68 results about "Feature term" patented technology

A method and system for extracting opinions about a subject of interest from a text document in which each sentence is analyzed individually to identify the opinions. The most relevant feature terms related to the subject are extracted from the document based on their relevancy scores. Candidate feature terms are definite noun phrases at the beginning of the sentences. For each sentence that refers to the subject or a feature term, the invention determines whether the sentence includes an opinion polarity about the subject or the feature term. The opinion polarity is detected by identifying opinion terms in the sentence using an opinion dictionary or an opinion rule base, parsing the sentence with an English parser to identify grammatical components in the sentence and their relationships, and finding a matching entry in the dictionary or the rule base.

A method and system for extracting opinions about a subject of interest from a text document in which each sentence is analyzed individually to identify the opinions. The most relevant feature terms related to the subject are extracted from the document based on their relevancy scores. Candidate feature terms are definite noun phrases at the beginning of the sentences. For each sentence that refers to the subject or a feature term, the invention determines whether the sentence includes an opinion polarity about the subject or the feature term. The opinion polarity is detected by identifying opinion terms in the sentence using an opinion dictionary or an opinion rule base, parsing the sentence with an English parser to identify grammatical components in the sentence and their relationships, and finding a matching entry in the dictionary or the rule base.

The invention relates to a Chinese text parallel data mining method based on hierarchy, comprising the steps of: step 1: a establishing vector space model of Chinese texts: performing work segmentation regarding to the entire Chinese text set to obtain a word segmentation form and a feature term set containing all removed duplicated terms in the text set of each text, then using the feature term set to count the term frequency-inverse document frequency (TFIDF) of each text, and establishing the text vector space model according to the TFIDF; step 2: performing dimension reduction regarding to a feature item vector of the text vector space model; and step 3: clustering texts using DCURE algorithm based on hierarchy. The method is efficient in word segmentation of Chinese texts with high accuracy, requires no input of parameters like radius of neighborhood for the clustering process, can mine irregular cluster and is insensitive to noise, employs distributed calculating, has high efficiency in mining mass texts and improves calculating speed of feature weight.

The invention discloses a feature selection method based on document frequency of within-class and between-class and term frequency statistics. The document frequency, the word frequency, the between-class concentration ratio and the within-class dispersity of a feature term are comprehensively considered to construct a feature selection assessment function based on DFCTFS (Document Frequency of within-class and between-class and Term Frequency Statistics); and the original feature space, which is subjected to text preprocessing, of a training set uses a feature selection assessment function which is put forward by the invention to select a certain ratio of feature terms in each class of the training set to form the feature term bank of the class, and the feature term bank of the trainingset is the union set of each class of feature term bank of the training set. The invention puts forward the feature selection method based on the DFCTFS, feature terms which are intensively distributed in the certain class of document, are evenly distributed in the class of documents and frequently appear can be diagnostically selected, and a Chinese text classification effect can be improved.

The present invention relates to a system and method for information process using artificially constructed apparatus. More specially, the present invention provides a system and method that can search for information in a document structure and provide precise results by analyzing the inputs and search results using the executing system and the knowledge structure of the think system. In one preferred embodiment of the present invention, the search terms are divided into subject terms and corresponding feature terms, and document entry files comprising respective subject terms and corresponding feature terms will provide access to documents including subject terms and corresponding feature terms.

A model training method, apparatus, and computer device are disclosed. The method includes: determining a common feature space of the source domain sample set and the target domain sample set; determining a positive correlation feature term and a negative correlation feature term in the common feature space according to the tag value of the determined tag sample in the source domain sample set andthe eigenvalue of the determined tag sample in the common feature space; predicting a label value of an uncertain label sample in the target domain sample set according to the positive correlation feature item and the negative correlation feature item; uncertain label samples with predicted label values are integrated with the source domain sample set, and a classification model is obtained by training the integrated sample set.

The invention discloses an electrical equipment name identification method. The method comprises the following steps: (1) constructing a power grid professional lexicon for storing vocabularies; The method comprises the following steps: directly adding a single vocabulary, directly adding more than two vocabularies, and screening and adding vocabularies; (2) realizing word segmentation: generatinga triplet search tree from the professional lexicon, and combining the digital search tree with the binary search tree to realize quick word segmentation; obtaining a character string array or a character string list; (3) dividing the name of the equipment to be identified and the name of the standard equipment into a character string array or a character string list according to the step (2), and extracting characteristic words which possibly conform to a place to which the equipment belongs and a voltage level; (4) screening a standard equipment name database according to the feature words;carrying out similarity calculation on two character string arrays obtained after word segmentation is carried out on the name of the to-be-identified equipment and the name of the standard equipment, and acquiring similarity value between 0-1; for the similarity value between 1 and 2, judging the character string meeting the condition by setting a threshold value, and selecting a corresponding data entry; And equipment recognition degree identification is realized.

The present invention discloses a method for acquiring a network service status based on microblog big data. The method comprises: using a part of microblogs of a microblog dataset as a training dataset, using the remaining microblogs as a testing dataset, and preprocessing the training dataset and the testing dataset; performing marking, initialization operation, word partitioning and word pausing on training data, performing feature selection on the training dataset to obtain a feature term dictionary, generating feature vectors according to the feature term dictionary to obtain a feature vector set, and performing training on the feature vectors to obtain an SVM classifier; and acquiring a preset keyword library; presorting testing data, performing initialization operation, word partitioning and word pausing on testing data of which the presorting fails; according to the feature term dictionary, generating feature vectors of the testing data of which the presorting fails, to obtain a feature vector set; performing classification by using the SVM classifier to obtain a classification result, and integrating the classification result and a presorting result. The method effectively reduces scale and complexity of network big data.

The invention relates to the field of artificial intelligence, and provides a question and answer corpus generation method and device based on a text generation model, computer equipment and a storagemedium. The method comprises the steps of obtaining historical questions and a standard document, extracting keywords in the standard document and paraphrasing sentences corresponding to the keywords, performing word segmentation processing on the historical questions, identifying and discarding entity nouns in the historical questions to obtain syntactic feature words of the historical questions, combining the syntactic feature words with the keywords, and inputting the combined data into a pre-trained text generation model to obtain a target question corresponding to the keyword, wherein the text generation model by training based on a training sample marked with the keyword and syntax feature words are obtained, and according to the target question corresponding to the keyword and a paraphrasing statement corresponding to the keyword, a question-answer pair comprising the target question sentence and the paraphrasing sentence is constructed so as to improve the quality of the target question sentence and the question-answer pair.

The invention relates to the field of natural language processing, in particular to a term definition discriminating and analysis method based on the Internet. The problems that one term has multiple definitions, and the definition standardability and accuracy are poor are mainly solved. According to the technical scheme, the method is characterized by comprising the steps of obtaining term definitions to be discriminated and analyzed and a reference paraphrase, expressing the term definitions, calculating similarity, obtaining a term definition template, calculating term definition reliability and selecting a discriminating and analysis result. The constructed reference paraphrase considers the characteristics of term definition accuracy and professionality, a quintuple expression method of the term definitions is utilized for calculating the term definition similarity, the similarity among term definition feature words and semantic similarity among the definitions are considered, and the similarity among the term definitions is described better. The matching template of the term definitions is concluded to adjust the similarity among the term definitions, and the reliability of the term definitions is more accurate. According to the method, the good discriminating and analysis effect is achieved, and the problem that the term definitions are nonstandard and inaccurate can be solved.

The invention relates to the field of big data, and discloses an official document recommendation method and device based on a graph structure, computer equipment and a medium. The official document recommendation method comprises the steps: acquiring multiple official documents, screening feature words according to TF-IDF, and recording the feature words as keyword tags of the official documentscorresponding to the feature words; screening out text topics of which the selection probabilities are greater than or equal to a preset probability through the text topic-keyword distribution probability matrix of the official document, and recording the screened text topics as topic tags of the official document corresponding to the screened text topics; generating official document attributes according to the keyword tags and the topic tags; acquiring record data of the official document, and establishing an official document recommendation library based on a graph structure through a Neo4jframework according to the record data of the official document and official document attributes; and receiving retrieval content input by the user from the official document recommendation library,and outputting a target official document according to the similarity sequence calculated by the SimRank. According to the official document recommendation method and device, the target official document with the highest relevancy with the retrieval content input by the user can be recommended to the user.

The invention provides a method and a device for checking a knowledge base triad. The method comprises the steps that M terms used for representing a first relation in a corpus are used as target feature terms, and first weight values of the target feature terms are acquired; according to the first weight values, the confidence of a to-be-checked triad in the first relation in a knowledge base isacquired; and whether the to-be-checked triad is credible is determined according to the confidence. According to the method, whether the to-be-checked triad is credible is determined by acquiring theconfidence of the to-be-checked triad, separate or batch checking can be realized, checking efficiency is improved, manual checking cost in practical application can be saved, and the efficiency of constructing a high-quality knowledge base is substantially improved; and moreover, it is accurate to check the credible degree of the triad through the confidence, universality is high when information checking is performed on different types of knowledge base triads, and the method can be applied to triad checking of any knowledge base.

The invention belongs to the technical field of visual information processing, and discloses a multipurpose visual communication design information processing system and method, and the system comprises a visual design material importing module, a parameter configuration module, a main control module, a drawing module, an editing module, a color correction module, a design evaluation module, a printing module, and a projection module. According to the invention, corresponding color correction and display are carried out on the original content through the color correction module, so that the visual comfort of a user is improved and the watching experience of the user is improved. Meanwhile, the evaluation module can be designed to evaluate the first expected feature words from the first expected feature words. A first preset number of first expected feature words is determined as the target feature words for evaluating the target visual design interface, the number of the feature wordsfor evaluating the visual design interface can be reduced, and the first expected feature words can be determined according to the judgment duration of the target user, so that the quantification ofthe evaluation result is realized, and the reliability of the evaluation result of the visual design interface is improved.

The invention discloses a keyword extraction method and device, equipment and a medium, and relates to the field of data processing. The method comprises the steps of obtaining a first comment text from a plurality of comment texts, wherein an emotion tag of the first comment text is a first emotion tag; performing word segmentation on the first comment text to obtain a feature word set of the first comment text; calculating an information entropy set of the first comment text according to the feature word set, wherein information entropies in the information entropy set are obtained through calculation according to feature words in the feature word set; and determining a keyword of the first comment text according to the information entropy set. According to the method and device, the information entropy set of the comment texts can be obtained, the keywords in the comment texts are determined according to the information entropy set, and the information entropy represents the difference between the comment texts, so that the keywords obtained through the information entropy have higher interpretability for the sentiment classification result, and the modeling effect and interpretability are improved.

An output apparatus includes a processor configured to receive an input word expressing a feature of a matter; and, by inputting the input word to a generation model trained on relation between a feature term extracted based on a descriptive text describing the matter and an associative text associated with the matter, the associative text being generated from the descriptive text describing the matter, output an associative text corresponding to the input word.

The invention relates to a dependency relationship recognition method and device based on a data table and computer equipment in the field of data analysis. The method comprises the steps of obtaininga dependency relationship identification task carrying a source table identifier and a target table identifier, and reading a source field name in a source table corresponding to the source table identifier according to the dependency relationship identification task; calling multiple threads to perform word segmentation on the source field name in parallel to obtain feature words in the source field name; obtaining description words corresponding to the feature words from a preset word bank; generating a reference field name by utilizing the feature words and the description words; calculating a first similarity between the source field name and the reference field name, marking the reference field name corresponding to the first similarity meeting a preset condition as an intermediate field name, and recording the intermediate field name; and searching a target field name corresponding to the intermediate field name from the target table, and when the target table comprises the target field name, determining a dependency relationship between the source table and the target table. By adopting the method, the identification accuracy of the dependency relationship between the datatables can be improved.

The invention discloses a text retrieval method based on mixed expansion of front and back parts of matrix weighted association rules, First, the user queries and retrieves the document set to construct the related document set of the first-checked user. Then, the weighted value and frequency of the item set are fused with the total weighted value of the feature words and the total number of documents of the first-checked user. The frequent item set containing the original query term is mined. The candidate item set is pruned by item weight sorting, and by use of a confidence-correlation evaluation frame is used to excavate association rules from frequent item sets. Finally, the consequent association rules of the original query term and the consequent association rules of the original query term are used as extension words, and the extension words are combined with the original query term to retrieve the document set again to obtain the final retrieval result document and return it tothe user. The invention adopts the pruning method based on item weight sorting, improves the mining efficiency, adopts the weighted association rule before and after parts mixed expansion technology,and improves the text information retrieval performance.

The invention relates to the technical field of data analysis, and discloses a task allocation method and device, a storage medium and computer equipment, and the method comprises the steps: obtaining user information of all users and task information of a to-be-processed task; extracting feature words from the user information of each user, and extracting keywords from the task information; converting the feature word of each user into a first feature vector by using a pre-trained word vector model, and converting the keyword into a second feature vector; respectively calculating a cosine distance between the first feature vector and the second feature vector of each user, and screening out a plurality of target feature vectors of which the cosine distances are greater than a preset similarity threshold; inquiring first users corresponding to the target feature vectors respectively, screening out a second user from the multiple first users according to a preset screening rule, and distributing the to-be-processed task to the second user. According to the method and the device, the second user matched with the task is screened out from a large number of users, and the task allocation accuracy is improved.

The invention provides an information association method and device, electronic equipment and a storage medium, and relates to the field of artificial intelligence such as natural language processing, deep learning and big data processing, and the method can comprise the steps: building corresponding test question banks for different courses; respectively determining feature word sets corresponding to different courses according to the corresponding test question banks; obtaining a to-be-associated question, and extracting feature words from the question; and determining the course corresponding to the question according to the extracted feature words and the feature word set corresponding to each course. According to the scheme, manpower and time cost can be saved.

The invention provides a text abstract intelligent extraction method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a plurality of feature statements from a plurality of texts, dividing feature words for each feature statement, and obtaining a plurality of feature words; classifying the plurality of feature words into different class clusters through clustering analysis; classifying the feature statement to which each feature word belongs into a corresponding class cluster; and extracting a fixed number of feature statements from each class cluster to form an overall abstract of the plurality of texts, wherein the clustering analysis process comprises the following steps: respectively carrying out word vector representation on the plurality of feature words to obtain a plurality of feature vectors; weighting each feature vector according to the importance degree to obtain a plurality of weighted vectors; calculating the similarity between every two weighting vectors; and performing clustering operation according to the similarity to obtain the number of clustering centers, and dividing the plurality of feature words into a plurality of class clusters according to the number of clustering centers.

The present invention relates to a system and method for information process using artificially constructed apparatus. More specially, the present invention provides a system and method that can search for information in a document structure and provide precise results by analyzing the inputs and search results using the executing system and the knowledge structure of the think system. In one preferred embodiment of the present invention, the search terms are divided into subject terms and corresponding feature terms, and document entry files comprising respective subject terms and corresponding feature terms will provide access to documents including subject terms and corresponding feature terms.

The embodiment of the invention belongs to the field of artificial intelligence and digital medical treatment, is applied to the field of intelligent medical treatment, and relates to an auxiliary diagnosis method, device based on an inquiry session, computer equipment and a storage medium. The method comprises the steps that a conversation text generated in the inquiry process is acquired, and the conversation text comprises an inquiry session between a doctor and a patient; performing feature word extraction on the inquiry conversation through a trained first feature extraction model to obtain a target feature word in each inquiry conversation; performing feature statement extraction on the conversation text through a trained second feature extraction model to obtain a target feature statement in the inquiry process; and performing differential recognition on the target feature word and the target feature statement in the conversation text, and displaying the recognized information as auxiliary diagnosis information in the inquiry process. In addition, the invention also relates to a block chain technology, and the conversation text can be stored in a block chain. Through the auxiliary diagnosis information, the misdiagnosis rate of doctors can be reduced.

The embodiment of the invention discloses a logistics object information processing method and device and a computer system. The method comprises the following steps: determining the text descriptioninformation of a to-be-classified target logistics object, processing the text description information, and determining a contained target feature word; generating a feature word vector correspondingto the target logistics object according to the inclusion condition of the text description information for each target feature word; and inputting the feature word vector into a coding classificationmodel to obtain corresponding classification feature information. Through the embodiment of the invention, automatic classification of logistics object codes can be realized, and the error probability is reduced while the labor cost is reduced.

The invention discloses an object classification method and a classification model construction method and device, and relates to the technical field of computers. One specific embodiment of the object classification method comprises the steps of obtaining initial feature data of a to-be-classified object, wherein the initial feature data comprises identification information data and attribute information data of the to-be-classified object; carrying out word segmentation on the identification information data and the attribute information data to obtain a feature word set, and the feature word set comprising at least one feature word; and performing vector representation on the feature words in the feature word set, and determining a target category to which the to-be-classified object belongs based on a trained classification model. According to the object classification method, the word vectors can be input into the trained classification model according to the feature word set of the initial feature data and the vector representation of the feature words, so that the target category to which the to-be-classified object belongs can be automatically, quickly and accurately determined.

The invention discloses a text generation method and device, and relates to the technical field of computers. A specific embodiment of the method comprises the steps of determining a feature word set corresponding to a product attribute and a comment word set corresponding to a user comment according to a reference abstract text, and training and determining a target text extraction model based on the feature word set, the comment word set and a general word set; forming the abstract text containing the recommendation words automatically by utilizing the target text extraction model according to the detailed information of the target product and the user evaluation, and recommending the product by utilizing the generated abstract text, so that the recommendation accuracy and the marketing effect of the recommended target product are improved.

The invention discloses a text similarity calculation method based on x2-C, and particularly relates to the field of text information processing. According to the method, a convolutional neural network CNN is used to classify a test data set; calculating an initial weight of each feature word in the detection sample according to the TF-IDF; calculating a domain correlation factor by using an x < 2>-C algorithm, calculating an initial weight by using the word position factor alpha in combination with the domain correlation factor to obtain a feature word weight, establishing a word bank by using all feature words of the detection sample, and expressing the detection sample as an initial text vector in combination with the word bank and the feature word weight; utilizing a word2vec tool tocalculate the similarity degree among the words in the word bank and form a word meaning similarity degree matrix; the initial text vector is calculated by using the matrix to obtain the text vector,and finally the text vector is calculated by using a cosine similarity algorithm to obtain the similarity between the texts, so that the association degree between the feature words and the field of the feature words, the semantic relationship between the feature words and the position information of the feature words are increased, and the accuracy of text similarity calculation is improved.

According to a feature selection method based on a covariance measurement factor, on the basis of an original triangular comparison measurement algorithm (TCM), the concept of the covariance measurement factor is introduced, and the correlation between features and categories is further measured on the document frequency level by calculating covariance values of feature words and the categories. When the performance of the method is verified, a naive Bayes algorithm is used for classification operation, and a macro F1 and a micro F1 are used for evaluating the classification effect. According to the method, feature words highly related to the categories can be better screened out, the method is a reliable feature selection algorithm, and the classification accuracy and efficiency are improved.

The invention discloses a classification method of information flow material creative text, includes extracting frequent feature word sets and their associated training text based on association analysis and using the association analysis results, The present invention greatly reduces the time complexity and improves the accuracy of classification, improves the determination of the number of the nearest neighbors, and greatly reduces the time complexity of classification.

The invention discloses a text matching method and device, computer equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: constructing a question feature word set QU, a text feature word set QC and a term set T; performing vectorization processing to obtain a feature vector QE and a term vector TE; performing linear conversion to obtain a key matrix K, a query matrix Q, a value matrix V and a matrix KT; calculating a non-normalized weight matrix AQT, and then performing self-multiplication and normalization processingto obtain a plurality of sub-matrices; and performing equalization processing on the plurality of sub-matrices, performing normalization processing to obtain an influence matrix, performing matrix multiplication on the key matrix K and the query matrix Q to obtain a self-attention matrix A, performing calculation to obtain output of a self-attention module, and performing matching according to the output. According to the method, the matching between known terms is emphasized, the matching between non-terms is reduced, and the effect of improving the matching accuracy is achieved.