knowledge point has changed, obtaining change information; The target server sends the change information to the server that has searched the first knowledge point.

4. A knowledge query system, characterized in that: include: The first server is used to: Receive query statements sent by the client; If the first server does not search for a first knowledge point whose relevance to the query statement is higher than a first threshold value from the first database corresponding to the first server, and a second knowledge point whose relevance to the query statement is higher than a second threshold value and less than or equal to the first threshold value is searched from the first database, the query statement and the identification ID of the second knowledge point are sent to a second server, the second server includes a server of an upper branch of the first server, and / or a server at the same level as the first server; The Mth server is used for: Obtaining a query range from an Mth database corresponding to the Mth server, the query range including a plurality of knowledge points belonging to the same source document as the second knowledge point and associated knowledge points whose relevance to the second knowledge point is greater than or equal to a third threshold, the associated knowledge points not belonging to the same source document as the second knowledge point, and M being a positive integer starting from 2; If the first knowledge point is found in the query range, sending the first knowledge point to the M-1th server, so that the M-1th server sends the first knowledge point to the client; If the first knowledge point is not found in the query scope, and the Mth server is not a head office server, the query statement and the identifier of the second knowledge point are sent to the M+1th server, M=M+1 is set, and the Mth server is returned to step M. The Mth server obtains the query scope from the second database corresponding to the Mth server, and the Mth server includes the server of the upper branch of the M-1th server, and / or the server at the same level as the M-1th server; If the first knowledge point is not found in the query scope, and the Mth server is a head office server, it is determined that the first knowledge point corresponding to the query statement is a knowledge point to be supplemented.

5. The knowledge query system according to claim 4, characterized in that: The head office server is also used for: If the first knowledge point corresponding to the query statement is a knowledge point to be supplemented, the head office server calculates the relevance between the query statement and the preset business tags of each branch; Sending the query statement to a target server of a branch whose corresponding preset business tag has a correlation with the query statement that is greater than or equal to a fourth threshold, wherein the supplementary knowledge task includes the query statement; The target server is used to: Determining the knowledge type of the supplementary knowledge constituting the first knowledge point based on the query statement; Sending a supplementary knowledge task to an editing terminal corresponding to a knowledge type included in the supplementary knowledge task, wherein the supplementary knowledge task includes: a knowledge type of the supplementary knowledge and supplementary direction information of the supplementary knowledge; receiving the supplementary knowledge uploaded by each of the editing terminals, and combining the supplementary knowledge uploaded by each of the editing terminals into the first knowledge point; The first knowledge point is synchronized to a database corresponding to the server that has searched for the first knowledge point.

6. The knowledge query system according to claim 5, characterized in that: The target server is also used to: If a change is detected in the first knowledge point, obtaining change information; The change information is sent to the server that has searched the first knowledge point.

Images