75 results about "Lexical correspondence" patented technology

The invention discloses a system and method for matching colors according to emotions of bullet screen contents, and relates to the field of Internet bullet screens. The method comprises the steps: enabling vocabularies, the number of appearance times of which exceeds a preset number, in a bullet screen to be stored in an emotion vocabulary library, and enabling Chinese privatives to be stored in an antiphrasis vocabulary library; setting an emotion for each vocabulary in the emotion vocabulary library, enabling the emotions to be divided into different levels, and enabling the vocabulary of each level to be corresponding to one weighted value; adding the weighted values of all vocabularies in the emotion vocabulary library in a bullet screen edited by a user, and multiplying the weighted value of the vocabulary existing in the emotion vocabulary library by (-1) if the vocabulary exists in the emotion vocabulary library and a vocabulary before the vocabulary exists in the antiphrasis vocabulary library; dividing the total weighted value by the number of weighted values, and obtaining a mean weighted value; enabling each mean weighted value to be corresponding to one color, and displaying the bullet screen edited by the user in the corresponding color. The method can achieve the automatic matching of bullet screen colors according to the preset emotion vocabularies, is simple and convenient, and improves the interestingness of the bullet screen.

The invention relates to the technical field of electronic equipment, in particular to a dictation reporting and reading method and device, and the method comprises the steps: determining outline vocabularies matched with the current age of a user, the outline vocabularies being vocabularies needing to be mastered by a user group of the age group to which the current age belongs; obtaining audio data corresponding to the outline words from the dictionary audio data, and forming a voice reading audio data set according to the audio data corresponding to all the outline words, wherein the audiodata at least comprises pronunciation audios of the outline vocabularies; when a dictation instruction triggered by a user is received, a dictation task is read from the dictation instruction, and thedictation task comprises at least one to-be-listened sketch word; and obtaining audio data corresponding to the to-be-listened sketch words from the voice reading audio data set, and performing reading based on the audio data corresponding to the to-be-listened sketch words. By implementing the embodiment of the invention, the dictation effect of the user can be effectively improved.

The invention discloses a synonym data mining method and system. The method includes the steps of: obtaining word pairs in a dictionary, a video file library and a search log record and similarity values of the word pairs, and building a candidate synonym library where the word pairs and the similarity values are associated; according to data information in the candidate synonym library, training and obtaining a synonym model; substituting the similarity value corresponding to each word in the candidate synonym library into the synonym model to obtain an output numerical value; and storing the word pairs whose output numerical values are larger than a preset threshold value in the synonym library. Thus, the synonym data mining method and system solve the problem that video file watching limits for different watching groups during media playing cannot be realized.

The embodiments of the invention provide a punctuation adding method and device and a device for adding punctuations. The method specifically comprises the steps of: acquiring a text to be processed; and adding punctuations for the text to be processed via a neural network conversion model to obtain a punctuation adding result corresponding to the text to be processed, wherein the neural network conversion model is obtained by training parallel corpuses, the parallel corpuses comprise a source corpus and a destination corpus, and the destination corpus comprise punctuations corresponding to words in the source corpus. The punctuation adding method and device can improve the punctuation adding accuracy.

The invention relates to a method for generating a document abstract. The method includes the steps that a document set is preprocessed, a vocabulary set is processed through a latent Dirichlet modelor a vector space model, weights corresponding to vocabularies are obtained, all the weights of the vocabularies corresponding to each sentence in a sentence set are added, and corresponding internalinformation amount scores are obtained; according to preset similarity threshold values, similar sentences and the number of the similar sentences corresponding to each sentence are determined, corresponding importance scores are obtained through calculation, the numbers of the similar sentences of all the sentences are compared with the numbers of the similar sentences corresponding to all the similar sentences of all the sentences, diversity scores of all the sentences are obtained through calculation, and then comprehensive scores of all the sentences are obtained through calculation; finally, screening is conducted according to the comprehensive scores of all the sentences and preset abstract length, and the document abstract is generated. In addition, the invention provides a device for generating the document abstract. According to the method and device for generating the document abstract, the redundancy of the abstract is generally reduced.

The invention relates to a keyword automatic extension querying method and the device thereof; the method includes the steps: 1) a database is created and includes keywords, vocabularies and identification codes; 2) the keyword is corresponding to at least a word; 3) the related keyword is corresponding to an identification code;4) the identification code in the database corresponding to the keyword is determined according to the keyword inputted by the user; 5) the related keyword corresponding to the identification code is extracted through the identification code; 6) the words corresponding to the related keyword are queried through the related keyword. The invention provides a keyword automatic extension querying method and the device thereof which solve the problems that only the word with the query character string can be queried; the synonym of the queried character string cannot be inputted; no relative queried character string is provided for the user; the querying device cannot provide the expected query result for the user quickly found in the prior art.

The invention discloses a vehicular site context anaphora resolution method and device, relates to the technical field of natural languages and mainly aims at improving the vehicular site context anaphora resolution accuracy. The method provided by the invention comprises the steps of resolving first vocabularies in first statement information from the first statement information, wherein the first vocabularies are anaphora vocabularies referring to site entities; searching second vocabularies corresponding to the first vocabularies from an entity database according to scene information corresponding to the first statement information, wherein the scene information is determined according to context content of the first statement information, the second vocabularies are site entity vocabularies having corresponding relationship with the first vocabularies, and the second vocabularies and the attribute information corresponding to the second vocabularies are stored in the entity database; and replacing the first vocabularies by the second vocabularies, thereby generating second statement information. The method is used for the vehicular site context anaphora resolution.

The invention discloses a field conversion method for an input method and a client. The field conversion method comprises the following steps: receiving an input field based on mandarin and displaying the field in a client interface; splitting the field into a plurality of vocabularies according to natural semanteme, searching characteristic vocabularies corresponding to the vocabularies obtained by splitting by traversing in a special word bank according to the vocabularies, and correspondingly storing the vocabularies based on mandarin and the corresponding characteristic vocabularies by the special word bank in a character string manner; associatively displaying the searched characteristic vocabularies corresponding to the vocabularies obtained by splitting at the periphery of the vocabularies in the client interface for selection and confirmation; receiving the selected and confirmed characteristic vocabularies corresponding to the vocabularies, replacing the corresponding vocabularies in the field with the characteristic vocabularies, and displaying the replaced field in the client interface. According to the method, the characteristic vocabularies can be output by inputting the vocabularies based on mandarin, so that any person who is unfamiliar with the characteristic vocabularies or a non-mandarin input method can conveniently write the characteristic vocabularies.

The invention relates to a method and a system for extracting and displaying vocabularies of digital publications. The method comprises: inputting a digital publication; counting total vocabulary amount in the digital publication and word frequency of each vocabulary, sorting all the vocabularies in the digital publication according to values of word frequency, and adding vocabulary information to form a glossary index and adding related information of the vocabularies to form a related data table; or sorting the vocabularies in a word frequency dictionary according to values of word frequency, according to the sorted vocabulary sequence in the word frequency dictionary, extracting the corresponding vocabularies in sequence in the digital publication, to obtain a glossary index and a related data table; according to the glossary index, determining the number of word frequency sections and the number of vocabularies included in each word frequency section, and displaying the numbers; making vocabulary information and related information corresponding to the vocabularies which are included in each word frequency section form a data package, the packages being used for downloading and learning. The method and the system can reduce language barrier caused by new words in reading, and improve reading quality and improve learning efficiency of vocabularies.

The invention provides a power grid dispatching professional language semantic relation extraction method and device and electronic equipment, and the method comprises the steps: collecting a dispatching history corpus generated in the operation process of a power grid; extracting vocabularies from each corpus text in the scheduling history corpus to obtain a plurality of vocabularies contained inthe corpus text, and constructing a vocabulary vector corresponding to each vocabulary according to the arrangement sequence of the plurality of vocabularies in the corpus text; Constructing a semantic vector corresponding to each vocabulary based on the vocabulary vector corresponding to each vocabulary and a preset neural network model; and calculating semantic similarity among the vocabulariesaccording to the semantic vector corresponding to each vocabulary so as to determine a semantic relationship among the vocabularies. According to the method, the semantic relation of the power grid dispatching major can be quickly and accurately extracted by means of the neural network model, the subjective influence of dispatching personnel is avoided, and the work load of the dispatching personnel is reduced.

The invention relates to a method for generating business rules, an electronic device and a readable storage medium. The method comprises steps: extracting preset identification words and preset key words in the natural language rules; determining a Java rule template corresponding to a preset identification vocabulary in the natural language rule according to a preset identification vocabulary and a mapping relationship between the preset identification vocabulary and the Java rule template, and transforming the preset key words in the natural language rule into corresponding codes, and filling the transformed codes into corresponding positions in the determined Java rule template; the populated Java rule template being compiled to generate executable Java rules. The invention does not need to be manually written by a professional developer, the human cost is lower, and the efficiency of generating business rules is improved.

The embodiment of the invention provides a video semantics labeling method based on a bullet screen. The method includes: obtaining all words in the bullet screen of a target video and corresponding time stamps; averagely dividing the target video into a preset number of time slices; generating an initial topic set, which contains topics corresponding to all the time slices, and an initial plot set, which contains plots corresponding to all the time slices, according to preset probability correspondence relationships of words and the topics and the plots; generating a dictionary vocabulary setand a vocabulary distribution matrix; calculating temporal a priori information of the dictionary vocabulary set; using a preset total probability formula of bullet-screen vocabulary to calculate probability that each piece of the dictionary vocabulary corresponds to each topic and each plot; generating plot-topic distribution matrices of the time slices; merging adjacent similar time slices intoone time slice; determining plots corresponding to all time slices; and labeling the target video. By applying the scheme provided by the embodiment of the invention for video semantics labeling, labeling on video semantics is enabled to be more accurate.

The present disclosure provides a medical file processing method and device. The medical file processing method includes the steps: determining a plurality of character strings corresponding to a plurality of description classes according to a preset file format corresponding to the medical file; segmenting the plurality of character strings and labeling the vocabulary categories according to a preset vocabulary, and recording the symptom vocabulary and diagnosis vocabulary; identifying the character string corresponding to the description class according to a sentence pattern template corresponding to the description class, and confirming the symptom vocabulary in the vocabulary of the unlabeled vocabulary classes; recording the vocabulary relation between the symptom vocabulary each of the description classes and the vocabulary of the diagnosis vocabulary. The medical document processing method provided by the present invention can analyze the relation between symptoms and diagnosisbased on a large number of medical record files.

An inputting method having an assistant translation function is used for giving a suggestion about related contents when translating. The method is characterized by reading a sentence to be translated; recognizing all likely terms in the sentence to be translated; acquiring translation of the terms from a dictionary of the field corresponding to the sentence; acquiring translation of non-term words from a common bilingual electronic dictionary, establishing a high-priority tree, and placing probability of translation result-to-word correspondence in the high-priority tree; simultaneously matching a normal-word tree and the high-priority tree in the input method when the user inputs letters, and comprehensively sequencing the match result according to the probability; clearing a high-priority tree word list of the sentence when completing inputting the sentence; and repeating the above steps until completing inputting all sentences to be translated. The invention integrates word stocks of professional fields to the inputting method, and effectively improves the inputting efficiency of texts in a translation process.

The invention provides a word vector training method and apparatus, and belongs to the technical field of machine learning. The word vector training method comprises the steps of obtaining a newly added vocabulary library, wherein vocabularies in the newly added vocabulary library and vocabularies in an old vocabulary library form a new vocabulary library, and the vocabularies in the old vocabulary library correspond to old word vectors; performing initialization processing on the vocabularies in the new vocabulary library, thereby enabling the word vectors of the vocabularies, belonging to the old vocabulary library, in the new vocabulary library to be the old word vectors, and enabling the word vectors of the vocabularies, belonging to the newly added vocabulary library, in the new vocabulary library to be random word vectors; and updating the word vectors of the vocabularies in the new vocabulary library according to a first Huffman tree corresponding to the new vocabulary library and a second Huffman tree corresponding to the old vocabulary library respectively. According to the word vector training method and apparatus provided by the invention, the training efficiency of the word vectors is improved.

The invention discloses a sentence pattern recommendation method and device, an electronic device and a computer-readable storage medium. The method includes: searching a plurality of similar words corresponding to an input word from a similar word map; s for each similar word, the similar words and the historical input text of the input words are spliced to obtain the candidate sentence patternscorresponding to the similar words; The reasonableness of candidate sentences is calculated; According to the reasonableness of the candidate sentence patterns, the reasonable sentence patterns are selected for recommendation. The invention searches the similar words corresponding to the input words from the similar word map, splices the similar words and the history input text to obtain the candidate sentence pattern, further recommends the candidate sentence pattern with high reasonableness to the developer, and inspires the developer to carry out sentence pattern configuration. Thus, the workload of the developer for sentence configuration is reduced, the sentence configuration is more comprehensive, and in the interactive system, all the user instructions can be matched to the appropriate sentence, the user intention can be accurately understood, and the user instructions can be executed.

A semantic analysis method, semantic analysis and non-transitory computer-readable medium are provided in this disclosure. The semantic analysis method includes the following operations: inputting a voice and recognizing the voice to generate an input sentence, wherein the input sentence includes a plurality of vocabularies; selecting at least one key vocabulary from the vocabularies according to a word property corresponding to each vocabulary; establishing a parse tree according to the input sentence and finding a plurality of associated sub-sentences; calculating an associated feature vector between the associated sub-sentences; concatenating the associated feature vector and the vocabulary vector corresponding to each vocabulary to generate a vocabulary feature vector corresponding to each vocabulary; and analyzing the vocabulary feature vector to generate an analysis result by a semantic analysis model, wherein the analysis result includes a slot type corresponding to each vocabulary and an intent corresponding to the input sentence.

The embodiment of the invention provides a text similarity calculation method and device. According to the embodiment of the invention, vocabularies in two texts are trained by using a first vocabulary vector training model to obtain a vocabulary vector corresponding to each vocabulary, then the cosine similarity of two vocabulary vectors is calculated, and finally the similarity of the two textsis calculated by using the maximum cosine similarity of the word vocabularies. The vocabulary vector comprises context information of the corresponding vocabulary, so that the cosine similarity of thevocabulary vectors can reflect the meaning similarity of the corresponding vocabularies, thus the similarity, which is calculated by using the cosine similarity, of the two texts can accurately reflect the meaning similarity of the two texts, that is, the accuracy of calculation for the text similarity can be improved by using the cosine similarity, and thus the limitations brought about by a circumstance that the similarity of two texts can only be determined by using the same vocabularies in the prior art are overcome.

The embodiment of the invention discloses a voice processing method, a device, a storage medium and electronic equipment. The method includes the following steps: collecting an input appraisal voice set, and identifying each pronunciation and vocabulary in the appraisal voice set; acquiring a text vocabulary corresponding to target pronunciation and vocabulary by a transliteration method when it is detected that there exists the unrecognizable target pronunciation and vocabulary in the appraisal voice set; and outputting the text vocabulary. Thus, by the embodiment of the invention, intelligence for pronunciation and vocabulary identification can be improved by identifying the text vocabulary corresponding to the unrecognizable target pronunciation and vocabulary.

The invention discloses a text sentiment analysis method, an electronic device and a storage medium, and belongs to the technical field of natural language processing. The method comprises the following steps: acquiring an input text, wherein the input text comprises a plurality of first vocabularies; performing semantic analysis on the input text to obtain a plurality of vocabulary feature vectors, and performing dependency syntax analysis on the input text to obtain dependency syntax information corresponding to each first vocabulary; performing weighted calculation according to the plurality of vocabulary feature vectors and the dependency syntax information corresponding to each first vocabulary to obtain a plurality of target feature vectors; and performing sentiment classification based on the plurality of target feature vectors to obtain a sentiment analysis result corresponding to the input text. According to the technical scheme, the dependency relationship between vocabularies in the input text can be fused, and the accuracy of text sentiment analysis can be improved by adopting the syntactic information to carry out weighted recognition on noise in the syntactic information.

The invention relates to a news automatic labeling method based on an LDA model. The news automatic labeling method includes the steps: extracting text data features at a semantic level, and having better effect in practical application; proposing improvements for an LDA model, utilizing point mutual information for quantizing the subject term relation, obtaining the co-occurrence relation betweensubject terms by calculating the weights of the subject terms, and setting a threshold value to select the optimal subject term. For the news automatic labeling method, keywords with high accuracy are selected according to the strength of the representation ability of vocabularies corresponding to different topics, and mutual information can be introduced to improve a topic-lexical item matrix, so that the accuracy of an LDA model in news document automatic label application is improved, and the correlation between subject terms is better described.

The invention provides a statement similarity determination method and device, an electronic device and a readable storage medium. The method comprises the following steps of: determining first semantic feature vectors corresponding to the first vocabularies in the input statement and first weights of the first semantic feature vectors, and determining a second semantic feature vector corresponding to each second vocabulary in the standard statement and a second weight of each second semantic feature vector, and calculating the similarity between the input statement and the standard statementbased on the first semantic feature vector, the first weight, the second semantic feature vector and the second weight. According to the method and the device, the semantic dependency relationship ofthe vocabularies far away from each other is represented through the first weight and the second weight, so that the similarity between the input statement and the standard statement is determined, the statement similarity determination accuracy is improved, the condition that the intelligent statement responds not to questions in the response process is reduced, and the intelligent statement response accuracy is improved.

The invention relates to the field of big data, and discloses an index code generation method, a device, equipment and a storage medium. The method comprises the steps of obtaining an index name of aservice demand; performing semantic word segmentation processing on the index names by adopting a preset natural language processing model to obtain a plurality of index vocabularies; obtaining a multi-dimensional index table and an index dimension code mapping table, comparing the multi-dimensional index table with each index vocabulary, and determining a dimension index and a measurement index corresponding to each index vocabulary; searching atomic index codes corresponding to each dimension index and each measurement index from the index dimension code mapping table, and sorting the atomicindex codes according to a preset rule; and splicing the atom index codes based on the sorting sequence of the atom index codes to obtain the index codes of the demand index names. The invention alsorelates to a blockchain technology, and the index code is stored in the blockchain. According to the invention, the standardization degree of the business index management architecture is improved.

An instant translation system and method is provided, which is suitable for an electronic information processing platform with word database. Users set establishment conditions of new word database in advance and extract all the word data conforming to the set conditions from the word database and store the data in the new word database, thus when users pick up target inquiring data of translation operation to be executed, analyzing all the target inquiring words included in the target inquiring data based on the data stored in the new word database, extracting the separately corresponding word meaning explanation data of all the target inquiring words from the word database and outputting and displaying the extracted word meaning explanation data corresponding to target inquiring words for reference of users, thereby realizing the effect of instant group translation, not only improving the data translation efficiency, but also facilitating operation of users and making instant translation functions more humanized.

A patent data retrieval system comprises a database for storing a relation between specialized vocabulary and patent classification number, a determination patent classification number module for determining a patent classification number corresponding to a specialized vocabulary input by the searching staff according to the relation in the database, and a searching module for retrieval in the patent database according to the determined patent classification number to find related patent document, a display module for displaying the patent document as a reference to the searching staff; the system is capable of helping the searching staff search with a patent classification number through determining the classification number of the specialized vocabulary input by the searching staff.

A patent data retrieval system comprises a database, a patent class number determining module, a retrieval module and a display module, wherein the database is used for storing corresponding relations between specialized words and patent class numbers, the patent class number determining module is used for determining the patent class numbers corresponding to the specialized words according to the corresponding relations in the database and the specialized words input by retrieval personnel, the retrieval module is used or carrying out retrieval in a patent database according to the determined patent class numbers to retrieve related patent documents, and the display module is used for displaying the patent documents so as to provide a reference for retrieval personnel. Thus, the patent data retrieval system can help retrieval personnel to carry out retrieval through the patent class numbers by determining the patent class numbers of the specialized words input by retrieval personnel.

The invention provides a measuring method and device for creative personality traits based on an electroencephalogram signal. The measuring method includes the steps of segmenting vocabulary of a natural language material, and obtaining the occurrence probability of each segmented vocabulary in the context of a text; obtaining an electroencephalogram signal of a subject when hearing the audio of the natural language material, segmenting the electroencephalogram signal, and obtaining a corresponding electroencephalogram response segment of each vocabulary; obtaining an impact response functionof the occurrence probability according to the occurrence probability of each vocabulary and the corresponding electroencephalogram response segment of each vocabulary; and obtaining the test score ofthe personality traits of the subject based on a pre-trained creativity personality traits prediction model according to the impact response function. According to the measuring method and device, automatic measurement of the creative personality traits is achieved, effects of external factors are not prone to appearing, and measurement is more accurate.

The invention belongs to the technical field of computer information processing, and provides a natural language processing method and apparatus, an electronic device and a computer readable medium. The method comprises the steps of performing word segmentation processing on characters in text data to obtain characters and / or vocabularies; inputting the text data and the domain attributes corresponding to the text data into a character vector model to obtain a character vector; inputting the text data and the corresponding domain attribute into a vocabulary vector model to obtain a word vector; determining a first weight corresponding to the text and / or a second weight corresponding to the vocabulary based on the text data; determining a sentence semantic vector of the text data through the character vector, the first weight and / or the word vector and the second weight; and performing natural language processing on the real-time text data based on the sentence semantic vector. According to the method, the semantic expression capability of sentences can be effectively improved.

The invention discloses an intelligent keyword extraction method and device, computer equipment and a storage medium; the method comprises the steps: converting an initial text inputted by a user into text coding information, obtaining a statement vector matrix of each statement change, extracting a vocabulary vector from the statement vector matrix, and whitening the vocabulary vector to obtain a standard unit vector corresponding to each vocabulary vector; then calculating the similarity between the standard unit vector and the statement identification vector of the corresponding statement vector matrix, and screening the word segmentation result of the initial text according to the similarity calculation result to obtain a target vocabulary meeting the vocabulary screening rule as a keyword extraction result. The method belongs to the technical field of semantic analysis, can accurately obtain the standard unit vector corresponding to the vocabulary in the initial text, and extracts the target vocabulary from the initial text as the keyword extraction result based on the similarity between the standard unit vector and the statement identification vector of the corresponding statement vector matrix. Therefore, the accuracy of keyword extraction from the text is greatly improved.