56results about How to "Extraction implementation" patented technology

The invention relates to an entity relationship extracting method of the Zang language. The method comprises the following steps: extracting training linguistic data from the Zang-Chinese text linguistic data information; constructing a Zang word vector model; acquiring an entity relationship characteristic vector from the Zang word vector model; using the entity relationship characteristic vector as an input to construct an entity relationship classification model based on a neural network; and applying multiple layers of characteristic extractions to the entity relationship characteristic vector, thereby finally acquiring a Zang language entity relationship classification. The extraction of the Zang language entity relationship is achieved by constructing the Zang word vector model, researching and solving lexical semantic characteristics and sentence characteristic vector expression methods of the Zang language entity relationship, and further constructing the Zang language entity relationship classification model, accordingly increasing the accuracy in the Zang language entity relationship classification, and providing technical supports and services to the researches in the fields of the Zang language knowledge mapping, question-answering system, information extraction, information search, and the like.

A method for intelligently checking financial documents, a deive and a storage medium are provided in the invention which relates information extraction tecnology. The method comprises: combining natural language processing with information extraction, in the aspect of paragraph text, through effective word segmentation, part-of-speech analysis, named entity recognition, dependency syntax analysisand semantic role analysis, the effect of text information extraction and paragraph text semantic understanding being greatly improved, and the unstructured text data being transformed into structured financial tuples. For tables, use python, the best parser for text tables Docx transforms the table into a data structure that can be processed by computer, then divides the table into regions according to the header information, understands the table semantics, accurately realizes the table element location and element extraction, and transforms the semi-structured table data into structured financial tuples; Finally, the parsed financial structured tuples can be checked with the data of financial statements, and also can be checked for context consistency.

The invention discloses a fault detection and identification method for civil aircraft system based on LSTM-AE depth learning framework and relates to the technical field of condition monitoring and fault diagnosis of complex civil aircraft system, and can be used to realize the monitoring and identification of flight faults. The invention comprises the following steps: selecting time series dataof multi-state parameters of an aircraft in flight under a certain stable condition, and according to the characteristics of the monitored object, the time series data of state parameters under suitable conditions are selected for the training of the system reconstruction model, then the fault-free state of civil aircraft system is modeled and reconstructed by making full use of the long-time series-dependent memory ability of LSTM model. The fault monitoring and identification are realized by further analyzing the reconstruction error of its state parameters. The invention solves the problemof insufficient fault monitoring means of civil aircraft system, utilizes the advantage of deep learning in big data analysis to mine massive operation and maintenance data of civil aircraft, and provides important support for fault monitoring of civil aircraft system and route fault isolation.

The invention provides an entity attribute extraction method and system oriented to an open web page. The method includes the steps: extracting a text of the open web page to obtain a candidate text set of a target entity; extracting the value of a target entity attribute from the candidate text set in a rule-based or statistical-based mode according to the frequency of the target entity attribute in a training text set. Extraction accuracy and recall rate of the entity attribute of the open web page can be improved, and the method is independent of a web page structure and can adapt to change of open web page types.

The invention discloses a text classification and extraction method for Chinese news emergencies, belonging to the natural language processing field. The invention adopts the joint characterization based on the headline and the content information for the news text classification, and integrates the contribution degree of the part of speech to the text classification into the traditional TF-IDF algorithm, the weight of Word2Vec word vector is used to generate short text vector, which avoids the loss of information caused by the single use of title or content and the reduction of classificationaccuracy caused by the different lexical importance of the text. Finally, event instances are extracted by using event-driven news emergency extraction method, It not only overcomes the imbalance between positive and negative examples and data sparseness, but also solves the limitation of pre-defined event categories, and realizes event extraction, which is convenient for journalists and public opinion analysts to use the event extraction results to quickly analyze news.

The invention provides a named entity recognition model training method and device and an information extraction method and device. According to the training texts in the set, entities in the semi-annotation information of the training texts are matched with entities in the corresponding training texts; based on a matching result, an effective named entity label of the corresponding training textis obtained; vector representation of each training text in the training text set is obtained; and based on the vector representation of each training text in the training text set and the effective named entity labeling, a named entity recognition model is trained based on deep learning to obtain a target named entity recognition model.

The embodiment of the invention provides a document-level remote supervision relationship extraction method and system. The method comprises the following steps: acquiring remote supervision data, based on a trained pre-noise-reduction model, carrying out noise reduction processing on the remote supervision data to obtain target remote supervision data, acquiring the trained pre-noise-reduction model by training sample remote supervision data marked as a positive sample and sample remote supervision data marked as a negative sample, and inputting the target remote supervision data into a trained text encoder model to obtain a document level relationship extraction result, wherein the trained text encoder model is obtained by training noise-reduced sample document level remote supervision data. According to the embodiment of the invention, noise reduction is carried out on the remote supervision data in a pre-training mode, noise in the remote supervision data can be effectively filtered out, and the model is pre-trained by utilizing large-scale noise-reduced data, so that document-level remote supervision relationship extraction is realized, and the document-level relationship extraction effect is improved.

The invention provides a remote sensing image change detection method based on dense connection and geometric structure constraint. The method comprises the steps: carrying out the preprocessing of afront-and-back-stage change detection input image, and comprises the calculation of an image normalization parameter, the normalization of a front-and-back-stage image, and the color correction of a back-stage image; dCGC-CD model training and testing are carried out, and change area information extraction is carried out on the remote sensing images in the early and later periods; wherein the DCGC-CD model comprises a front-and-back-stage change feature differential encoding module and a multi-branch geometric structure constraint decoding module, and the front-and-back-stage image change feature differential encoding module comprises a branch for an input front-stage image and a branch for an input back-stage image; the multi-branch geometric structure constraint decoding module comprisesthree branches; wherein the first branch is used for early-stage edge prediction, the second branch is used for change area prediction, the third branch is used for later-stage edge prediction, and the structures of the first branch and the third branch and the structure of the second branch interact through a loss function to effectively constrain a final change detection result through geometric structure information.

The invention discloses a method for identifying and extracting water conservancy spatial relationship words. The method comprises the following steps: acquiring a spatial relationship seed set basedon quantitative statistical characteristics; constructing an original syntax mode; generalizing a syntax mode, and generalizing a plurality of original syntax modes for expressing the same kind of spatial relationship into one mode, so that the number of modes is reduced, and the abstraction degree is improved; and based on the generalized syntax mode, realizing extraction of the spatial relationship. According to the method, the spatial relationship extraction problem in the field of water conservancy is concerned, automatic identification of the spatial relationship, construction of a spatial relationship word set, acquisition of a spatial relationship syntactic mode and extraction of a spatial relationship tuple are realized by utilizing a weak supervision method, and a large amount ofmanpower and time are saved. Water conservancy data resource extraction oriented to the spatial relationship is realized, free texts in the water conservancy field are converted into structured data,and large-scale and professional spatial relationship supplementation is carried out on the atlas, so that more accurate query service is provided for users.

The invention is applied to the field of databases and provides a data extraction method, a data extraction device and a data management system. The data extraction method comprises the following steps: detecting whether basic data which does not establish a mapping relationship with a template account set is present in a sub account set; when detecting that the basic data which does not establish the mapping relationship with the template account set is present in the sub account set, establishing a mapping relationship between the basic data which does not establish the mapping relationship with the template account set and the template account set according to the configured automatic mapping rule; and extracting the basic data in the sub account set according to the mapping relationship. The data extraction method, the data extraction device and the data management system realize automatic data extraction of a database to the sub account set, avoid fussy and repeated manual operation, reduce the error rate and implementing cost, and guarantee timely and complete data extraction simultaneously.

The invention relates to an intelligent administrative law enforcement case information extraction and cause affirmation method. The method comprises the following steps: A, constructing a data set; c, preprocessing and vectorizing extracted data; D, training an extraction model based on a DGCNN; E, generating a short text based on UniLM, and training a generation model; f, inputting a law enforcement document text for prediction: processing an administrative law enforcement document needing to be converted through the steps A to C, inputting an obtained processing result into the trained extraction model, generating an extraction result, storing the extraction result, and further inputting the extraction result into the trained generation model to obtain a short text suitable for analysis of other downstream tasks. According to the method, the DGCNN is adopted as an extraction model, and the information of the text data is extracted by using the non-serialized neural network structure characteristics, so that the consumed time resources and computing resources are greatly reduced, and the extraction accuracy is improved.

The invention discloses an endocrine dyscrasia effusion extractor, which relates to the technical field of medical instruments and comprises a main body, a groove is formed in one side of the main body, a fixed column is fixedly connected to one side of the groove, a gear is movably connected to the other side of the fixed column, a threaded rod is fixedly connected to the other side of the gear, and a limiting block is fixedly connected to the outer surface of the main body. A rotating ring is movably connected to one side of the body, and a limiting groove is formed in one side in the rotating ring. According to the endocrine dyscrasia effusion extractor, the limiting block and the limiting groove are matched with each other, so that the rotating ring can be limited, and the rotating ring can rotate at a fixed position; then the rotating ring rotates to drive the inner gear ring and the gear to be meshed with each other, so that the threaded rod and the threaded hole are driven to be matched with each other, the movable rod can drive the piston to move, so that hydrops can be extracted. According to the invention, a traditional push-pull extractor is changed into a rotating extractor, so that the hydrops extractor can be used more smoothly and more conveniently.

The invention relates to an element extraction method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a to-be-recognized text corpus, inputting the to-be-recognized text corpus to an element extraction model, and carrying out named entity recognition of the to-be-recognized text corpus through a long text element extraction branch of the element extraction model, acquiring a long text extraction result for extraction of long sentences; and performing named entity recognition on the to-be-recognized text corpus through a short text element extraction branch of an element extraction model to obtain a short text extraction result, and performing constraint processing according to the long text extraction result and the short text extraction result to obtain structured data of the to-be-recognized text corpus. The combination of the long text extraction result and the short text extraction result is realized, the element extraction accuracy can be improved to 97% or above, and the problem of element overlapping can be solved.

The invention relates to a paint extractor which is used for extracting paint in a container. The paint extractor comprises a support seat, a mounting seat, a roller wheel and a hose pipe, wherein the support seat is provided with an arc-shaped guide part; the mounting seat is rotatably arranged on the support seat; the roller wheel is rotatably arranged on the mounting seat; the hose pipe is provided with an inlet and an outlet and is positioned between the arc-shaped guide part and the roller wheel; the hose pipe is extruded by the roller wheel; the inlet of the hose pipe is arranged in the container, and the outlet of the hose pipe is arranged outside the container; the paint extractor also comprises a power tool which is operationally disconnected with or in match connection with the mounting seat so as to drive the mounting seat to rotate; the roller wheel is driven to repeatedly extrude the hose pipe by the rotation of the mounting seat, so that the paint is extracted. The paint can be extracted without opening a cover of a paint bucket, so that the paint extractor is very convenient to operate. Furthermore, a mouth of the paint bucket is maintained to be clean, so that the trouble of cleaning is greatly eliminated.

The invention discloses a mining system for intelligently extracting data, including a system management module, the system management module is bi-directionally connected with a multi-channel data classification and extraction system, the multi-channel data classification and extraction system comprises a time data classification and extraction module, a data format classification extraction module and a data size classification extraction module, wherein an output end of the time data classification extraction module is connected with an input end of the data format classification extractionmodule, and an output end of the data format classification extraction module is connected with an input end of the data size classification extraction module, which relates to the technical field ofdata mining. The intelligent mining system of extracting data can realize the multi-channel classification extraction of the data to be extracted, and the purpose of increasing the pertinence and accuracy of data extraction is well achieved, so that the user can obtain the required data more quickly and accurately, thus greatly facilitating the data mining work of the user.

The invention discloses an oxygen bomb and an oxygen bomb air exhaust joint. The oxygen bomb air exhaust joint comprises a joint body; a through hole for arranging and sealing the periphery of an oxygen bomb head of the oxygen bomb in a sleeving manner is formed in the middle part of the joint body; and the side wall of the through hole is provided with a joint air pipe which extends in the radialdirection and is arranged in a penetrating way. The oxygen bomb air exhaust joint is suitable for the oxygen bomb assembled by a needle valve with a solid structure and the oxygen bomb head providedwith a connecting air pipe on the side wall, and is used for realizing the vacuumizing operation of the oxygen bomb. During vacuumizing, firstly, the joint body is arranged on the oxygen bomb head ina sleeving manner; then, the needle valve is unscrewed with respect to the oxygen bomb head, so that the oxygen bomb air pipe, the connecting air pipe and the joint air pipe communicate with each other in sequence; and then, the oxygen bomb air exhaust joint is connected with vacuumizing equipment to realize the vacuumizing of the oxygen bomb. After vacuumizing, the needle valve is screwed with respect to the oxygen bomb head, and the oxygen bomb air exhaust joint is taken down. The oxygen bomb air exhaust joint is suitable for the oxygen bomb and the needle valve thereof, and is high in structural strength, corrosion resistance, excellent in tightness and long in service life.

The invention relates to a device (5) and to a corresponding method for separating value documents (2), in particular banknotes, and to a value document processing system, wherein the device (5) has:an input means (4) which is designed to receive a stack (3) of value documents (2), a transport means (6), which is designed to withdraw a value document (2) from the stack (3) of value documents (2)located in the input means (4), and a transport element (7), which is designed to influence the withdrawal of the value document (2) from the stack (3) of value documents (2). Also provided is at least one sensor (8), which is designed to to measure a size of the stack (3) located in the input means (4), in particular a height of the stack (3) and/or a filling level of the input means (4), and a density of the stack (3). A control means is designed to control the transport means (6) and/or the transport element (7) on the basis of the measured size of the stack (3) and/or the measured densityof the stack (3).