CN116402137B - Satellite task mining prediction method, system and device based on public information - Google Patents
Satellite task mining prediction method, system and device based on public information Download PDFInfo
- Publication number
- CN116402137B CN116402137B CN202310646855.0A CN202310646855A CN116402137B CN 116402137 B CN116402137 B CN 116402137B CN 202310646855 A CN202310646855 A CN 202310646855A CN 116402137 B CN116402137 B CN 116402137B
- Authority
- CN
- China
- Prior art keywords
- model
- event
- information
- entity
- task
- 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 50
- 238000005065 mining Methods 0.000 title claims abstract description 35
- 238000000605 extraction Methods 0.000 claims abstract description 23
- 239000013598 vector Substances 0.000 claims description 32
- 238000004364 calculation method Methods 0.000 claims description 28
- 230000001364 causal effect Effects 0.000 claims description 28
- 238000012216 screening Methods 0.000 claims description 14
- 230000015654 memory Effects 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 12
- 238000004140 cleaning Methods 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 238000013145 classification model Methods 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 9
- 238000013075 data extraction Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000003058 natural language processing Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 210000005036 nerve Anatomy 0.000 claims description 5
- 238000003062 neural network model Methods 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000009193 crawling Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 239000000284 extract Substances 0.000 claims description 3
- 230000008520 organization Effects 0.000 claims description 3
- 238000012790 confirmation Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 9
- 238000010801 machine learning Methods 0.000 description 6
- 230000002093 peripheral effect Effects 0.000 description 3
- 230000002457 bidirectional effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000001965 increasing effect Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
- G06N5/025—Extracting rules from data
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
-
- 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)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a satellite task mining prediction method, a system and a device based on public information, which relate to the technical field of satellite task planning, and mainly comprise the following steps: based on public information and a domain word stock in the Internet, extracting structural information about a target event through an extraction model; generating a task association relation through a relation model based on the structured information and the event knowledge graph; and predicting the satellite observation task through a prediction model based on the task association relation and a satellite observation field knowledge base. According to the scheme, the satellite tasks are rapidly extracted from the Internet dynamic information as targets, event evolution reasoning is realized by utilizing event relations, automatic mining and prediction of satellite observation tasks are supported, auxiliary support is provided for subsequent manual confirmation, the scheme allows a user to make decision in advance to arrange related observation tasks, and reasonable, stable and ordered satellite observation plans are ensured.
Description
Technical Field
The invention relates to the technical field of satellite task planning, in particular to a satellite task mining prediction method, system and device based on public information.
Background
With the rapid development of internet technology, the speed of information disclosure is also increasing. Historical data suggests that certain public event scenarios tend to have a high degree of relevance to satellite observation tasks.
Meanwhile, with the continuous increase of the number of satellite tasks, the satellite task scheduling method is difficult to coordinate with the number of satellites, the traditional satellite scheduling method cannot meet the continuous increase of user task planning demands, and automatic mining prediction of satellite tasks is becoming more and more urgent. These predictions mainly include predictions of the total number of users in a future time window, as well as the corresponding task type, observation target, allowed time window, observation target type, number of observations targets, observation target mobility, observation feature elements, observation feature requirements, and user priorities, etc.
In addition, the current satellite task report has the problems of poor rationality, excessively high requirements for users and the like, and the problem of low satellite observation efficiency is further highlighted after the number of tasks is greatly increased.
Therefore, how to quickly mine and predict satellite tasks based on the disclosed information, so as to improve the observation efficiency and the observation quality is a big subject to be solved urgently.
Disclosure of Invention
The invention aims to provide a satellite task mining prediction method, system and device based on public information, so as to solve at least one of the technical problems in the prior art.
In order to solve the technical problems, the invention provides a satellite task mining prediction method based on public information, which comprises the following steps:
step 1, based on public information and a domain word stock in the Internet, extracting structural information about a target event through an extraction model; the extraction model comprises a first model and a second model;
step 2, generating a task association relation through a relation model based on the structured information and the event knowledge graph; the relation model comprises a third model and a fourth model;
step 3, predicting a satellite observation task through a prediction model based on the task association relation and a satellite observation field knowledge base; the predictive model includes a fifth model and a sixth model.
Through the steps, the hot events can be dynamically extracted from the public information such as the Internet, the satellite observation tasks related to the hot events can be automatically predicted, and the satellite tasks can be laid out in advance, so that the observation efficiency is improved.
In a possible embodiment, the step 1 includes:
step 11, automatically grabbing public data in a preset range at fixed time through a crawling program based on public information in the Internet, such as an Internet page;
step 12, for the public data, removing invalid information through data cleaning; the data cleaning belongs to the prior art, and comprises a box division method, a clustering method and a regression method, and can be used for eliminating noise;
step 13, performing word segmentation on the public data after data cleaning by a natural language processing method; the natural language processing method belongs to the prior art and comprises the processes of marking, deleting stop words, extracting trunks, embedding words and the like;
step 14, carrying out data filtering on the segmented public data based on the domain word stock, and screening out data meeting the domain requirements;
step 15, screening events consistent with the satellite observation task type from the data-filtered public data through a first model; the first model is a text classification model, the segmentation is carried out by taking the words of the event which are consistent with the satellite observation task type as granularity, for example, the text classification model is a text classification model capable of being shifted and learned, the text classification model belongs to the prior art, and the preliminary classification of the event is carried out by a hotword or high-frequency method;
Step 16, extracting the event through a second model for the event to obtain structural information; the second model is a pre-training model, the pre-training model belongs to the prior art, and fine-granularity event identification, event element extraction and event relation extraction can be performed, wherein the event elements comprise time, place, targets and the like.
Through the steps, key information disclosed in internet media pages such as a hot spot website, a hot spot forum and the like can be crawled, and after preprocessing, classification and recognition, events consistent with the satellite observation task types are extracted.
In a possible embodiment, the step 2 includes:
step 21, matching the target event with the historical similar event based on the structural information and the event knowledge graph;
and 22, carrying out event association query analysis based on the logic links in the event knowledge graph to generate a task association relation for the target event.
Through the steps, the task association relation of the current structural information can be generated from two aspects of the historical similar events and the event logic links, so that the follow-up prediction is facilitated.
In a possible embodiment, the step 21 includes:
Step 211, aiming at the structured information, carrying out embedded representation through a third model to obtain a word vector of a target event; the third model is a pre-training model, for example, a pre-training model represented by a bi-directional encoder; the embedded representation, also called word embedding, is a learning representation of text, belongs to the prior art, wherein words with the same meaning have similar representation forms, and can be represented as real value vectors in a predefined vector space;
step 212, obtaining the context characteristics of the target event through a fourth model based on the embedded and represented structured information; the fourth model is a neural network model, for example, a BiLSTM bidirectional circulating neural network model, belongs to the prior art, comprises 2 independent LSTM, inputs an input sequence into 2 LSTM neural networks in positive sequence and reverse sequence respectively for feature extraction, and takes a word vector formed by splicing 2 output vectors as a final feature expression of the word;
step 213, matching the target event with the historical similar event in the event knowledge graph through similarity calculation based on the word vector, the context feature and the event knowledge graph, thereby realizing entity link between the structured information and the event knowledge graph; the similarity calculation belongs to the prior art, such as cosine similarity, and the similarity is represented by calculating the distance between two vectors, and the closer the distance is, the larger the similarity is.
Through the steps, the historical events similar to the target event can be automatically matched according to the event similarity in a machine learning mode.
The event knowledge graph belongs to the prior art, comprises basic elements such as event nodes, event descriptions, event association and the like, and can project relationship network information to an event graph layer so as to abstract complex networks such as event descriptions, event storage, event prediction, event discovery, event relationships and the like. The relation among event nodes forms an event logic link, comprising a causal link and a compliant link, and is used as the basis of event reasoning.
In a possible embodiment, the step 22 includes:
step 221, constructing a participant P, a type T, a trigger word W and the like of an event as event description, and matching structural information of a target event to a most similar historical event node in an event knowledge graph through similarity calculation to obtain a trigger word SW and an entity list L of the corresponding most similar historical event of the target event;
step 222, based on the trigger word SW and the entity list L, reasoning is performed through a plurality of causal relationships in the causal link and a plurality of compliant relationships in the compliant link, so as to obtain a logical link reasoning result as a task association relationship; the logic link reasoning result comprises a causal link reasoning result and a compliant link reasoning result; the causal link reasoning result is entity information obtained in the entity list L after causal reasoning is performed according to the trigger word SW; the forward link reasoning result refers to entity information obtained in the entity list L after forward reasoning is carried out according to the trigger word SW.
By the method, reasoning can be carried out from two aspects of a causal logic link and a compliant logic link, the result possibly caused by the historical similar event is matched to the target event, and the task association relationship is generated, so that the accuracy of the task prediction of the subsequent satellite is ensured, and further task mining can be carried out on the basis of the task association relationship.
In a possible embodiment, the step 3 includes:
step 31, expanding an observation knowledge graph through a fifth model based on a knowledge base in the satellite observation field and a task association relation; the satellite observation field knowledge base comprises a task type, an observation target, an allowed time window, an observation target type, an observation target mobility, an observation feature element, an observation feature requirement, a user preference, a priority and the like;
step 32, predicting satellite observation tasks through a sixth model based on task association relations and the observation knowledge graph; the sixth model is a graph neural network and comprises a knowledge representation embedding layer, a characteristic convolution layer, an attention mechanism layer and an associated information prediction layer.
Through the steps, the associated information between the target event and the knowledge in the satellite observation field is mined, and the satellite observation task is predicted and generated.
In a possible embodiment, the step 31 includes:
step 311, identifying an entity mention set M from an entity set E and a text set D in a knowledge base of the satellite observation field;
step 312, mapping each entity mention mε M to a corresponding entity eE through a fifth model; if a certain m cannot be mapped to e, defining m as unlink; the fifth model includes a candidate entity generation model, a candidate entity ranking model, and an unlink mention prediction model.
Through the steps, link disambiguation can be completed for the entity, and the link between the entity in the target event and the entity in the knowledge base in the satellite observation field is standardized, so that the consistency of data is ensured, and the complete information of the target event is acquired from the knowledge base in the satellite observation field.
In one possible embodiment, the step 312 includes:
step 3121, inputting a certain entity mention M in the entity mention set M into the candidate entity generation model, and searching a plurality of candidate entities most similar to the M from the entity set E based on a dictionary to obtain a candidate entity set Em; the dictionary comprises a synonym dictionary and an ambiguous word dictionary; the synonym dictionary is used for mapping entity references into canonical entities existing in a knowledge base of the satellite observation field; the ambiguous word dictionary is used for obtaining an initial candidate entity set of entity names;
Step 3122, forming scoring entity pairs by each candidate entity in m and Em, inputting the scoring entity pairs into the candidate entity ranking model, and scoring and ranking each entity pair by the model according to the surface similarity and the context feature similarity;
step 3123, based on the unlink mention prediction model, screening out entity pairs with highest scores by an 'unlink mention' prediction algorithm, and defining the entity mention m in the entity pair as the best entity in the entity set E and the rest entity mention as unlink; the 'unlink mention' prediction algorithm belongs to the prior art and is a method for predicting unlink mention in a knowledge graph.
By the method, the best matching entity of the target event can be put into the entity set, so that ambiguity among entity links is eliminated.
In a possible embodiment, the step 32 includes:
step 321, carrying out vector organization on attribute information, constraint information and auxiliary information of a target entity in a task association relation and observation knowledge graph, and converting the attribute information, the constraint information and the auxiliary information into corresponding multidimensional matrixes through feature vector calculation; the feature vector calculation belongs to the prior art and is used for calculating the feature vector;
Step 322, performing graph nerve convolution calculation on the multidimensional matrix to obtain semantic representation, so as to enhance semantic features; the graph nerve convolution calculation belongs to the prior art, and is a matrix operation method;
step 323, calculating the attention of a certain target entity to the auxiliary information through an attention calculation rule, wherein the attention calculation rule is used for enhancing knowledge association; the attention calculation rule belongs to the prior art and is a data processing method in machine learning;
step 324, reserving a plurality of associated information record combinations to generate an associated information candidate set so as to reduce the operation amount;
step 325, mining out the associated information candidate entity of the target event by the multi-layer perceptron based on the associated information candidate set, and predicting to obtain a satellite observation task; the satellite observation tasks comprise task types, observation targets, allowed time windows, observation target types, observation target mobility, observation feature elements, observation feature requirements, user preferences, priorities and the like; the multi-layer perceptron belongs to the prior art and is a supervised machine learning algorithm.
Through the steps, satellite task observation elements in the knowledge base of the satellite observation field can be matched to the target event, so that satellite observation tasks aiming at the event can be predicted.
The application also provides a satellite task mining prediction system based on public information based on the same inventive concept, which comprises a data extraction module, a data processing module and a result generation module:
the data extraction module comprises an extraction model and a domain word stock, and extracts structural information about a target event based on public information in the Internet; the extraction model comprises a first model and a second model; the first model is used for screening out events consistent with the satellite observation task types; the second model is used for carrying out event extraction and generating structural information;
the data processing module comprises a database, an association relation unit and an observation task unit:
the database is used for storing event knowledge maps, a satellite observation field knowledge base and observation knowledge maps;
the association relation unit stores a relation model; the relation model generates a task association relation based on the structured information and an event knowledge graph; the relationship model includes a third model and a fourth model: the third model is used for carrying out embedded representation on the structured information to obtain a word vector of the target event; the fourth model is used for acquiring the context characteristics of the target event;
The observation task unit is stored with a prediction model; the prediction model predicts satellite observation tasks based on the task association relation; the prediction model includes a fifth model and a sixth model: the fifth model is used for mapping each entity mention in the knowledge base of the satellite observation field to a corresponding entity, expanding the observation knowledge graph, and comprises a candidate entity generation model, a candidate entity ranking model and an unlink mention prediction model; the sixth model is a graph neural network and is used for mining out the associated information candidate entity of the target event from the observation knowledge graph and predicting to obtain a satellite observation task;
and the result generation module is used for issuing the satellite observation task.
In a third aspect, based on the same inventive concept, the present application further provides a satellite task mining prediction device based on public information, including a processor, a memory, and a bus, where the memory stores instructions and data readable by the processor, and the processor is configured to call the instructions and data in the memory to execute any one of the satellite task mining prediction methods based on public information as described above, and the bus connects functional components to transfer information therebetween.
By adopting the technical scheme, the invention has the following beneficial effects:
according to the satellite task mining prediction method, system and device based on public information, which are provided by the invention, the satellite task is rapidly extracted by using the Internet dynamic information as a target, an evolvable knowledge graph comprising event compliance relations and event causal relations in the satellite observation field is constructed, accurate and comprehensive fact knowledge supply oriented to emergency prediction is realized, event evolution reasoning is realized by using the event relations, automatic mining and prediction of the satellite observation task are supported, and auxiliary support is provided for subsequent manual confirmation. The scheme allows the user to make a decision in advance to arrange related observation tasks, not only prevents the user tasks from being unable to be arranged due to resource shortage, but also prevents the existing satellite observation plans from being greatly adjusted after the user tasks arrive, thereby ensuring the reasonable, stable and orderly satellite observation plans.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings which are required in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are some embodiments of the invention and that other drawings may be obtained from these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a frame diagram of a satellite task mining prediction method based on public information according to an embodiment of the present invention;
FIG. 2 is a flowchart of a satellite task mining prediction method based on public information according to an embodiment of the present invention;
FIG. 3 is a frame diagram of step 1 according to an embodiment of the present invention;
FIG. 4 is a frame diagram of step 2 according to an embodiment of the present invention;
FIG. 5 is a block diagram of step 22 according to an embodiment of the present invention;
FIG. 6 is a frame diagram of step 3 according to an embodiment of the present invention;
FIG. 7 is a block diagram of step 31 according to an embodiment of the present invention;
FIG. 8 is a block diagram of step 32 according to an embodiment of the present invention;
fig. 9 is a diagram of a satellite task mining prediction system based on public information according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made apparent and fully in view of the accompanying drawings, in which some, but not all embodiments of the invention are shown. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
For the convenience of description of the specific embodiments, the specific concepts of the present invention will now be described as follows:
in order to better utilize and schedule satellites, from the standpoint of improving efficiency, it is considered to predict user tasks in advance. Particularly, for tasks with relaxed requirements on time windows, the tasks can be interacted with a user in advance to implement advanced operation, so that the situation that the tasks cannot be met due to resource shortage is avoided. In addition, the method provides possible operation choices for the user in advance, achieves the purpose of active observation of the system, increases the user friendliness of the system and is convenient for the user to operate. Therefore, intelligent prediction of tasks is essential for improving the efficiency of a constellation scheduling system.
The scheme solves the problems of automatic prediction and advanced triggering of the observation task from the aspect of the dynamic hot spot event of the Internet, and realizes the assumption that satellite tasks are automatically laid out in advance according to user behaviors. The above problem is first subdivided into three sub-problems: question 1, how to identify internet information and generate structured information; question 2, how to complete task association relation generation based on the structured information; and 3, completing satellite task mining and prediction based on the association relation. And then solving each problem one by one, as shown in a frame diagram in fig. 1, firstly identifying internet information from open source data to generate structural information, then generating task association relation based on the structural information, and finally carrying out task mining prediction based on the task association relation. The satellite observation related knowledge graph provides knowledge support for task association relation generation and task mining prediction.
The invention is further illustrated with reference to specific embodiments.
It should be further noted that the following specific examples or embodiments are a series of optimized arrangements of the present invention for further explaining specific summary, and all the arrangements may be combined or used in association with each other.
Embodiment one:
as shown in fig. 2, the satellite task mining prediction method based on public information provided in this embodiment includes the following steps:
step 1, based on public information and a domain word stock in the Internet, extracting structural information about a target event through an extraction model; the extraction model comprises a first model and a second model;
step 2, generating a task association relation through a relation model based on the structured information and the event knowledge graph; the relation model comprises a third model and a fourth model;
step 3, predicting a satellite observation task through a prediction model based on the task association relation and a satellite observation field knowledge base; the predictive model includes a fifth model and a sixth model.
Through the steps, the hot events can be dynamically extracted from the public information such as the Internet, the satellite observation tasks related to the hot events can be automatically predicted, and the satellite tasks can be laid out in advance, so that the observation efficiency is improved.
Further, as shown in fig. 3, the step 1 includes:
step 11, automatically capturing public data in a preset range, such as keywords, focus attention words, titles and the like, at fixed time by a crawling program, such as Scrapy, based on public information, such as internet pages of a hot spot website, a hot spot forum and the like;
Step 12, for the public data, removing invalid information through data cleaning; the data cleaning belongs to the prior art, and comprises a box division method, a clustering method and a regression method, and can be used for eliminating noise;
step 13, performing word segmentation on the public data after data cleaning by a natural language processing method; the natural language processing method belongs to the prior art and comprises the processes of marking, deleting stop words, extracting trunks, embedding words and the like;
step 14, carrying out data filtering on the segmented public data based on the domain word stock, and screening out data meeting the domain requirements;
step 15, screening events consistent with the satellite observation task type from the data-filtered public data through a first model; the first model is a text classification model capable of transfer learning, the text classification model is segmented by taking words of events consistent with the satellite observation task types as granularity, the text classification model belongs to the prior art, and the preliminary classification of the events is carried out through a hot focus word or high-frequency method;
step 16, extracting the event through a second model for the event to obtain structural information; the second model is a pre-training model, and the pre-training model belongs to the prior art, and can perform fine-granularity event identification, event element extraction and event relation extraction, wherein the event elements comprise time, place, target and the like, for example, the time is as follows: 2023, 1, 27; location: a fire area.
Through the steps, key information disclosed in internet media pages such as a hot spot website, a hot spot forum and the like can be crawled, and after preprocessing, classification and recognition, events consistent with the satellite observation task types are extracted.
Further, as shown in fig. 4, the step 2 includes:
step 21, matching the target event with the historical similar event based on the structured information and the event knowledge graph, for example, matching the satellite observation target E2 when the fire disaster happens to the target event E1 in a certain place with the fire disaster happens to other regions;
step 22, carrying out event association query analysis based on a logic link in an event knowledge graph to generate a task association relation for a target event, wherein the logic link comprises a causal link and a compliant link; the causal link refers to a historical event E3 obtained by causal reasoning, for example, a resident point is dangerous after a fire disaster occurs, and the resident point needs to be observed, so that an observation demand R1 observes nearby resident points; the compliant link refers to a historical event E4 obtained by carrying out compliant reasoning, for example, when a fire disaster occurs, other flammable areas need to be observed due to high ambient temperature, so that a nearby flammable point is observed by the investigation requirement R2.
Through the steps, the task association relation of the current structural information can be generated from two aspects of the historical similar events and the event logic links, so that the follow-up prediction is facilitated.
Further, the step 21 includes:
step 211, aiming at the structured information, carrying out embedded representation through a third model to obtain a word vector of a target event; the third model is a pre-training model represented by a bi-directional encoder; the embedded representation, also called word embedding, is a learning representation of text, belongs to the prior art, wherein words with the same meaning have similar representation forms, and can be represented as real value vectors in a predefined vector space;
step 212, obtaining the context characteristics of the target event through a fourth model based on the embedded and represented structured information; the fourth model is a BiLSTM bidirectional cyclic neural network model, belongs to the prior art, and comprises 2 independent LSTM, wherein an input sequence is respectively input into 2 LSTM neural networks in a positive sequence and a reverse sequence for feature extraction, and word vectors formed by splicing 2 output vectors are used as final feature expression of the word;
step 213, matching the target event with the historical similar event in the event knowledge graph through similarity calculation based on the word vector, the context feature and the event knowledge graph, thereby realizing entity link between the structured information and the event knowledge graph; the similarity calculation belongs to the prior art, such as cosine similarity, and the similarity is represented by calculating the distance between two vectors, and the closer the distance is, the larger the similarity is.
Through the steps, the historical events similar to the target event can be automatically matched according to the event similarity in a machine learning mode.
The event knowledge graph belongs to the prior art, comprises basic elements such as event nodes, event descriptions, event association and the like, and can project relationship network information to an event graph layer so as to abstract complex networks such as event descriptions, event storage, event prediction, event discovery, event relationships and the like. The relation among event nodes forms an event logic link, comprising a causal link and a compliant link, and is used as the basis of event reasoning.
Further, as shown in fig. 5, the step 22 includes:
step 221, constructing a participant P, a type T, a trigger word W and the like of an event as event description, and matching structural information of a target event to a most similar historical event node in an event knowledge graph through similarity calculation to obtain a trigger word SW and an entity list L of the corresponding most similar historical event of the target event;
step 222, based on the trigger word SW and the entity list L, reasoning is performed through a plurality of causal relationships in the causal link and a plurality of compliant relationships in the compliant link, so as to obtain a logical link reasoning result as a task association relationship; the logic link reasoning result comprises a causal link reasoning result and a compliant link reasoning result; the causal link reasoning result is entity information obtained in the entity list L after causal reasoning is performed according to the trigger word SW; the forward link reasoning result refers to entity information obtained in the entity list L after forward reasoning is carried out according to the trigger word SW.
By the method, reasoning can be carried out from two aspects of a causal logic link and a compliant logic link, the result possibly caused by the historical similar event is matched to the target event, and the task association relationship is generated, so that the accuracy of the task prediction of the subsequent satellite is ensured, and further task mining can be carried out on the basis of the task association relationship.
Further, as shown in fig. 6, the step 3 includes:
step 31, expanding an observation knowledge graph through a fifth model based on a knowledge base in the satellite observation field and a task association relation; the content of the knowledge base in the satellite observation field comprises task types, observation targets, allowed time windows, observation target types, observation target mobility, observation feature elements, observation feature requirements, user preferences, priorities and the like; the ontology modes of the knowledge base in the satellite observation field comprise M1 (entity 1, attribute relationship, entity 2), M2 (entity 2, association, entity 3), M3 (entity 3-attribute relationship-entity 4) and the like;
step 32, predicting satellite observation tasks through a sixth model based on task association relations and the observation knowledge graph; the sixth model is a graph neural network, which comprises a knowledge representation embedded layer, a characteristic convolution layer, an attention mechanism layer and an associated information prediction layer, and belongs to the prior art, and prediction operation can be performed through setting and training.
Through the steps, the entity linking technology is utilized to carry out matching calculation on the attention elements and the knowledge graph in the event, the corresponding entity of the target event in the knowledge graph is mined, and then the satellite observation task is predicted and generated based on ontology reasoning in the graph neural network.
Further, as shown in fig. 7, the step 31 includes:
step 311, identifying an entity mention set M from an entity set E and a text set D in a knowledge base of the satellite observation field;
step 312, mapping each entity mention mε M to a corresponding entity eE through a fifth model; if a certain m cannot be mapped to e, defining m as unlink; the fifth model includes a candidate entity generation model, a candidate entity ranking model, and an unlink mention prediction model.
Through the steps, link disambiguation can be completed for the entity, and the link between the entity in the target event and the entity in the knowledge base in the satellite observation field is standardized, so that the consistency of data is ensured, and the complete information of the target event is acquired from the knowledge base in the satellite observation field.
Further, the step 312 includes:
step 3121, inputting a certain entity mention M in the entity mention set M into the candidate entity generation model, and searching a plurality of candidate entities most similar to the M from the entity set E based on a dictionary to obtain a candidate entity set Em; the dictionary comprises a synonym dictionary and an ambiguous word dictionary; the synonym dictionary is used for mapping entity references into canonical entities existing in a knowledge base of the satellite observation field; the ambiguous word dictionary is used for obtaining an initial candidate entity set of entity names;
Step 3122, forming scoring entity pairs by each candidate entity in m and Em, inputting the scoring entity pairs into the candidate entity ranking model, and scoring and ranking each entity pair by the model according to the surface similarity and the context feature similarity;
step 3123, based on the unlink mention prediction model, screening out entity pairs with highest scores by an 'unlink mention' prediction algorithm, and defining the entity mention m in the entity pair as the best entity in the entity set E and the rest entity mention as unlink; the 'unlink mention' prediction algorithm belongs to the prior art and is a method for predicting unlink mention in a knowledge graph.
By the method, the best matching entity of the target event can be put into the entity set, so that ambiguity among entity links is eliminated.
Further, as shown in fig. 8, the step 32 includes:
step 321, performing vector organization on the task association relationship, the attribute information, constraint information (such as a region around a fire disaster concerned) and auxiliary information (such as an infrared abnormality of the fire disaster) of a target entity in the observation knowledge graph, and converting the task association relationship and the constraint information into a corresponding multidimensional matrix through feature vector calculation; the feature vector calculation belongs to the prior art and is used for calculating the feature vector;
Step 322, performing graph nerve convolution calculation on the multidimensional matrix to obtain semantic representation, so as to enhance semantic features; the graph nerve convolution calculation belongs to the prior art, and is a matrix operation method;
step 323, calculating the attention of a certain target entity to the auxiliary information through an attention calculation rule, wherein the attention calculation rule is used for enhancing knowledge association; the attention calculation rule belongs to the prior art and is a data processing method in machine learning;
step 324, reserving a plurality of associated information record combinations to generate an associated information candidate set so as to reduce the operation amount;
step 325, mining out the associated information candidate entity of the target event by the multi-layer perceptron based on the associated information candidate set, and predicting to obtain a satellite observation task, such as a nearby residential point target observation task; the satellite observation tasks comprise task types, observation targets, allowed time windows, observation target types, observation target mobility, observation feature elements, observation feature requirements, user preferences, priorities and the like; the multi-layer perceptron belongs to the prior art and is a supervised machine learning algorithm.
Through the steps, satellite task observation elements in the knowledge base of the satellite observation field can be matched to the target event, so that satellite observation tasks aiming at the event can be predicted.
Embodiment two:
as shown in fig. 9, the present embodiment provides a satellite task mining prediction system based on public information, which includes a data extraction module, a data processing module and a result generation module:
the data extraction module comprises an extraction model and a domain word stock, and extracts structural information about a target event based on public information in the Internet; the extraction model comprises a first model and a second model; the first model is used for screening out events consistent with the satellite observation task types; the second model is used for carrying out event extraction and generating structural information;
the data processing module comprises a database, an association relation unit and an observation task unit:
the database is used for storing event knowledge maps, a satellite observation field knowledge base and observation knowledge maps;
the association relation unit stores a relation model; the relation model generates a task association relation based on the structured information and an event knowledge graph; the relationship model includes a third model and a fourth model: the third model is used for carrying out embedded representation on the structured information to obtain a word vector of the target event; the fourth model is used for acquiring the context characteristics of the target event;
The observation task unit is stored with a prediction model; the prediction model predicts satellite observation tasks based on the task association relation; the prediction model includes a fifth model and a sixth model: the fifth model is used for mapping each entity mention in the knowledge base of the satellite observation field to a corresponding entity, expanding the observation knowledge graph, and comprises a candidate entity generation model, a candidate entity ranking model and an unlink mention prediction model; the sixth model is a graph neural network and is used for mining out the associated information candidate entity of the target event from the observation knowledge graph and predicting to obtain a satellite observation task;
and the result generation module is used for issuing the satellite observation task.
Embodiment III:
the embodiment provides a satellite task mining prediction device based on public information, which comprises a processor, a memory and a bus, wherein the memory stores instructions and data which can be read by the processor, the processor is used for calling the instructions and the data in the memory so as to execute any satellite task mining prediction method based on the public information, and the bus is connected with all functional components to transmit information.
In yet another embodiment, the present solution may be implemented by means of an apparatus, which may include corresponding modules performing each or several steps of the above-described embodiments. A module may be one or more hardware modules specifically configured to perform the respective steps, or be implemented by a processor configured to perform the respective steps, or be stored within a computer-readable medium for implementation by a processor, or be implemented by some combination.
The processor performs the various methods and processes described above. For example, method embodiments in the present solution may be implemented as a software program tangibly embodied on a machine-readable medium, such as a memory. In some embodiments, part or all of the software program may be loaded and/or installed via memory and/or a communication interface. One or more of the steps of the methods described above may be performed when a software program is loaded into memory and executed by a processor. Alternatively, in other embodiments, the processor may be configured to perform one of the methods described above in any other suitable manner (e.g., by means of firmware).
The device may be implemented using a bus architecture. The bus architecture may include any number of interconnecting buses and bridges depending on the specific application of the hardware and the overall design constraints. The bus connects together various circuits including one or more processors, memories, and/or hardware modules. The bus may also connect various other circuits such as peripherals, voltage regulators, power management circuits, external antennas, and the like.
The bus may be an industry standard architecture (ISA, industry Standard Architecture) bus, a peripheral component interconnect (PCI, peripheral Component) bus, or an extended industry standard architecture (EISA, extended Industry Standard Component) bus, etc., and may be classified as an address bus, a data bus, a control bus, etc.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (6)
1. The satellite task mining prediction method based on the public information is characterized by comprising the following steps of:
step 1, based on public information and a domain word stock in the Internet, extracting structural information about a target event through an extraction model; the extraction model comprises a first model and a second model; the first model is used for screening out events consistent with the satellite observation task types; the second model is used for carrying out event extraction and generating structural information; the step 1 specifically includes:
step 11, automatically grabbing the public data in a preset range at fixed time through a crawling program based on the public information in the Internet;
step 12, for the public data, removing invalid information through data cleaning;
step 13, performing word segmentation on the public data after data cleaning by a natural language processing method;
step 14, carrying out data filtering on the segmented public data based on the domain word stock, and screening out data meeting the domain requirements;
step 15, screening events consistent with the satellite observation task type from the data-filtered public data through a first model; the first model is a text classification model;
Step 16, extracting the event through a second model for the event to obtain structural information; the second model is a pre-training model;
step 2, generating a task association relation through a relation model based on the structured information and the event knowledge graph; the relation model comprises a third model and a fourth model; the third model is used for carrying out embedded representation on the structured information to obtain a word vector of the target event; the fourth model is used for acquiring the context characteristics of the target event;
the step 2 specifically includes:
step 21, matching the target event with the historical similar event based on the structural information and the event knowledge graph;
step 22, carrying out event association query analysis based on a logic link in the event knowledge graph to generate a task association relation for a target event;
the step 21 specifically includes:
step 211, aiming at the structured information, carrying out embedded representation through a third model to obtain a word vector of a target event; the third model is a pre-training model;
step 212, obtaining the context characteristics of the target event through a fourth model based on the embedded and represented structured information; the fourth model is a neural network model;
Step 213, matching the target event with the historical similar event in the event knowledge graph through similarity calculation based on the word vector, the context feature and the event knowledge graph;
the step 22 specifically includes:
step 221, constructing the participants, types and trigger words of the event as event descriptions, and matching the structured information of the target event to the most similar historical event nodes in the event knowledge graph through similarity calculation to obtain the trigger words and entity lists of the corresponding most similar historical events of the target event;
step 222, based on the trigger word and the entity list, reasoning is carried out through a plurality of causal relations in the causal link and a plurality of compliant relations in the compliant link, and a logical link reasoning result is obtained as a task association relation; the logic link reasoning result comprises a causal link reasoning result and a compliant link reasoning result; the causal link reasoning result is entity information obtained in an entity list after causal reasoning is performed according to the trigger word; the forward link reasoning result refers to entity information obtained in the entity list after forward reasoning is carried out according to the trigger word;
step 3, predicting a satellite observation task through a prediction model based on the task association relation and a satellite observation field knowledge base; the prediction model comprises a fifth model and a sixth model; the fifth model is used for mapping each entity mention in the knowledge base of the satellite observation field to a corresponding entity and expanding an observation knowledge map; and the sixth model is used for mining out the associated information candidate entity of the target event from the observation knowledge graph and predicting to obtain a satellite observation task.
2. The method according to claim 1, wherein the step 3 comprises:
step 31, expanding an observation knowledge graph through a fifth model based on a knowledge base in the satellite observation field and a task association relation;
step 32, predicting satellite observation tasks through a sixth model based on task association relations and the observation knowledge graph; the sixth model is a graph neural network and comprises a knowledge representation embedding layer, a characteristic convolution layer, an attention mechanism layer and an associated information prediction layer.
3. The method according to claim 2, wherein said step 31 comprises:
step 311, identifying an entity mention set from an entity set and a text set of a knowledge base in the satellite observation field;
step 312, mapping each entity mention to a corresponding entity by a fifth model; if an entity mention cannot be mapped to an entity, defining the entity mention as unlinkable; the fifth model includes a candidate entity generation model, a candidate entity ranking model, and an unlink mention prediction model.
4. A method according to claim 3, wherein said step 32 comprises:
step 321, carrying out vector organization on attribute information, constraint information and auxiliary information of a target entity in a task association relation and observation knowledge graph, and converting the attribute information, the constraint information and the auxiliary information into corresponding multidimensional matrixes through feature vector calculation;
Step 322, performing graph nerve convolution calculation on the multidimensional matrix to obtain semantic representation;
step 323, calculating the attention of a certain target entity to the auxiliary information through an attention calculation rule;
step 324, reserving a plurality of association information record combinations to generate an association information candidate set;
and 325, mining out the associated information candidate entity of the target event through the multi-layer perceptron based on the associated information candidate set, and predicting to obtain a satellite observation task.
5. The satellite task mining and predicting system based on the public information is characterized by comprising a data extraction module, a data processing module and a result generation module:
the data extraction module comprises an extraction model and a domain word stock, and extracts structural information about a target event based on public information in the Internet; the extraction model comprises a first model and a second model; the first model is used for screening out events consistent with the satellite observation task types; the second model is used for carrying out event extraction and generating structural information;
the specific process comprises the following steps:
step 11, automatically grabbing the public data in a preset range at fixed time through a crawling program based on the public information in the Internet;
Step 12, for the public data, removing invalid information through data cleaning;
step 13, performing word segmentation on the public data after data cleaning by a natural language processing method;
step 14, carrying out data filtering on the segmented public data based on the domain word stock, and screening out data meeting the domain requirements;
step 15, screening events consistent with the satellite observation task type from the data-filtered public data through a first model; the first model is a text classification model;
step 16, extracting the event through a second model for the event to obtain structural information; the second model is a pre-training model;
the data processing module comprises a database, an association relation unit and an observation task unit:
the database is used for storing event knowledge maps, a satellite observation field knowledge base and observation knowledge maps;
the association relation unit stores a relation model; the relation model generates a task association relation based on the structured information and an event knowledge graph; the relationship model includes a third model and a fourth model: the third model is used for carrying out embedded representation on the structured information to obtain a word vector of the target event; the fourth model is used for acquiring the context characteristics of the target event;
The specific process comprises the following steps:
step 21, matching the target event with the historical similar event based on the structural information and the event knowledge graph;
step 22, carrying out event association query analysis based on a logic link in the event knowledge graph to generate a task association relation for a target event;
the step 21 specifically includes:
step 211, aiming at the structured information, carrying out embedded representation through a third model to obtain a word vector of a target event; the third model is a pre-training model;
step 212, obtaining the context characteristics of the target event through a fourth model based on the embedded and represented structured information; the fourth model is a neural network model;
step 213, matching the target event with the historical similar event in the event knowledge graph through similarity calculation based on the word vector, the context feature and the event knowledge graph;
the step 22 specifically includes:
step 221, constructing the participants, types and trigger words of the event as event descriptions, and matching the structured information of the target event to the most similar historical event nodes in the event knowledge graph through similarity calculation to obtain the trigger words and entity lists of the corresponding most similar historical events of the target event;
Step 222, based on the trigger word and the entity list, reasoning is carried out through a plurality of causal relations in the causal link and a plurality of compliant relations in the compliant link, and a logical link reasoning result is obtained as a task association relation; the logic link reasoning result comprises a causal link reasoning result and a compliant link reasoning result; the causal link reasoning result is entity information obtained in an entity list after causal reasoning is performed according to the trigger word; the forward link reasoning result refers to entity information obtained in the entity list after forward reasoning is carried out according to the trigger word;
the observation task unit is stored with a prediction model; the prediction model predicts satellite observation tasks based on the task association relation; the prediction model includes a fifth model and a sixth model: the fifth model is used for mapping each entity mention in the knowledge base of the satellite observation field to a corresponding entity, expanding the observation knowledge graph, and comprises a candidate entity generation model, a candidate entity ranking model and an unlink mention prediction model; the sixth model is a graph neural network and is used for mining out the associated information candidate entity of the target event from the observation knowledge graph and predicting to obtain a satellite observation task;
And the result generation module is used for issuing the satellite observation task.
6. The satellite task mining prediction device based on public information is characterized by comprising a processor, a memory and a bus, wherein the memory stores instructions and data which can be read by the processor, the processor is used for calling the instructions and the data in the memory to execute the method according to any one of claims 1-4, and the bus is used for connecting all functional components to transmit information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310646855.0A CN116402137B (en) | 2023-06-02 | 2023-06-02 | Satellite task mining prediction method, system and device based on public information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310646855.0A CN116402137B (en) | 2023-06-02 | 2023-06-02 | Satellite task mining prediction method, system and device based on public information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116402137A CN116402137A (en) | 2023-07-07 |
| CN116402137B true CN116402137B (en) | 2023-09-12 |
Family
ID=87012619
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310646855.0A Active CN116402137B (en) | 2023-06-02 | 2023-06-02 | Satellite task mining prediction method, system and device based on public information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116402137B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116992294B (en) * | 2023-09-26 | 2023-12-19 | 成都国恒空间技术工程股份有限公司 | Satellite measurement and control training evaluation method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112235029A (en) * | 2020-08-24 | 2021-01-15 | 成都天奥集团有限公司 | Automatic operation management method for large-scale low-orbit satellite constellation operation and control system |
| CN112668613A (en) * | 2020-12-07 | 2021-04-16 | 中国西安卫星测控中心 | Satellite infrared imaging effect prediction method based on weather forecast and machine learning |
| CN114880484A (en) * | 2022-05-11 | 2022-08-09 | 军事科学院系统工程研究院网络信息研究所 | Satellite communication frequency-orbit resource map construction method based on vector mapping |
| CN115391545A (en) * | 2022-04-26 | 2022-11-25 | 航天宏图信息技术股份有限公司 | Knowledge graph construction method and device for multi-platform collaborative observation task |
| CN115422373A (en) * | 2022-09-05 | 2022-12-02 | 中国人民解放军国防科技大学 | Knowledge graph and user intention task decomposition method for satellite task planning |
| WO2023086613A1 (en) * | 2021-11-15 | 2023-05-19 | Urugus S.A. | Dynamic order and resource management method and system for geospatial information provision |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2946305B1 (en) * | 2013-01-16 | 2021-02-03 | Tata Consultancy Services Limited | A system and method for smart public alerts and notifications |
| US20200019435A1 (en) * | 2018-07-13 | 2020-01-16 | Raytheon Company | Dynamic optimizing task scheduling |
-
2023
- 2023-06-02 CN CN202310646855.0A patent/CN116402137B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112235029A (en) * | 2020-08-24 | 2021-01-15 | 成都天奥集团有限公司 | Automatic operation management method for large-scale low-orbit satellite constellation operation and control system |
| CN112668613A (en) * | 2020-12-07 | 2021-04-16 | 中国西安卫星测控中心 | Satellite infrared imaging effect prediction method based on weather forecast and machine learning |
| WO2023086613A1 (en) * | 2021-11-15 | 2023-05-19 | Urugus S.A. | Dynamic order and resource management method and system for geospatial information provision |
| CN115391545A (en) * | 2022-04-26 | 2022-11-25 | 航天宏图信息技术股份有限公司 | Knowledge graph construction method and device for multi-platform collaborative observation task |
| CN114880484A (en) * | 2022-05-11 | 2022-08-09 | 军事科学院系统工程研究院网络信息研究所 | Satellite communication frequency-orbit resource map construction method based on vector mapping |
| CN115422373A (en) * | 2022-09-05 | 2022-12-02 | 中国人民解放军国防科技大学 | Knowledge graph and user intention task decomposition method for satellite task planning |
Non-Patent Citations (1)
| Title |
|---|
| 新一代指挥信息系统的知识中心原型构想;易侃;王菁;崔隽;葛唯益;汪霜玲;阮国庆;;指挥信息系统与技术(03);14-20 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116402137A (en) | 2023-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Knowledge engineering with big data | |
| Zhang et al. | Integrated applications of building information modeling and artificial intelligence techniques in the AEC/FM industry | |
| Li et al. | Multi-attribute decision making with generalized fuzzy numbers | |
| Ma et al. | Big graph search: challenges and techniques | |
| CN112015868B (en) | Question-answering method based on knowledge graph completion | |
| CN116402137B (en) | Satellite task mining prediction method, system and device based on public information | |
| Wang et al. | Enhancing spatial and textual analysis with EUPEG: An extensible and unified platform for evaluating geoparsers | |
| CN112231418A (en) | Power standard knowledge graph construction method and device, computer equipment and medium | |
| CN113435582B (en) | Text processing method and related equipment based on sentence vector pre-training model | |
| Mills et al. | Tracing with less data: active learning for classification-based traceability link recovery | |
| Bedi et al. | CitEnergy: A BERT based model to analyse Citizens’ Energy-Tweets | |
| CN112583640A (en) | Service fault detection method and device based on knowledge graph | |
| Feng et al. | A pretraining numerical reasoning model for ordinal constrained question answering on knowledge base | |
| Zhang et al. | Transformer-based approach for automated context-aware IFC-regulation semantic information alignment | |
| Li et al. | Emotion analysis model of microblog comment text based on CNN-BiLSTM | |
| Embarak | Explainable artificial intelligence for services exchange in smart cities | |
| Sahoh et al. | Automatic semantic description extraction from social big data for emergency management | |
| CN110232108B (en) | Man-machine conversation method and conversation system | |
| US11314793B2 (en) | Query processing | |
| Yin et al. | A deep natural language processing‐based method for ontology learning of project‐specific properties from building information models | |
| CN115062619A (en) | Chinese entity linking method, device, equipment and storage medium | |
| Zhu et al. | An intelligent collaboration framework of IoT applications based on event logic graph | |
| KR102096328B1 (en) | Platform for providing high value-added intelligent research information based on prescriptive analysis and a method thereof | |
| Karimi et al. | Multi-task recommendation system for scientific papers with high-way networks | |
| Guo et al. | A Review of Knowledge Graph Research in Military Domain |
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 |