111 results about "Entity relation extraction" patented technology

The invention discloses a system and method for constructing information-analysis-oriented knowledge maps. The system comprises a data acquisition module, a text extraction module, an entity recognition module, a semantic analysis module and an entity-relation extraction module, wherein the data acquisition module is used for carrying out cleaning and simple preprocessing on acquired data and outputting the data to the text extraction module; the text extraction module is used for carrying out data cleaning and preprocessing on structured and unstructured data and conveying clean data to the entity recognition module; the entity recognition module is used for segmenting words of a text, marking the word characteristics of the segmented words, then extracting terms and conveying extracted results to the semantic analysis module; the semantic analysis module is used for analyzing and extracting relation among bodies, generating a semantic metadata model by a body construction tool and outputting the semantic metadata model to the entity-relation extraction module; and the entity-relation extraction module is used for finally generating knowledge map language by extracting taxonomic relation and non-taxonomic relation. The system and method disclosed by the invention have the advantages that by combination of syntactic training and association rules, not only are external input and artificial intervention reduced, but also the entity relation can be continuously recognized.

The invention discloses an open Chinese entity relation extraction method using dependency analysis. According to the method, firstly, sentences are subjected to dependency analysis; then, a Chinese grammar heuristic rule and the dependency analysis result are combined for extracting relation words; next, the named entity position is determined according to the distance; and finally, the triple output is carried out. The experiment is carried out on SogouCA and SogouCS language databases. The result shows that the method provided by the invention is applicable to large-scale language databases, and has good transportability. The method provided by the invention fundamentally overcomes the limitation of intrinsic properties of complicated Chinese grammar, diverse expression modes, rich semantics and the like.

The invention relates to a text data-oriented threat intelligence knowledge graph construction method. The functions of automatically extracting the key information from the text threat intelligence data and constructing the threat intelligence knowledge graph are realized. The invention provides a threat intelligence knowledge graph construction method for text data. The method comprises the following steps: defining an ontology structure in the threat intelligence field; using a threat intelligence named entity recognition model based on multiple factors and a threat intelligence entity relation extraction model based on a graph neural network to obtain threat intelligence entities and relation triples from text data, and finally storing information through a graph database to form a threat intelligence knowledge graph.

The invention discloses a method for extracting relations among named entities in Internet massive data. The method comprises the following steps: crawling network information and constructing a corpus; preprocessing texts; extracting keywords representing short document features; acquiring 'entity-relation modes' representing entity relations; annotating the relations, and finding new 'entity-relation pairs' from a large quantity of unstructured texts by using the modes; and evaluating the entity-relation pairs. The invention also discloses a system for implementing the method for extracting the relations among the named entities in Internet massive data. The system comprises a network information crawling module, an information preprocessing module, a feature word extraction module and an entity relation extraction and evaluation module. The method and the system have the advantages of greater convenience in extension of a relation lookup system, high running efficiency and the like.

The invention provides a control method of Chinese entity relationship extraction based on word co-occurrence, which comprises the steps of: a. calculating a word correlation degree by statistics of co-occurrence frequency of words; b. calculating word similarity according to the word correlation degree; and c. determining an entity relationship according to the word similarity. The invention also provides a corresponding control device. A corpus used by the method is a news corpus, so that the corpus can be directly created by extracting news texts and titles via the current mature webpage analysis technology without a large amount of manpower participating in the corpus creation; and the method is capable of obtaining such information as word frequency used for calculating the word correlation degree, word position considered when calculating the matching similarity, part of speech of the words and whether the words are verbs, and the like, by utilizing a shallow language rule, for example, participle and part of speech marking in natural language processing and simple statistical techniques, and the method can combine the semantic information of the words with a traditional mode matching method.

The invention discloses an electronic medical record entity relation extraction method and apparatus, and belongs to the field of medical data mining. The method comprises the steps of obtaining a matrix after electronic medical record natural statement mapping through a convolutional neural network model and word vectorization representation; inputting tested electronic medical record natural statements to the trained convolutional neural network model to obtain eigenvectors; and inputting the eigenvectors to a trained classifier, and extracting an entity relation of the tested electronic medical record natural statements. Therefore, the advantages of the convolutional neural network model are utilized, the relation among entities in the electronic medical record natural statements is mined, and a technical way is provided for automatically learning electronic medical record information.

The invention relates to a method for automatically extracting character relations from a Chinese text or a text set, and belongs to the technical field of computer science and information extraction. In the method, by means of sentence meaning model characteristics, the relation attribute affiliation is determined; the character relations scattered in the text or the text set are automatically extracted by combining methods such as relation attribute disambiguation, character relation strength calculation and the like; and the character relations are organized by a character relation network, and the character relations comprising character relation attributes and the relation strength are displayed in a character relation graph manner. According to the method, sentence meaning model characteristics are introduced, so that the accuracy of the method for extracting entity relations is improved, and the method for extracting the character relations is enriched. Besides, as the number of texts about a central character in the text set is increased, the method can extract character relations of the central character more accurately and comprehensively, and the application range is wider.

The invention provides a multi-entity relationship joint extraction method and device based on text generation, wherein the method comprises the steps of expressing each word in a sentence to be processed through a coding vector, and obtaining a word embedding vector of each word; performing feature extraction on the word embedding vector of each word, and obtaining a high-grade feature representation vector of each word; and decoding the advanced feature representation vector, generating a target entity or a relational word at each moment to obtain a generation sequence, and generating the words generated at each three consecutive moments in the generation sequence to form an entity relationship triple. According to the method, an entity relationship extraction task is converted into a text generation task, the entity and the relation words are used as the target text to be generated, and one or more groups of relation triplets are generated, so that the joint extraction of the entityand the relation is achieved, the entity can repeatedly appear in the multiple triplets, and the entity overlapping and entity relation extraction problems under the multiple relations are solved.

Methods and systems for trend aware self-correcting entity relationship extraction are disclosed. For example, a method can include receiving a selected entity, determining a plurality of entities related to the selected entity, determining a plurality of most probable entities, calculating relevance scores, and displaying a subset of the plurality of most probable entities. The selected entity can be received on a network-based transaction system. The plurality of entities related to the selected entity can be determined based on a relationship score. The relationship score can represent navigation transitions, aggregated over time, between the selected entity and each of the plurality of entities. The plurality of most probable entities can be determined based on probabilities. Relevance scores can be calculated for each of the plurality of most probable entities. Finally, the subset of the plurality of most probable entities to be displayed can be determined according to the relevance scores.

The invention discloses a medical knowledge graph generation method, a storage medium and a server. The method includes the steps that input medical information is received, and named entity identification is conducted on the input medical information to obtain entities in the medical information; entity relation extraction is conducted on the identified entities to obtain the corresponding relation between the entities; a medical knowledge graph corresponding to the medical information is generated according to the obtained entities and the corresponding relation between the entities. The method has the advantage of automatically generating the corresponding medical knowledge graph according to the input medical information; labor is saved, and convenience is provided for users.

The invention provides a semantic entity relation extraction method and device, and electronic equipment. The method relates to the fields of artificial intelligence and information extraction technology of natural language processing. The method comprises: identifying each word node of an input text; constructing a dependency characteristic of each word node; and when more than two word nodes arein a coordinating relation, extracting relational triples of candidate nodes by recursively calling pre-stored semantic rules; wherein the pre-stored semantic rules comprise a pre-modification structure rule and a verb-related rule. Compared with the prior art, the method and the device avoid extraction omission due to complicated rules and incomplete definitions by using a recursive method, andcan improve the accuracy of entity relation extraction.

The invention discloses an entity relation extraction method based on a neural network, using the algorithm of machine learning and neural network model, Input a Chinese sentence into the program model, the model will give a special label to the entity words or statements, that is, the entity in the text can be extracted, and then through a classification algorithm for the extracted entity to do relationship classification, entity relationship classification is completed. Specifically, assign an ID to each word that appears in the Chinese text, Then the IDs corresponding to these sentences aretransformed into input vectors of the neural network model, and the results obtained through bilstm and CRF layer are mapped to corresponding entity tags to complete entity extraction. Finally, the entities in the text are classified by machine learning classification algorithm, and finally such a triple form of an entity of the entities-relationsientities is obtained. . This method only needs training text and input statements to complete the extraction of relational entities, which is a flexible and convenient method.

The invention discloses an entity relation extraction method and device. All entities having target relations with each sub-entity are stored in an obtained complete knowledge base, that is, attributevalues of all the stored sub-entities are complete, language data matched with an alternative relation mode obtained through analysis as well as entity pairs contained in the language data are determined through a corpus on the basis, the alternative relation has no semantic drift if the entity pairs exist in the knowledge base and can be determined as the target relation mode, and then the entity pair of a target relation is extracted according to the target relation mode. Understandably, the alternative relation mode has semantic drift if the entity pairs do not exist in the knowledge base,then the alternative relation mode is abandoned, and error propagation caused by iteration on the basis of the alternative relation mode is avoided.

The invention discloses a system and a method for topic meta search based on unsupervised entity relation extraction. The system comprises a topic model building module which is used for building various topic models, a matching search engine module and a search result processing module, wherein the matching search engine module is used for matching suitable member search engines so as to conduct topic search in accordance with different topic models when a user selects topics which are required to be searched, and for search results returned by member search engines, the search result processing module is used for conducting entity relation extraction for characteristic words of search results by using the unsupervised entity relation extraction algorithm and for returning results meeting conditions to the user in accordance with the similarity of the extracted relation on calculation and topics. On the premise of the recall ratio, according to results of unsupervised relation extraction, the similarity of topics is determined and the high recall ratio is obtained.

The invention discloses a knowledge representation learning framework based on multi-class cross entropy comparison completion coding. The framework mainly comprises a semantic structure feature extraction module (S) and an automatic comparison completion coding module (G). The semantic structure feature extraction module (S) is responsible for extracting low-level and high-level semantic structure features from entities and relationships and fusing the low-level and high-level semantic structure features to obtain low-level and high-level semantic structure features; the automatic comparison completion coding module (G) is responsible for predicting an entity context vector, setting positive and negative samples and a sampling method (C3NCE) of the positive and negative samples, calculating a multi-class cross entropy comparison loss function, obtaining vector representation of knowledge graph entities and relations by optimizing a target function training model, and completing a triple completion task. The framework provided by the invention can quickly, stably and accurately complement the triple of the missing information in the knowledge graph, well completes the knowledge representation learning task, greatly improves the accuracy and efficiency of knowledge graph construction, and is wide in application prospect.

The invention discloses a joint extraction method for named entities and relationships in a judicial domain, which is an entity relation extraction method of a BILSTM network and an attention mechanism set based on a BERT pre-training language model, realizes joint learning of two tasks through parameter sharing, and fully utilizes the relation between the tasks to optimize a result. A BERT pre-training language model is selected to train word vectors to complete conversion work of the data set word vectors; more complete context feature information is acquired by using a BILSTM neural network so as to extract text depth word vector features; Finally, category labels of the characters are acquired through a softmax classifier to realize entity recognition, and an association relationship is judged between the current character and the previous character by utilizing an attention mechanism to realize combined extraction of the entity and multiple relationships.

The present invention relates to the technical field of artificial intelligence, and more particularly, to an entity relationship joint extraction method based on reinforcement learning. Firstly, theunstructured text, segmented words, and training word vectors for entity relation extraction can be obtained, and are input in LSTM by taking words as a unit, because the same entity in a sentence mayappear in different forms in different locations, and we do not know where the entities really useful for relationship extraction, so we can use reinforcement learning method to select these entities; after the entity selection is completed, if there is a consecutive one, we need to merge it into one entity. Finally, after removing the redundancy, if two entities are picked out, the word vectorsof these two entities and the sentence vectors of the final output of LSTM are stitched together, and the relations are classified by a fully connected neural network, otherwise the sentence is considered to be noisy.

The invention provides an entity relation extraction method. The entity relation extraction method includes the steps of firstly, marking a negative sample in a data set according to the description information of the entity, so as to divide the negative sample into a real negative sample and an uncertain sample; then, providing the indeterminate sample relation label to construct a new training set; finally, extracting the relationship features of the new training set according to the bi-directional gating loop unit to obtain the entity relationship. By labeling the negative samples of the dataset according to the entity description information, the dataset is effectively optimized. A new training set is constructed by endowing the indeterminate sample relation label to improve the accuracy of the training set and further improve the precision of the relationship of the extracted entities.