CN114969382B - Entity generation method based on event chain inference of event graph - Google Patents
Entity generation method based on event chain inference of event graph Download PDFInfo
- Publication number
- CN114969382B CN114969382B CN202210848876.6A CN202210848876A CN114969382B CN 114969382 B CN114969382 B CN 114969382B CN 202210848876 A CN202210848876 A CN 202210848876A CN 114969382 B CN114969382 B CN 114969382B
- Authority
- CN
- China
- Prior art keywords
- event
- entity
- node
- inference
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 28
- 230000004044 response Effects 0.000 claims description 65
- 238000012549 training Methods 0.000 claims description 56
- 238000012545 processing Methods 0.000 claims description 32
- 238000012986 modification Methods 0.000 claims description 20
- 230000004048 modification Effects 0.000 claims description 20
- 230000005540 biological transmission Effects 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 230000001174 ascending effect Effects 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 238000012790 confirmation Methods 0.000 claims description 2
- 239000000126 substance Substances 0.000 claims 2
- 238000012423 maintenance Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
- G06Q50/2057—Career enhancement or continuing education service
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Human Resources & Organizations (AREA)
- Educational Technology (AREA)
- Educational Administration (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Animal Behavior & Ethology (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an entity generation method based on event chain inference of a case map, which comprises the following steps: acquiring all first event connection nodes which are directly connected with the initial event node in the event graph; taking the first event connection node as a first event inference node, and generating a first event chain according to the first event inference node and corresponding event logic; determining a first entity inference node connected according to the first event inference node, and determining a second entity inference node connected according to the second event inference node; and determining a first entity department corresponding to the first entity inference node in the entity inference set and a second entity department corresponding to the second entity inference node according to the main information, sending the corresponding event label and the first time sequence information to the corresponding first entity department, and sending the corresponding event label and the second time sequence information to the corresponding second entity department.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to an entity generation method based on event chain inference of a case map.
Background
Events are one of the core concepts of human society, and social activities of people tend to be event-driven. The evolution laws and patterns that occur sequentially in time and space between events are a valuable piece of knowledge. However, the existing typical knowledge graph takes the entity and the attribute and relationship thereof as the research core, and lacks the description of the important human knowledge of the affairs logic. In order to make up for the deficiency, a affair map is generated, and the affair map can reveal the evolution law and the development logic of events, and describe and record human behavior activities. In the graph structure, the event graph is a directed cyclic graph, wherein nodes represent events, and directed edges represent evolutionary relations among the events.
The events have a progressive evolution relationship, and the nodes with progressive events form a corresponding event chain, and different events may have different entities.
For example, if an initial node fails for power supply, at this time, event logics at the extended part of the event node where the power supply fails include at least two event logics, the first event logic is to maintain the power supply station, the second event logic is to supply power to a target power supply area in a standby power supply manner, at this time, different corresponding entities are obtained in different logics, and a series of operations are performed on the power supply failure through the corresponding entities.
In the current technical scheme, corresponding participating entities cannot be reminded and corresponding time sequence information can be generated in the sequential evolution process of events according to a case map event chain.
Disclosure of Invention
The invention overcomes the defects of the prior art and provides an entity generation method based on the event chain inference of the event map, which can remind corresponding participating entities and generate corresponding time sequence information in the sequential evolution process of the event according to the event chain of the event map.
In order to solve the technical problems, the technical scheme of the invention is as follows:
the invention provides an entity generation method based on event chain inference of a case map, which comprises the following steps:
the method comprises the steps of S1, determining an initial event node corresponding to a current power emergency in a matter graph, and acquiring all first event connection nodes which are directly connected with the initial event node in the matter graph, wherein the first event connection nodes are any one of event nodes or entity nodes;
s2, if the first event connection node is an entity node, the first event connection node is used as a first entity inference node;
s3, if the first event connection node is an event node, taking the first event connection node as a first event inference node, and generating a first event chain according to the first event inference node and corresponding event logic;
s4, taking other event nodes except the initial event node and the first event inference node in the first event chain as second event inference nodes, determining the connected first entity inference nodes according to the first event inference nodes, and determining the connected second entity inference nodes according to the second event inference nodes;
s5, adding first time sequence information to all the first entity reasoning nodes, adding second time sequence information to all the second entity reasoning nodes, and generating an entity reasoning set according to all the first entity reasoning nodes and the second entity reasoning nodes, wherein the first time sequence information is smaller than the second time sequence information;
s6, obtaining event labels and main body information in the electric power emergency, determining a first entity department corresponding to a first entity inference node in an entity inference set and a second entity department corresponding to a second entity inference node according to the main body information, sending the corresponding event labels and the first time sequence information to the corresponding first entity department, and sending the corresponding event labels and the second time sequence information to the corresponding second entity department.
The invention has the beneficial effects that:
(1) According to the method, a corresponding first event chain is determined according to the difference of the electric power emergency, a corresponding participating entity is determined, and a corresponding first entity inference node and a second entity inference node in a case graph are determined by combining the event chain, wherein the first entity inference node is a node for solving a first level corresponding to the electric power emergency, the response time is required to be fast, the processing efficiency is required to be high, the second entity inference node is a node for solving a second level corresponding to the electric power emergency, the response time can be slower than that of the node of the first level, corresponding time sequence information is generated, the first time sequence information with short time can be added to all the first entity inference nodes, the second time sequence information with long time can be added to all the second entity inference nodes, and the corresponding participating entity is effectively reminded in the time dimension; in addition, the method also can determine a corresponding entity department for solving the emergency power event by combining the event label and the main body information in the power emergency event, and sends corresponding first time sequence information and second time sequence information to the corresponding entity department so as to remind the corresponding department of response time;
(2) The invention also can combine the quantity of the event logic and the event grade corresponding to the initial event node to determine the quantity of the labels, if the electric power emergency is more serious and the grade is higher, the corresponding response nodes are more, so as to reasonably determine the corresponding participating entities; in addition, the scheme can also adjust the event grade of the standard quantity value set in the calculation formula by combining the feedback of the user to obtain the event grade of the non-standard quantity value, so that the result calculated by the calculation formula is more in line with the actual requirement;
(3) The method comprises the steps of recording a first sending time, a second sending time, a first processing time and a second processing time to obtain a corresponding response time period, comparing the response time period with time information in corresponding time sequence information, and if the response time period is larger than a time requirement value of the time sequence information, indicating that the response speed of a department is slow, training the department to improve the subsequent response speed of the department; in addition, the scheme can also be used for calculating the training data of the relevant departments by combining the difference value between the response time period and the time requirement value corresponding to the time sequence information, carrying out targeted training on the corresponding departments, and improving the subsequent response time of the relevant departments so as to solve the power emergency more effectively.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
Fig. 1 is a schematic diagram of a case map provided by the present invention.
Detailed Description
In order that the manner in which the present invention is attained and can be more readily understood, a more particular description of the invention briefly summarized above may be had by reference to the embodiments thereof which are illustrated in the appended drawings.
The invention provides an entity generation method based on event chain inference of a case map, which comprises the following steps of S1-S6:
the method includes the steps of S1, determining an initial event node corresponding to a current power emergency in a matter graph, and obtaining all first event connection nodes which are directly connected with the initial event node in the matter graph, wherein the first event connection nodes are any one of event nodes or entity nodes.
Referring to fig. 1, when a power emergency occurs, according to the scheme, an initial event node corresponding to a current power emergency in a case graph is determined first, and then all first event connection nodes directly connected to the initial event node are found.
In practical applications, the power emergency may be a power emergency fault of a type such as "a transformer of a substation has failed".
Wherein the first event connection node includes any one of an event node or an entity node. The event node is a corresponding node of some response events which need to be done in order to solve the power emergency, for example, maintenance needs to be performed by a maintenance department, for example, emergency power supply needs to be performed on a fault area by an emergency power supply department, and the like; the entity node is a node corresponding to a main body that needs to execute a corresponding response event, for example, a corresponding "maintenance department", "emergency power supply department", and the like.
In some embodiments, the S1 includes S11-S13:
and S11, if the power emergency input by the user is judged, performing word segmentation processing on the power emergency to obtain a plurality of emergency keywords.
According to the scheme, after the user inputs the power emergency, word segmentation processing is carried out on the power emergency to obtain a plurality of emergency keywords.
For example, the power emergency may be "a transformer of a substation has a fault", and the corresponding emergency keyword after word segmentation may be "a substation", "a transformer" and "a fault".
And S12, extracting an event keyword classification table corresponding to the event map, wherein the event keyword classification table is provided with preset keywords corresponding to each event node.
The scheme is preset with an event keyword classification table corresponding to the event map, wherein the event keyword classification table comprises preset keywords corresponding to each event node. The preset keywords are, for example, "substation", "transformer", "fault", and the like.
And S13, comparing the emergency keywords with preset keywords, taking event nodes corresponding to the electric power emergency as initial event nodes, and taking all nodes directly connected with the initial event nodes in the event graph as first event connection nodes.
According to the scheme, the emergency keywords 'transformer substation', 'transformer', 'fault' are compared with preset keywords, an event node corresponding to the power emergency is found to serve as an initial event node, the initial event node corresponds to 'transformer fault', and then all nodes directly connected with the initial event node in a matter graph serve as first matter connection nodes.
According to the scheme, the first event connection node corresponding to the power emergency can be obtained through the embodiment. It is understood that there may be a plurality of first transaction connection nodes, and the power emergency is different, and the corresponding first transaction connection nodes are also different.
And S2, if the first event connection node is an entity node, taking the first event connection node as a first entity inference node.
It can be understood that if the first transaction connection node is an entity node, the present solution classifies the first transaction connection node into the entity nodes, that is, the first transaction connection node is used as the first entity inference node.
And S3, if the first event connection node is an event node, taking the first event connection node as a first event inference node, and generating a first event chain according to the first event inference node and the corresponding event logic.
It can be understood that, if the first event connection node is an event node, the first event connection node is used as a first event inference node, and a first event chain is generated according to the first event inference node and the corresponding event logic.
In some embodiments, said S3 comprises S31-S35:
and S31, taking the initial event node as a label 0 point, and sequentially carrying out ascending labels on the first event inference node and other event nodes connected with the first event inference node under the extension of each event logic.
Referring to fig. 1, there are 2 event logics corresponding to the initial event node in fig. 1, for example, respectively corresponding to "maintenance department is needed", and "emergency power supply department is needed to perform emergency power supply on fault area", in the present solution, the initial event node is taken as a label 0 point, and under the extension of each event logic, labels such as labels "1", "2", and "3" in fig. 1 are sequentially performed on the first event inference node and other event nodes connected to the first event inference node in an ascending order.
And S32, acquiring the number of the event logics and the event grade corresponding to the initial event node, and calculating to obtain the number of labels according to the number of the event logics and the event grade.
According to the scheme, the event grade corresponding to the power event is calculated firstly, and then the number of the labels is determined by combining the number of the event logics.
In practical applications, event nodes corresponding to one power event often have more, such as "nodes needing maintenance", "nodes needing maintenance tools", "nodes for logistics support", and the like, and the meaning of the number of labels in the present solution is that if the power event is more serious and the level is higher, the number of corresponding response nodes is more, and at this time, the number of labels may be up to 3 (label 3 in fig. 1); if the power event is less severe and the level is less high, the corresponding responding node does not need to be as many, and the number of labels may be cut off to 1 or 2 (label 2 in fig. 1).
It should be noted that the event level of the power emergency event can be divided into 3 levels, which are 1 level, 2 levels and 3 levels, and the higher the event level of the power emergency event is, the more urgent and important the power emergency event is; it can be understood that the more complicated the power emergency, the more the number of the required event logics is, for example, a small loose fitting fault can be only corresponding to one event logic to complete a task, and the transformer fault corresponding to a substation can affect regional power supply, and the influence range is generally wide, and the maintenance is more complicated, and the number of the corresponding event logics is usually at least 2.
It is understood that the greater the number of event logics and the higher the event level, the greater the number of corresponding reference numerals.
In some embodiments, the step S32 (obtaining the number of the event logics and the event level corresponding to the initial event node, and obtaining the number of labels according to the number of the event logics and the event level) includes:
each initial event node has an event rank of a standard magnitude set in advance thereto, and the index value is calculated by the following formula,
wherein, the first and the second end of the pipe are connected with each other,in the case of the numerical values of the reference numerals,to the extent of the amount of event logic,is a standard quantity of event logic that is,the value is normalized for the quantity,event ratings of standard magnitude.
Number of event logics in the above formulaThe larger the size, the more complicated the explanation of the power emergency, and the corresponding numerical values of the labelsThe larger; the higher the event grade of the electric power emergency, the more urgent and important the electric power emergency is, and the corresponding event grade with standard magnitudeI.e. greater, e.g. 3 greater than 2, the corresponding index valueThe larger.
And displaying the label numerical value, and if receiving confirmation information, taking the calculated label numerical value as the number of labels. The scheme can also show the calculated label number value to the user, the user can confirm or feed back according to the actual condition and the actual requirement, and if the user feels that the current label number value is more appropriate, the scheme can take the calculated label number value as the label number.
And if the modification information is received, taking a modification value corresponding to the modification information as the final label number, and adjusting the event grade of the standard quantity value according to the modification value and the label number to obtain the event grade of the non-standard quantity value. It can be understood that, if the user feels that the current label number is not accurate enough, the user may input the modification information, and after receiving the modification information of the user, the server will take the modification value corresponding to the modification information as the final label number. Meanwhile, the method can adjust the event grade of the standard quantity value according to the modified numerical value and the standard number value to obtain the event grade of the non-standard quantity value, so that the result calculated by the calculation formula is more in line with the actual requirement.
In some embodiments, if modification information is received, taking a modification value corresponding to the modification information as a final number of labels, and adjusting the event class of the standard quantity according to the modification value and the label value to obtain an event class of a non-standard quantity, where the method includes:
and if the modified numerical value is larger than the label numerical value, obtaining an enlargement coefficient according to the difference value of the modified numerical value and the label numerical value, and carrying out forward adjustment on the event grade of the standard numerical value according to the enlargement coefficient to obtain the event grade of the non-standard numerical value after the forward adjustment. It can be understood that if the modified value is greater than the numerical value of the label, which indicates that the calculated result is smaller, the server calculates an increase coefficient to perform forward adjustment on the event level of the standard value, so as to obtain the event level of the non-standard value after the forward adjustment.
And if the modified numerical value is smaller than the label numerical value, obtaining a turn-down coefficient according to the difference value of the modified numerical value and the label numerical value, and carrying out negative adjustment on the event grade of the standard numerical value according to the turn-down coefficient to obtain the event grade of the non-standard numerical value after negative adjustment. It can be understood that if the modified value is smaller than the label value, which indicates that the calculated result is larger, the server calculates a reduction coefficient to perform negative adjustment on the event level of the standard value, so as to obtain a negative-adjusted event level of the non-standard value.
The event rating of the adjusted non-standard magnitude is calculated by the following formula,
wherein, the first and the second end of the pipe are connected with each other,for a forward adjusted event level of non-standard magnitude,in order to modify the value of the numerical value,in order to adjust the weight value in the forward direction,for negative-scaled non-standard magnitude event ratings,the weight value is adjusted for the negative direction.
In the above-mentioned formula,andthe difference value between the modified numerical value input by the user and the calculated numerical value is larger, which shows that the amplitude needing to be adjusted is larger, and finally, the proper value is obtainedAndso that the numerical value of the label calculated by the above calculation formulaSMore accords with the actual demand.
And S33, after judging that the labels corresponding to the event nodes correspond to the number of the labels, stopping labeling the event nodes, and taking the event nodes with the labels corresponding to the preset number of the labels as cut-off points of corresponding event logics.
For example, referring to the above event logic in fig. 1, the label corresponding to the event logic in fig. 1 is marked with 3, and the number of labels calculated by the present solution is 3, then the present solution will take the event node of the label corresponding to the preset number of labels (label 3) in fig. 1 as the cut-off point of the corresponding event logic.
And S34, when the labels corresponding to the event nodes do not correspond to the label number and the event node under the extension of the corresponding event logic is the last one, taking the last event node as the cut-off point of the corresponding event logic.
For example, referring to the next event logic in fig. 1, the label corresponding to the event logic in fig. 1 is labeled as 2, and the number of labels calculated in the present solution is 3, then, at this time, the label corresponding to the event node does not correspond to the preset label number, and the event node under the extension of the corresponding event logic is the last one, and the event node corresponding to the label corresponding to the preset label number (label 2) in fig. 1 is taken as the cut-off point of the corresponding event logic in the present solution.
And S35, taking the mark 0 point as a starting point, counting event nodes corresponding to the starting point, the cut-off point and a middle point from the starting point to the cut-off point under each event logic to generate a first event chain.
It can be understood that, with a start point and a cut-off point, the scheme can obtain an event node corresponding to a middle point between the start point and the cut-off point to generate a first event chain. Referring to FIG. 1, there are two first chains of events, one 0-3 above and one 0-2 below.
And S4, taking other event nodes except the initial event node and the first event inference node in the first event chain as second event inference nodes, determining the connected first entity inference nodes according to the first event inference nodes, and determining the connected second entity inference nodes according to the second event inference nodes.
Referring to fig. 1, after the first event chain is obtained, the scheme uses other event nodes in the first event chain except the initial event node and the first event inference node as second event inference nodes. And then, determining a first entity inference node connected by using the first event inference node, and determining a second entity inference node connected by using the second event inference node.
In some embodiments, the S4 comprises:
and taking the event nodes corresponding to the label points 0 and 1 in each first event chain as second event inference nodes. Referring to fig. 1, event nodes corresponding to point No. 2 and point No. 3 in fig. 1 are second event inference nodes.
And if the entity node directly connected with the first event inference node exists, taking the entity node directly connected with the first event inference node as the first entity inference node. Referring to fig. 1, 3 entities corresponding to a point labeled 0 and a point labeled 1 in fig. 1 are first entity inference nodes.
And if the entity node directly connected with the second event inference node exists, taking the entity node directly connected with the second event inference node as the second entity inference node. Referring to fig. 1, 3 entities corresponding to the point labeled 2 and the point labeled 3 in fig. 1 are second entity inference nodes.
And S5, adding first time sequence information to all the first entity reasoning nodes, adding second time sequence information to all the second entity reasoning nodes, and generating an entity reasoning set according to all the first entity reasoning nodes and the second entity reasoning nodes, wherein the first time sequence information is smaller than the second time sequence information.
It can be understood that the first entity inference node is a node of a first level corresponding to the power emergency, the response time needs to be fast, and the processing efficiency needs to be high, and the second entity inference node is a node of a second level corresponding to the power emergency, and the response time may be slower than that of the node of the first level.
According to the scheme, first timing information is added to the first entity reasoning node, and second timing information is added to all the second entity reasoning nodes, wherein the first timing information is 1 hour for example, and the second timing information is 5 hours for example.
S6, acquiring event tags and main body information in the power emergency, determining a first entity department corresponding to a first entity inference node in an entity inference set and a second entity department corresponding to a second entity inference node according to the main body information, sending the corresponding event tags and the first time sequence information to the corresponding first entity department, and sending the corresponding event tags and the second time sequence information to the corresponding second entity department.
It is understood that, in the above embodiments, only the processing logic corresponding to the solution of the power emergency is determined, but the processing logic is not determined in the physical department. According to the scheme, event tags and main body information in the power emergency are acquired, then a first entity department corresponding to a first entity reasoning node in an entity reasoning set and a second entity department corresponding to a second entity reasoning node are determined by the main body information, finally, the corresponding event tags and the first time sequence information are sent to the corresponding first entity department, and the corresponding event tags and the second time sequence information are sent to the corresponding second entity department, so that the corresponding entity departments can respond timely to solve the power emergency.
In some embodiments, the S6 includes S61-S64:
and S61, determining all corresponding related departments according to the main body information, and performing word segmentation processing on the related departments to obtain department keywords.
If the main information is, for example, "a zone-12 transformer substation-128 transformer", then the scheme determines that all the corresponding associated departments may be all associated departments responsible for "a zone-12 transformer substation-128 transformer", for example, a maintenance a department, an emergency power supply B department, a logistics C department, and the like.
In addition, the scheme can also perform word segmentation processing on the relevant departments to obtain department keywords, wherein the department keywords are 'maintenance', 'emergency power supply', 'logistics', and the like.
S62, sequentially calling the first entity keywords of each first entity inference node in the entity inference set, comparing the department keywords with the first entity keywords, and determining the associated department corresponding to the first entity inference node as a first entity department.
According to the scheme, the first entity keywords of each first entity reasoning node in the entity reasoning set are sequentially called, wherein the first entity keywords are 'maintenance', 'emergency power supply', 'logistics', and then are compared with the department keywords to find the corresponding entity department as the first entity department.
Illustratively, the first entity key is "repair", then the first entity department ultimately determined may be repair A.
S63, second entity keywords of each second entity inference node in the entity inference set are sequentially called, the department keywords are compared with the second entity keywords, and the associated department corresponding to the second entity inference node is determined to serve as a second entity department.
The same as step S62, the present solution may find the corresponding entity department as the second entity department.
Illustratively, the second entity key is "logistics," then the first entity department ultimately determined may be a logistics C department.
And S64, recording the time respectively transmitted by the first entity department and the second entity department when the event label, the first time sequence information and the second time sequence information are transmitted to the first entity department and the second entity department, and obtaining the first transmission time and the second transmission time.
It can be understood that, in order to calculate the response time of each entity department, the first sending time and the second sending time when the event tag, the first timing information and the second timing information are sent to the first entity department and the second entity department are recorded.
On the basis of the above embodiment, the method further comprises S65-S67:
and S65, receiving feedback information sent by the management terminal, wherein the feedback information comprises a first processing time when the first entity department completes the processing task corresponding to the corresponding event label, and a second processing time when the second entity department completes the processing task corresponding to the corresponding event label.
It is understood that the first processing time and the second processing time are response times of corresponding entity departments.
And S66, obtaining a first response time period according to the first processing time and the first sending time, and obtaining a second response time period according to the second processing time and the second sending time.
It is understood that, by using the first and second transmission timings and the first and second processing timings, the first and second response time periods corresponding to the entity departments may be calculated.
S67, comparing the first response time period with first time sequence information, and comparing the second response time period with second time sequence information to obtain corresponding department training data.
For example, the first time sequence information may be 1 hour, and the first response time period is 3 hours, which means that the first response time period is greater than the first time sequence information, and the response speed of the department is slow, then the scheme may obtain corresponding department training data, train the department training data, and improve the subsequent response speed of the department.
In some embodiments, S67 (the comparing the first response time period with first timing information and the second response time period with second timing information to obtain corresponding department training information) includes:
s671, if the first response time period is less than the time requirement value of the first timing information, not generating a corresponding department training time.
It is understood that if the first response time period is less than the time requirement value of the first timing information, it indicates that the response time of the corresponding department is fast, and it is not necessary to train it.
And S672, if the first response time period is larger than the time requirement value of the first time sequence information, generating a first training requirement coefficient according to the first response time period and the time requirement value of the first time sequence information, and determining first training course time corresponding to the event label according to the first training requirement coefficient.
It is understood that if the first response time period is greater than the time requirement value of the first time sequence information, which indicates that the response speed of the department is slow, training needs to be performed on the department to improve the subsequent response speed of the department.
The scheme generates a first training requirement coefficient by using the first response time period and the time requirement value of the first time sequence information, and then calculates a first training course time corresponding to an event label (such as maintenance of a transformer).
It is understood that the greater the difference in the time information between the first response time period and the first timing information, the greater the corresponding first training requirement coefficient.
S673, if the second response time period is smaller than the time requirement value of the second time sequence information, no corresponding department training time is generated.
Similarly to step S671, if the second response time period is smaller than the time requirement value of the second timing information, it indicates that the response time of the corresponding department is fast, and it is not necessary to train it.
And S674, if the second response time period is larger than the time requirement value of the second time sequence information, generating a second training requirement coefficient according to the second response time period and the time requirement of the second time sequence information, and determining second training course time corresponding to the event label according to the second training requirement coefficient.
Similarly to step S672, if the second response time period is greater than the time requirement value of the second time series information, which indicates that the response speed of the department is slow, the department needs to be trained to improve the subsequent response speed of the department.
The scheme generates a second training requirement coefficient by using the second response time period and the time requirement of the second time sequence information, and then calculates a second training course time corresponding to the event label (such as emergency power supply).
It is understood that the larger the difference between the second response time period and the second timing information is, the larger the corresponding second training requirement coefficient is.
And S675, counting the first training course time of all the first entity departments and the second training course time of all the second entity departments, and generating department training data.
It is to be appreciated that after the first training course time and the second training course time are obtained, department training data can be generated according to the first training course time and the second training course time, and training can be performed on corresponding departments.
In some embodiments, calculating a first training lesson time for the first entity department and a second training lesson time for the second entity department by the following formulas specifically includes:
wherein, the first and the second end of the pipe are connected with each other,for a first training session time for a first entity department,in order to be the first moment of processing,is the first moment of time of transmission,is a time requirement value of the first timing information,is a value that is normalized for the time,it is the second moment of the processing,is the second moment of time of transmission,is a time requirement value of the second timing information,for a second training session time for a second entity department,training the curriculum time for the benchmark.
In the above-mentioned formula,which represents a first period of time in response to the first,a difference value representing the first response time period and corresponding time information of the first time sequence information, the larger the difference value is, the corresponding first training course time of the first entity departmentThe higher the need; in the same way, the method has the advantages of,which represents the second period of time of the response,a difference value of the corresponding time information representing the second response time period and the second time sequence information, wherein the larger the difference value is, the time of the second training course of the corresponding second entity departmentThe higher the need.
In addition to the above embodiments, the present invention may have other embodiments; all technical solutions formed by adopting equivalent substitutions or equivalent transformations fall within the protection scope of the claims of the present invention.
Claims (9)
1. The entity generation method based on the event chain inference of the event graph is characterized by comprising the following steps:
the method comprises the steps of S1, determining an initial event node corresponding to a current power emergency in a matter graph, and acquiring all first event connection nodes which are directly connected with the initial event node in the matter graph, wherein the first event connection nodes are any one of event nodes or entity nodes;
s2, if the first event connection node is an entity node, the first event connection node is used as a first entity inference node;
s3, if the first event connection node is an event node, taking the first event connection node as a first event inference node, and generating a first event chain according to the first event inference node and corresponding event logic;
s4, taking other event nodes except the initial event node and the first event inference node in the first event chain as second event inference nodes, determining the connected first entity inference nodes according to the first event inference nodes, and determining the connected second entity inference nodes according to the second event inference nodes;
s5, adding first time sequence information to all the first entity reasoning nodes, adding second time sequence information to all the second entity reasoning nodes, and generating an entity reasoning set according to all the first entity reasoning nodes and the second entity reasoning nodes, wherein the first time sequence information is smaller than the second time sequence information;
s6, acquiring event tags and main body information in the power emergency, determining a first entity department corresponding to a first entity inference node in an entity inference set and a second entity department corresponding to a second entity inference node according to the main body information, sending the corresponding event tags and the first time sequence information to the corresponding first entity department, and sending the corresponding event tags and the second time sequence information to the corresponding second entity department;
the S3 comprises the following steps:
taking the initial event node as a label 0 point, and sequentially labeling the first event inference node and other event nodes connected with the first event inference node in an ascending order under the extension of each event logic;
acquiring the number of the event logics and the event grade corresponding to the initial event node, and calculating to obtain the number of labels according to the number of the event logics and the event grade;
after judging that the labels corresponding to the event nodes correspond to the number of the labels, stopping labeling the event nodes, and taking the event nodes with the labels corresponding to the number of the preset labels as cut-off points of corresponding event logics;
when the number of the labels corresponding to the event nodes is judged not to correspond to the number of the labels and the event node under the extension of the corresponding event logic is the last one, taking the last event node as the cut-off point of the corresponding event logic;
and taking the point labeled as 0 as a starting point, and counting event nodes corresponding to the starting point, the cut-off point and a middle point from the starting point to the cut-off point under each event logic to generate a first event chain.
2. The method for entity generation based on event graph chain inference as claimed in claim 1,
the S1 comprises:
if the user inputs the power emergency, performing word segmentation processing on the power emergency to obtain a plurality of emergency keywords;
extracting an event keyword classification table corresponding to the event map, wherein the event keyword classification table is provided with preset keywords corresponding to each event node;
and comparing the emergency keywords with preset keywords, taking event nodes corresponding to the power emergency as initial event nodes, and taking all nodes directly connected with the initial event nodes in the event graph as first event connection nodes.
3. The method for entity generation based on event graph chain inference of claim 1,
the obtaining of the number of the event logics and the event grades corresponding to the initial event nodes and the calculating of the number of labels according to the number of the event logics and the event grades comprise:
each initial event node has an event rank of a standard magnitude set in advance thereto, and the index value is calculated by the following formula,
wherein the content of the first and second substances,the numerical values of the reference numerals are given,as to the amount of the event logic that is to be executed,is a standard quantity of event logic that is,the value is normalized for the quantity,an event rating of a standard magnitude;
displaying the label numerical value, and if confirmation information is received, taking the calculated label numerical value as the number of labels;
and if the modification information is received, taking a modification value corresponding to the modification information as the final label number, and adjusting the event grade of the standard quantity value according to the modification value and the label number to obtain the event grade of the non-standard quantity value.
4. The method for entity generation based on event graph chain inference as claimed in claim 3,
if receiving the modification information, taking a modification value corresponding to the modification information as a final label number, and adjusting the event grade of the standard quantity value according to the modification value and the label number value to obtain an event grade of a non-standard quantity value, including:
if the modified numerical value is larger than the label numerical value, obtaining an enlargement coefficient according to the difference value of the modified numerical value and the label numerical value, and carrying out forward adjustment on the event grade of the standard numerical value according to the enlargement coefficient to obtain the event grade of the non-standard numerical value after the forward adjustment;
if the modified numerical value is smaller than the label numerical value, obtaining a turn-down coefficient according to the difference value of the modified numerical value and the label numerical value, and carrying out negative adjustment on the event grade of the standard numerical value according to the turn-down coefficient to obtain the event grade of the non-standard numerical value after negative adjustment;
the event rating of the adjusted non-standard magnitude is calculated by the following formula,
wherein, the first and the second end of the pipe are connected with each other,for a forward adjusted event level of non-standard magnitude,in order to modify the value of the numerical value,in order to adjust the weight value in the forward direction,for a negatively adjusted event rating of non-standard magnitude,the weight value is adjusted for the negative direction.
5. The method for entity generation based on event graph chain inference as claimed in claim 1,
the S4 comprises the following steps:
taking event nodes corresponding to other label points except the label 0 point and the label 1 point in each first event chain as second event reasoning nodes;
if the entity node directly connected with the first event inference node exists, taking the entity node directly connected with the first event inference node as the first entity inference node;
and if the entity node directly connected with the second event inference node exists, taking the entity node directly connected with the second event inference node as the second entity inference node.
6. The method for entity generation based on event graph chain inference of claim 1,
the S6 comprises the following steps:
determining all corresponding associated departments according to the main body information, and performing word segmentation processing on the associated departments to obtain department keywords;
sequentially calling a first entity keyword of each first entity reasoning node in the entity reasoning set, comparing the department keyword with the first entity keyword, and determining an associated department corresponding to the first entity reasoning node as a first entity department;
sequentially calling a second entity keyword of each second entity reasoning node in the entity reasoning set, comparing the department keyword with the second entity keyword, and determining an associated department corresponding to the second entity reasoning node as a second entity department;
when the event label, the first time sequence information and the second time sequence information are sent to the first entity department and the second entity department, the time sent by the first entity department and the time sent by the second entity department are recorded, and the first sending time and the second sending time are obtained.
7. The method for entity generation based on event graph event chain inference as claimed in claim 6, further comprising:
receiving feedback information sent by a management end, wherein the feedback information comprises a first processing time when a first entity department completes a processing task corresponding to a corresponding event label and a second processing time when a second entity department completes the processing task corresponding to the corresponding event label;
obtaining a first response time period according to the first processing time and the first sending time, and obtaining a second response time period according to the second processing time and the second sending time;
and comparing the first response time period with first time sequence information, and comparing the second response time period with second time sequence information to obtain corresponding department training data.
8. The method for entity generation based on event graph chain inference of claim 7,
the comparing the first response time period with the first time sequence information and the comparing the second response time period with the second time sequence information to obtain corresponding department training information includes:
if the first response time period is smaller than the time requirement value of the first time sequence information, no corresponding department training time is generated;
if the first response time period is larger than the time requirement value of the first time sequence information, generating a first training requirement coefficient according to the first response time period and the time requirement value of the first time sequence information, and determining first training course time corresponding to an event label according to the first training requirement coefficient;
if the second response time period is smaller than the time requirement value of the second time sequence information, no corresponding department training time is generated;
if the second response time period is larger than the time requirement value of the second time sequence information, generating a second training requirement coefficient according to the second response time period and the time requirement of the second time sequence information, and determining second training course time corresponding to the event label according to the second training requirement coefficient;
and counting the first training course time of all the first entity departments and the second training course time of all the second entity departments to generate department training data.
9. The method for entity generation based on event graph chain inference of claim 8,
calculating a first training course time of a first entity department and a second training course time of a second entity department by the following formulas, which specifically comprises the following steps:
wherein the content of the first and second substances,for a first training session time for a first entity department,in order to be the first moment of processing,is the first moment of time of transmission,is a time requirement value of the first timing information,is a value that is normalized for the time,in order to be the second moment of processing,is the second moment of time of transmission,is a time requirement value of the second timing information,for a second training session time for a second entity department,training the curriculum time for the benchmark.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210848876.6A CN114969382B (en) | 2022-07-19 | 2022-07-19 | Entity generation method based on event chain inference of event graph |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210848876.6A CN114969382B (en) | 2022-07-19 | 2022-07-19 | Entity generation method based on event chain inference of event graph |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114969382A CN114969382A (en) | 2022-08-30 |
CN114969382B true CN114969382B (en) | 2022-10-21 |
Family
ID=82969593
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210848876.6A Active CN114969382B (en) | 2022-07-19 | 2022-07-19 | Entity generation method based on event chain inference of event graph |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114969382B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115630846B (en) * | 2022-12-07 | 2023-08-01 | 速度科技股份有限公司 | Processing method suitable for natural disaster risk comprehensive monitoring data |
CN116467468B (en) * | 2023-05-05 | 2024-01-05 | 国网浙江省电力有限公司 | Power management system abnormal information handling method based on knowledge graph technology |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114281972A (en) * | 2021-12-28 | 2022-04-05 | 亿迅信息技术有限公司 | Dialog control method, system storage medium and server based on subject object tracking and cognitive inference |
CN114417004A (en) * | 2021-11-10 | 2022-04-29 | 南京邮电大学 | Method, device and system for fusing knowledge graph and case graph |
WO2022134794A1 (en) * | 2020-12-22 | 2022-06-30 | 深圳壹账通智能科技有限公司 | Method and apparatus for processing public opinions about news event, storage medium, and computer device |
CN114707004A (en) * | 2022-05-24 | 2022-07-05 | 国网浙江省电力有限公司信息通信分公司 | Method and system for extracting and processing case-affair relation based on image model and language model |
CN114722974A (en) * | 2022-06-07 | 2022-07-08 | 国网浙江省电力有限公司信息通信分公司 | Multi-dimensional map fusion method based on matter logic and entity knowledge |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111784192A (en) * | 2020-07-17 | 2020-10-16 | 塔盾信息技术(上海)有限公司 | Industrial park emergency plan executable system based on dynamic evolution |
CN112215458A (en) * | 2020-09-01 | 2021-01-12 | 青岛海信网络科技股份有限公司 | Disaster analysis method and electronic device |
-
2022
- 2022-07-19 CN CN202210848876.6A patent/CN114969382B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022134794A1 (en) * | 2020-12-22 | 2022-06-30 | 深圳壹账通智能科技有限公司 | Method and apparatus for processing public opinions about news event, storage medium, and computer device |
CN114417004A (en) * | 2021-11-10 | 2022-04-29 | 南京邮电大学 | Method, device and system for fusing knowledge graph and case graph |
CN114281972A (en) * | 2021-12-28 | 2022-04-05 | 亿迅信息技术有限公司 | Dialog control method, system storage medium and server based on subject object tracking and cognitive inference |
CN114707004A (en) * | 2022-05-24 | 2022-07-05 | 国网浙江省电力有限公司信息通信分公司 | Method and system for extracting and processing case-affair relation based on image model and language model |
CN114722974A (en) * | 2022-06-07 | 2022-07-08 | 国网浙江省电力有限公司信息通信分公司 | Multi-dimensional map fusion method based on matter logic and entity knowledge |
Non-Patent Citations (2)
Title |
---|
事件知识图谱构建技术与应用综述;项威;《计算机与现代化》;20200115(第01期);全文 * |
基于应急实例本体模型的应急案例推理方法;蔡玫等;《情报杂志》;20160618(第06期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN114969382A (en) | 2022-08-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114969382B (en) | Entity generation method based on event chain inference of event graph | |
CN108764707A (en) | A kind of data assessment system and method | |
US6490565B1 (en) | Environmental certification system and method | |
WO2023221343A1 (en) | Safety hazard information code generation method based on violation portraits and safety portraits of electric power members | |
CN115730749B (en) | Power dispatching risk early warning method and device based on fusion power data | |
CN110471898A (en) | Dissemination method can be traced in a kind of information credit management method and Figures | |
CN113434575A (en) | Data attribution processing method and device based on data warehouse and storage medium | |
Abdelgawad et al. | Accessibility of Norwegian municipalities websites: a qualitative system dynamics approach | |
Adderly | Reviewing power outage trends, electric reliability indices and smart grid funding | |
CN113902596B (en) | Post-class service method and system using information matching | |
CN108132976B (en) | Emergency handling method and system for railway station emergency | |
Bandyopadhayay et al. | Forecasting day-ahead price of electricity–a dynamic regression approach | |
De Dios et al. | Just how good is unemployment as a measure of welfare? A note | |
CN116596462B (en) | Enterprise management normative research and development project management system based on block chain | |
Asrari et al. | A stochastic hybrid method to forecast operating reserve: comparison of fuzzy and classical set theory | |
Roe et al. | Risk assessment and management for interconnected critical infrastructure systems at the site and regional levels in California's Sacramento-San Joaquin Delta | |
GUNAWAN et al. | The analysis of extended producer responsibility (EPR) for e-waste management Policy Drivers and Challenges in Singapore | |
Gillingham et al. | Using Behavioral Science to Target LMI and High-Value Solar Installations (Final Technical Report) | |
Bollinger | Public relations, business and the press | |
Salleh et al. | RENEWABLE ENERGY ADOPTION FOR ENERGY COMPLEMENTARY: APPLICATION OF OFFSHORE WIND TURBINES IN SABAH, MALAYSIA. | |
Yuliawati et al. | GENRE Campaign, Youth and Instagram | |
JP2022138957A (en) | Prediction system and prediction method | |
Gessese | The Use of Big Data in Education: Challenges and Policy Recommendations for Ethiopia | |
CN113450222A (en) | Block chain technology-based method for managing deposit duration bonds | |
CN110097483A (en) | A kind of shield training platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |