Energy science and technology intelligence system

Abstract

The invention provides an energy science and technology intelligence system, which comprises a basic data layer, a data platform layer and a business application layer. Wherein the basic data layer is used for determining an energy science and technology information data source and acquiring energy data from the energy science and technology information data source; the data center layer is used for cleaning, extracting and fusing energy data, constructing an energy knowledge graph and providing energy data services, and extraction comprises extraction of energy technology terms, extraction of technology affiliation relationships, extraction of technology dependency relationships, extraction of energy industries or products and extraction of industrial upstream and downstream relationships; the energy data service comprises an energy data retrieval service and an energy data analysis service; and the application layer is used for providing an energy science and technology information analysis result for the target client based on the energy data retrieval service and the energy data analysis service. Therefore, the energy industry can be served more accurately by constructing an energy science and technology intelligence information industry chain.

Description

Technical Field

[0001] The present invention relates to the field of energy technology, and in particular to an energy technology intelligence system. Background Art

[0002] With the development of the times, scientific and technological information resources have become a key force in promoting scientific and technological innovation. Nowadays, various emerging technologies are emerging in an endless stream, which has shortened the life cycle of science and technology and greatly increased the speed of transformation of scientific and technological achievements. At the same time, with the rapid development of technologies such as big data and artificial intelligence, human society is also rapidly developing from informatization to dataization and intelligence. In the face of the above development background, how to ensure the concentration of energy innovation knowledge is a difficult problem that needs to be solved in the field of energy technology. Summary of the invention

[0003] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, the present invention proposes an energy science and technology intelligence system, including: a basic data layer, a data middle platform layer and a business application layer.

[0005] The basic data layer is used to determine the energy science and technology intelligence data source and obtain energy data from the energy science and technology intelligence data source;

[0006] The data middle platform layer is used to clean, extract and integrate the energy data, build an energy knowledge graph, and provide energy data services; wherein the extraction includes the extraction of energy technical terms, the extraction of technical affiliations, the extraction of technical dependencies, the extraction of energy industries or products, and the extraction of upstream and downstream relationships of industries; the energy data services include energy data retrieval services and energy data analysis services;

[0007] The application layer is used to provide energy technology intelligence analysis results to the target client based on the energy data retrieval service and the energy data analysis service.

[0008] According to the energy science and technology intelligence system provided by the present invention, the energy data is cleaned, extracted, and integrated, an energy knowledge graph is constructed, and energy data services are provided, including:

[0009] Standardizing the energy data to obtain cleaned energy data;

[0010] The cleaned energy data is subjected to knowledge extraction and data fusion to construct an energy knowledge graph, and data modeling is performed through graph node discovery, anomaly point monitoring, information transmission, center point discovery and other technologies to provide energy data services; wherein, the knowledge extraction includes at least one of keyword extraction, concept extraction, entity extraction, relationship extraction, and attribute extraction, the relationship extraction includes classification relationship extraction and correlation relationship extraction, and the classification relationship extraction includes concept hyponym and hyponym relationship extraction and keyword hyponym and hyponym relationship extraction.

[0011] According to the energy science and technology intelligence system provided by the present invention, the energy science and technology intelligence analysis results include at least one of energy strategy analysis results, energy industry chain analysis results, energy technology analysis results, energy field talent analysis results, energy enterprise portrait analysis results and energy industry dynamics analysis results.

[0012] According to the energy science and technology intelligence system provided by the present invention, the data middle platform layer is also used to perform one or more of file processing, content translation, offline calculation, data transmission, data retrieval, word segmentation and part-of-speech tagging, subject identification, sentiment analysis and text keyword processing on the energy data.

[0013] According to the energy science and technology intelligence system provided by the present invention, the file processing of the energy data includes:

[0014] When the energy data is in the form of multiple documents, the multiple documents are archived and stored to support front-end query, batch export and other operations;

[0015] The content translation of the energy data includes:

[0016] Identifying the language of the energy data, and translating the energy data into the target language if the language of the energy data is not the target language; wherein the target language is a language selected by a user or a set language;

[0017] The offline calculation of the energy data includes:

[0018] In the case of batch processing of the energy data, a computing framework based on hive and Spark is used to perform offline computing on the energy data;

[0019] The data transmission of the energy data comprises:

[0020] Using Kafka as a message middleware to transmit the energy data;

[0021] The retrieving the energy data includes:

[0022] Create indexes, store data and retrieve data for the energy data using ElasticSearch ES technology;

[0023] The word segmentation and part-of-speech tagging processing of the energy data includes:

[0024] Using a conditional random field model, the energy data is segmented and POS tagged in at least one language; wherein the conditional random field model has been previously learned on a large-scale tagging training forecast;

[0025] The subject identification of the energy data includes:

[0026] Using an end-to-end deep learning model to identify entities in the energy data;

[0027] The performing sentiment analysis on the energy data includes:

[0028] Using a preset polarity classification system, sentiment analysis is performed on the energy data in at least one language to obtain sentiment information corresponding to the energy data; wherein the preset polarity classification system includes positive, negative and neutral, or includes strongly positive, generally positive, neutral, generally negative, and strongly negative;

[0029] The performing text keyword processing on the energy data includes:

[0030] Using at least one syntactic analysis method to determine at least one first candidate keyword corresponding to the energy data; wherein the at least one syntactic analysis method includes one or more of word segmentation, stop word filtering, clause extraction, and phrase recognition;

[0031] Determine at least one second candidate keyword corresponding to a target word frequency according to the word frequency of each of the first candidate keywords, and / or determine at least one second candidate keyword corresponding to a target document frequency according to the document frequency of each of the first candidate keywords; wherein the target word frequency is a word frequency within a first set range, and the target document frequency is a document frequency within a second set range;

[0032] Based on each of the second candidate keywords, a graph-based TextRank algorithm is used to determine a target keyword.

[0033] According to the energy science and technology intelligence system provided by the present invention, the basic data layer is also used to store the energy data in a variety of data storage methods according to the data volume, type, service object and application scenario of the energy data.

[0034] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0036] Figure 1 A schematic diagram of the structure of an energy technology information system provided by an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of the architecture of energy knowledge graph construction and application provided by an embodiment of the present invention;

[0038] Figure 3 A schematic diagram of an architecture for energy data collection and fusion provided by an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of the architecture of an energy technology intelligence system in a scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0041] In the relevant technologies, there are the following difficulties in centralizing energy innovation knowledge:

[0042] Difficulty 1: Data sources are heterogeneous, multi-source, multi-language, and multi-modal, and the knowledge network updates quickly. A single model for extracting structured knowledge is not sufficient, and it is necessary to combine and optimize multiple combined models to extract domain knowledge.

[0043] Difficulty 2: There is relatively little publicly labeled data in the energy field. The relevant extraction algorithms based on deep learning require a lot of labeled data. Zero-sample or small-sample learning poses a great obstacle to knowledge extraction.

[0044] Difficulty 3: The analysis of energy strategies in specific regions lacks quantitative indicators for judgment, and it is difficult to build an energy technology analysis model. It is necessary to combine the wisdom of field personnel and machine learning to explore and discover energy field knowledge and knowledge relationships, and infer and discover indicator dimensions that can provide quantitative calculations.

[0045] In response to the above problems, an embodiment of the present invention provides an energy science and technology intelligence system to achieve knowledge association through high-quality multilingual translation, intelligent extraction and fusion of energy field knowledge, and automatic construction of large-scale knowledge graphs, thereby building a scientific and technological cognitive brain engine to serve science and technology intelligence analysis.

[0046] The energy technology information system according to an embodiment of the present invention will be described below with reference to the accompanying drawings.

[0047] Figure 1 A schematic diagram of the structure of an energy technology information system provided by an embodiment of the present invention.

[0048] like Figure 1 As shown, the energy technology intelligence system includes: basic data layer, data middle platform layer and business application layer.

[0049] Among them, the basic data layer is used to determine the energy science and technology intelligence data source and obtain energy data from the energy science and technology intelligence data source;

[0050] The data middle platform layer is used to clean, extract, and integrate energy data, build energy knowledge graphs, and provide energy data services. The extraction includes the extraction of energy technical terms, the extraction of technical affiliations, the extraction of technical dependencies, the extraction of energy industries or products, and the extraction of upstream and downstream relationships of industries. Energy data services include energy data retrieval services and energy data analysis services.

[0051] The application layer is used to provide energy technology intelligence analysis results to the target clients based on energy data retrieval services and energy data analysis services.

[0052] As a possible implementation method, energy data can be standardized first to obtain cleaned energy data, and then knowledge extraction and data fusion are performed on the cleaned energy data to construct an energy knowledge graph, and data modeling is performed through graph node discovery, anomaly point monitoring, information transmission, center point discovery and other technologies to provide energy data services; wherein, knowledge extraction includes at least one of keyword extraction, concept extraction, entity extraction, relationship extraction, and attribute extraction, relationship extraction includes classification relationship extraction and correlation relationship extraction, and classification relationship extraction includes concept hyponym and hyponym relationship extraction and keyword hyponym and hyponym relationship extraction.

[0053] Optionally, the architecture of energy knowledge graph construction and application is as follows Figure 2As shown in the figure, after extracting energy data, semantic representation learning and knowledge reasoning can be performed. Semantic representation learning includes shallow representation learning, graph representation learning and graph neural network. Shallow representation learning mainly predicts relationships through TransE model, DistMult model, ConvE model, ComplEx model, etc. Knowledge reasoning includes cognitive reasoning and question answering, and cognitive reasoning includes cognitive graph and concept expansion. After that, application and / or generation can be performed. Application includes algorithms (entity linking and knowledge graph verification), alignment (entity alignment and cross-language alignment), completion (knowledge base completion) and visualization (knowledge visualization), and generation includes summary generation.

[0054] As an example, Neo4j can be used as a database to store and manage energy data. The entities, attribute information and semantic relationships between entities contained in the energy data are stored through the graph data storage framework provided by Neo4j, and a data operation engine is developed based on it to provide functions such as adding, deleting, modifying, querying and analyzing energy data.

[0055] As a possible implementation method, the energy science and technology intelligence analysis results may include at least one of energy strategy analysis results, energy industry chain analysis results, energy technology analysis results, energy field talent analysis results, energy enterprise portrait analysis results and energy industry dynamics analysis results.

[0056] As an example, the basic data layer can mainly study energy science and technology intelligence data source discovery technology and energy data acquisition technology, wherein the energy science and technology intelligence data source discovery technology is used to realize the discovery of energy-related data sources from a large number of websites or databases. Energy data acquisition technology is used to obtain energy data from corresponding data sources through corresponding network protocol parsing programs and / or data acquisition tools.

[0057] The data middle platform layer can mainly study the technologies of energy data cleaning, extraction, fusion and data services, including the extraction of energy technology terminology, the extraction of technology affiliations, the extraction of technology dependencies, the extraction of energy industries or products, the extraction of upstream and downstream industry relationships, as well as energy knowledge graph construction technology, energy data retrieval technology for external services and energy data analysis technology.

[0058] The application layer can mainly study energy strategy analysis, energy industry chain analysis, energy technology analysis, energy field talent analysis, energy enterprise portrait analysis and energy industry dynamics analysis. Among them, energy strategy analysis mainly analyzes the construction of energy strategy topic system, energy strategy classification and energy strategy sentiment judgment; energy industry chain analysis mainly analyzes the construction of energy industry chain system, the construction and extraction of energy market indicators and the identification of energy industry projects; energy technology analysis mainly studies the construction of energy technology system, the identification of information technology in the energy field, energy technology maturity analysis, the evaluation of energy technology industrialization application, the construction of energy technology research institution system, the correspondence between energy technology system and legal patent language; energy field talent analysis mainly analyzes the identification of energy field talent, the evaluation of energy field talent technology; energy enterprise portrait analysis mainly analyzes the construction of energy enterprise technology evaluation system, the construction of energy enterprise business evaluation system, and the value evaluation of energy enterprises; energy industry dynamics analysis mainly collects and analyzes real-time dynamic information in the industry to timely discover market trends and track major events.

[0059] As a possible implementation method, the data middle platform layer can also be used for one or more of file processing, content translation, offline calculation, data transmission, data retrieval, word segmentation and part-of-speech tagging, subject identification, sentiment analysis and text keyword processing of energy data.

[0060] The file processing of energy data may be to archive and store multiple documents when the energy data is in multiple documents, so as to support front-end query, batch export and other operations.

[0061] The content translation of the energy data may be to identify the language of the energy data, so as to translate the energy data into the target language when the language of the energy data is not the target language, wherein the target language is a language selected by the user or a set language.

[0062] Offline calculation of energy data can be performed by using a computing framework based on hive and Spark when batch processing of energy data is performed.

[0063] For data transmission of energy data, Kafka may be used as a message middleware to transmit energy data.

[0064] Data retrieval for energy data may be performed by using ES (Elasticsearch) technology to create indexes, store data, and retrieve data for energy data.

[0065] The energy data may be segmented and POS tagged in at least one language using a conditional random field model, wherein the conditional random field model has been previously learned on a large-scale annotation training set.

[0066] To identify the subject of energy data, an end-to-end deep learning model can be used to identify the entities in the energy data.

[0067] Sentiment analysis of energy data may be performed by using a preset polarity classification system to perform sentiment analysis on the energy data in at least one language to obtain sentiment information corresponding to the energy data, wherein the preset polarity classification system includes positive, negative and neutral, or includes strongly positive, generally positive, neutral, generally negative, and strongly negative.

[0068] Text keyword processing of energy data may be performed by first using at least one syntactic analysis method to determine at least one first candidate keyword corresponding to the energy data, wherein the at least one syntactic analysis method includes one or more of word segmentation, stop word filtering, clause extraction, and phrase recognition, and then determining at least one second candidate keyword corresponding to a target word frequency based on the word frequency of each first candidate keyword, and / or determining at least one second candidate keyword corresponding to a target document frequency based on the document frequency of each first candidate keyword, wherein the target word frequency is a word frequency within a first set range, and the target document frequency is a document frequency within a second set range, thereby determining the target keyword based on each second candidate keyword using a graph-based TextRank algorithm.

[0069] The embodiment of the present invention does not limit the setting of the first setting range and the second setting range. Optionally, the setting time period can be set according to manual experience, for example, the first setting range can be set to a word frequency greater than a certain value (high word frequency), and the second setting range can be set to a document frequency less than a certain value (low document frequency), or the setting time period can be dynamically adjusted according to actual application requirements, which is not limited in this embodiment.

[0070] As a possible implementation method, the basic data layer can also be used to store energy data in a variety of data storage methods according to the amount, type, service object and application scenario of energy data. Figure 3As shown in the figure, the energy source data can be obtained from the internal business system data and / or external data through microservices, distributed, message middleware, virtualization and other infrastructures. The internal business system data includes application and review information, and the external data includes global Chinese and English paper databases, scientific and technological dynamic information, scientific and technological project information, search engines, and others. After that, task monitoring and task scheduling are carried out, and then the data collection and fusion pipeline is realized through user configuration, distributed crawlers, information extraction, data deduplication, data fusion, and entity disambiguation.

[0071] The energy science and technology intelligence system proposed in the present invention includes: a basic data layer, a data middle platform layer and a business application layer. Among them, the basic data layer is used to determine the energy science and technology intelligence data source and obtain energy data from the energy science and technology intelligence data source; the data middle platform layer is used to clean, extract and fuse energy data, build an energy knowledge graph, and provide energy data services, wherein the extraction includes the extraction of energy technology terms, the extraction of technical affiliations, the extraction of technical dependencies, the extraction of energy industries or products, and the extraction of upstream and downstream relationships of the industry, and the energy data services include energy data retrieval services and energy data analysis services; the application layer is used to provide energy science and technology intelligence analysis results to the target client based on energy data retrieval services and energy data analysis services. Therefore, by building an energy science and technology intelligence information industry chain, it is possible to serve the energy industry more accurately.

[0072] In order to more clearly illustrate the energy technology information system of the present invention, an example is now given to illustrate.

[0073] Figure 4 A schematic diagram of the architecture of an energy technology intelligence system in a scenario provided by an embodiment of the present invention.

[0074] like Figure 4As shown, the basic data layer can obtain energy patent data, energy expert data, energy enterprise data, etc., and then the data middle platform layer performs data structure standardization, data cleaning, data field processing, data fusion, data translation, etc. on the acquired energy data, so as to construct a knowledge graph, perform natural language processing and algorithm modeling. Among them, the process of constructing a knowledge graph includes knowledge representation, knowledge modeling and knowledge storage. Knowledge representation is to extract concepts, entities, attributes, and relationships. Knowledge modeling includes concept definition, concept specification, attribute definition and attribute specification. Natural language processing includes word segmentation and part-of-speech tagging, text Keyword extraction, subject identification, semantic understanding and sentiment analysis, etc.; algorithm modeling includes establishing enterprise science and technology innovation value assessment model, energy field technology analysis model, technology chain map model, industrial chain map model and energy strategy analysis model, etc., and then generating multiple subsystems, including data processing subsystem, machine translation subsystem, data semantic retrieval subsystem and energy industry chain technology chain data governance subsystem, etc. Finally, the application layer displays the results of energy strategy analysis, energy industry chain analysis, energy technology analysis, energy field talent analysis, energy enterprise portrait analysis and energy industry dynamics analysis.

[0075] In this system, energy data can include energy patent data, energy expert data, energy enterprise data, etc. The data sources are paper database, patent database, scientific research project information disclosed on the Internet, science and technology dynamic information, industry news information, etc.

[0076] In this system, neural network machine translation, intelligent semantic technology, knowledge graph construction and other technologies can be comprehensively applied to clean, extract, and fuse energy data, integrate them into knowledge graphs, and provide services to the outside world through search, visual analysis and other forms. Specifically, the following technologies are included:

[0077] Data cleaning: mainly standardizes the collected data, and is used to extract, clean and reorganize data to improve data quality. The cleaned data will be stored in the original database of the data center.

[0078] File processing: archive and store the collected PDF files, and support front-end query, batch export and other operations.

[0079] Content translation: Machine translation of each piece of data, translating the article into a language that users can recognize. The integrated machine translation subsystem can provide instant language recognition, text translation, document translation and other functions.

[0080] Offline computing: It is generally a process of batch processing of data. For offline computing, two computing frameworks based on Hive and Spark are mainly used.

[0081] Data transmission: Kafka is mainly used as the message middleware. Kafka is a message publishing and subscription system that is suitable for heterogeneous clusters, easy to expand outward, decoupled, and has flexibility and peak processing capabilities. It also supports offline data processing and real-time data processing, providing real-time query capabilities for high-concurrency businesses.

[0082] Data retrieval: ES technology is used to support data index creation, data storage, and data retrieval.

[0083] Knowledge graph technology: By accumulating various entities and relationships such as people, institutions, and enterprises from data information, knowledge-level association analysis can be easily carried out. Since relational databases are not easy to expand and are not conducive to analyzing various associations and semantic relationships between related entities, this system uses Neo4j, a non-relational database based on the Java virtual machine, as a database for storing and managing data information. The entities, attribute information, and semantic relationships between entities contained in the data information are stored through the graph data storage framework provided by Neo4j, and functions such as adding, deleting, modifying, checking, and analyzing data information are provided, enabling faster queries than relational databases.

[0084] Word segmentation and part-of-speech tagging: The most basic underlying technology in natural language processing, the goal is to convert natural language from a character sequence into a basic unit word sequence that carries semantic information. This technology is widely used in information retrieval, information extraction, machine translation, text classification, intelligent question answering, automatic summarization and other fields, and its performance will directly affect the effect of upper-level applications. This system can support word segmentation and part-of-speech tagging in multiple languages.

[0085] For word segmentation and part-of-speech tagging tasks, this system can use the conditional random field model to learn on a large-scale annotated training corpus. Then, based on the annotation sequence output by the model in actual applications, the results of word segmentation and part-of-speech tagging are reversely mapped.

[0086] Subject recognition: A basic task required for natural language upper-layer applications, it refers to identifying entities with specific meanings in text, including names of people, places, institutions, proper nouns, etc. It is an important tool for applications such as information extraction, question-answering systems, semantic understanding, and machine translation, and plays a fundamental role in natural language processing. This system can support subject recognition in Chinese and English.

[0087] Entity extraction methods include artificial organization rule methods, machine learning-based methods, and the relatively popular deep learning-based methods. For news data with a wide range of content, this system uses an end-to-end deep learning model to improve the algorithm's adaptive capabilities, and can run on a certain range of corpora and achieve good recognition results.

[0088] In order to provide efficient performance support for GPU (Graphics Processing Unit) like CNN (Convolutional Neural Networks), and remember as much input information as possible with a simple structure like LSTM (Long Short-Term Memory), this system can use "expanded" CNN technology to capture more input information while keeping the algorithm complexity low, thereby improving the performance of sequence labeling tasks.

[0089] Sentiment analysis: Sentiment analysis, or sentiment tendency analysis, is the analysis of subjective attitudes or opinions expressed in natural language texts, identifying whether the opinions expressed in the article belong to the three categories of positive, negative, and neutral. Alternatively, a more fine-grained polarity classification system, such as strongly positive, generally positive, neutral, generally negative, and strongly negative, can be used to describe the intensity of emotional expression. This system can support sentiment analysis in multiple languages.

[0090] Text keyword extraction: The keyword extraction method adopted by this system is a general method independent of different languages, industries and application fields. For any input text, after removing the universal text commonalities of the scene in which it is located, the keyword features that can prominently represent the theme of the text are extracted and output.

[0091] This technology combines syntactic analysis technology with concept identification and extraction technology to ensure the practicality of keyword extraction from three aspects: using various syntactic analysis methods including word segmentation, stop word filtering, clause extraction, phrase recognition and other technologies to expand the range of candidate keywords; comparing and analyzing the word frequency and document frequency of candidate keywords, and outputting the parts with high word frequency and low document frequency; using the graph-based TextRank algorithm to establish a statistical learning model to output the best keywords.

[0092] Among them, the main principle of the TextRank algorithm is: divide each document into several words, use these words as the vertices of the graph, use the connection between words as the edge, and construct a graph structure model through the connection between points and edges. After that, set the initial weight of each vertex to 1, and calculate the weight of each point through the link relationship between points. The basic idea is that the more vertices connected to the current vertex, and the more important the connected vertices are, the more important the vertex is. The TextRank algorithm is used to calculate the score of each vertex. This value represents the importance of the word in the text, and the first K words with relatively large weights in the document are selected as the keywords of the document.

[0093] In summary, knowledge association can be achieved through high-quality multilingual translation, and large-scale knowledge graphs can be automatically constructed through intelligent extraction and fusion of energy field knowledge, thereby building a scientific and technological innovation cognitive brain engine to serve scientific and technological intelligence analysis. Specifically, based on the technical capabilities of neural network machine translation, high-quality translation of cross-language scientific and technological information can be achieved, and enterprise-level professional document translation capabilities in the field of science and technology can be built to achieve cross-language application of scientific and technological information data in the energy industry, and online acquisition and dynamic fusion of multi-source, dynamic, heterogeneous fragmented knowledge can be achieved, so as to improve the accuracy, comprehensiveness and timeliness of acquiring and utilizing scientific and technological information in the energy industry. At the same time, the knowledge representation of representation learning and symbolic logic is integrated to extract knowledge and fuse data for patent data, enterprise data, scientific and technological information data, industrial data, investment and financing information in the energy field, and construct a knowledge graph in the field of science and technology. Through graph-related node discovery, anomaly detection, information transmission, center point discovery and other technologies, data modeling is carried out to achieve a comprehensive display of multi-dimensional energy intelligence. And combined with knowledge-driven and data-driven, a large-scale knowledge graph is built, and a scientific and technological innovation cognitive brain engine is built to achieve knowledge-rich user understanding and interactive knowledge adaptation.

[0094] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0095] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0096] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute the instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0098] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0099] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0100] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0101] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. An energy technology intelligence system, characterized in that: include: Basic data layer, data middle platform layer and business application layer; The basic data layer is used to determine the energy science and technology intelligence data source and obtain energy data from the energy science and technology intelligence data source; The data middle platform layer is used to clean, extract and integrate the energy data, build an energy knowledge graph, and provide energy data services; wherein the extraction includes the extraction of energy technical terms, the extraction of technical affiliations, the extraction of technical dependencies, the extraction of energy industries or products, and the extraction of upstream and downstream relationships of industries; the energy data services include energy data retrieval services and energy data analysis services; The application layer is used to provide energy technology intelligence analysis results to the target client based on the energy data retrieval service and the energy data analysis service.

2. The system according to claim 1, characterized in that The energy data is cleaned, extracted, and integrated, an energy knowledge graph is constructed, and energy data services are provided, including: Standardizing the energy data to obtain cleaned energy data; The cleaned energy data is subjected to knowledge extraction and data fusion to construct an energy knowledge graph, and data modeling is performed through graph node discovery, anomaly point monitoring, information transmission, center point discovery and other technologies to provide energy data services; wherein, the knowledge extraction includes at least one of keyword extraction, concept extraction, entity extraction, relationship extraction, and attribute extraction, the relationship extraction includes classification relationship extraction and correlation relationship extraction, and the classification relationship extraction includes concept hyponym and hyponym relationship extraction and keyword hyponym and hyponym relationship extraction.

3. The system according to claim 1, characterized in that The energy science and technology intelligence analysis results include at least one of energy strategy analysis results, energy industry chain analysis results, energy technology analysis results, energy field talent analysis results, energy enterprise portrait analysis results and energy industry dynamics analysis results.

4. The system according to claim 1, characterized in that The data middle platform layer is also used to perform one or more of file processing, content translation, offline calculation, data transmission, data retrieval, word segmentation and part-of-speech tagging, subject identification, sentiment analysis and text keyword processing on the energy data.

5. The system according to claim 4, characterized in that The file processing of the energy data includes: When the energy data is in the form of multiple documents, the multiple documents are archived and stored to support front-end query, batch export and other operations; The content translation of the energy data includes: The language of the energy data is identified, and when the language of the energy data is not a target language, the energy data is translated into a target language; wherein the target language is a language selected by a user or a set language.

6. The system according to claim 4, characterized in that The offline calculation of the energy data includes: In the case of batch processing of the energy data, a computing framework based on hive and Spark is used to perform offline computing on the energy data; The data transmission of the energy data comprises: Using Kafka as a message middleware to transmit the energy data; The retrieving the energy data includes: The energy data is indexed, stored and retrieved using the Elastic Search ES technology.

7. The system according to claim 4, characterized in that The word segmentation and part-of-speech tagging processing of the energy data includes: Using a conditional random field model, the energy data is segmented and POS tagged in at least one language; wherein the conditional random field model has been previously learned on a large-scale tagging training forecast; The subject identification of the energy data includes: An end-to-end deep learning model is used to identify entities in the energy data.

8. The system according to claim 4, characterized in that The performing sentiment analysis on the energy data includes: A preset polarity classification system is used to perform sentiment analysis on the energy data in at least one language to obtain sentiment information corresponding to the energy data; wherein the preset polarity classification system includes positive, negative and neutral, or includes strongly positive, generally positive, neutral, generally negative, and strongly negative.

9. The system according to claim 4, characterized in that The performing text keyword processing on the energy data includes: Using at least one syntactic analysis method to determine at least one first candidate keyword corresponding to the energy data; wherein the at least one syntactic analysis method includes one or more of word segmentation, stop word filtering, clause extraction, and phrase recognition; Determine at least one second candidate keyword corresponding to a target word frequency according to the word frequency of each of the first candidate keywords, and / or determine at least one second candidate keyword corresponding to a target document frequency according to the document frequency of each of the first candidate keywords; wherein the target word frequency is a word frequency within a first set range, and the target document frequency is a document frequency within a second set range; Based on each of the second candidate keywords, a graph-based TextRank algorithm is used to determine a target keyword.

10. The system according to claim 1, characterized in that The basic data layer is also used to store the energy data in a variety of data storage methods according to the data volume, type, service object and application scenario of the energy data.

Images