35 results about "Topic sentence" patented technology

The invention discloses a network public feelings collecting method; hot words are extracted from a network forum, related information documents are extracted according to the hot words to form a hot event document set which is related to the hot words; the key sentences in the hot event document set are clustered to obtain a plurality of view topic sentence sets on one hot event; then a feelings word stock with different types is established, feelings marking is carried out on the feelings words of the key sentence in the view topic sentence set; the feelings tendency value of each key sentence is calculated according to the feelings marking; the feelings tendency value of each netizen on the view topic is calculated according to netizen ID mark of the key sentence to obtain the network public feelings about some hot event, thus achieving the collection of the network public feelings. The invention not only can effectively find the public feelings information in the network, but also can timely and accurately extract netizen view points aiming at concrete public feelings, and makes up the defect that the previous public feelings system can only find the public feelings information, but the view can not be analyzed.

The invention discloses an automatic extraction method of a conversation text topic, comprising preprocessing conversation text data and detecting question-answer pairs in the preprocessed conversation text; segmenting topics of the conversation text, clustering language chunk groups after topic segmentation and extracting a topic sentence from the clustered language chunk groups. The conversation text topic is more accurate by adopting the extraction method, enabling a user to search or retrieve conversation records of interest from the extracted topic sentence, thereby improving the user experience.

The invention provides an aspect displaying method and device. The method includes analyzing aspect sentences from review data, and determining emotional tendency and review dimensionality of each aspect sentence; calculating review similarity of the aspect sentences; according to the review similarity, clustering aspects to obtain clusters of similar aspect sentences; extracting topic sentences from the clusters of similar aspect sentences, screening the topic sentences to obtain topic sentences rich in lexical expression and having typical points, and displaying the topic sentences as sentence-level aspects. The aspect displaying method and device has the advantages that single aspects are embodied through sentence-level review tags overall aspects are embodied through abstracts generated from the review tags, the displaying mode is richer and more flexible, user reading expectations are better met, demand satisfaction is evident, universality is good, extension to other types of review data can be achieved at low labor cost, and data output efficiency is improved.

The invention provides a domain-oriented Chinese text topic sentence generation method. The method is characterized by comprising the following steps of establishing a corresponding domain knowledge map for a domain-oriented text data set, using a deep neural network model for extracting semantic information from texts, classifying the texts according to topic sentence patterns, and finally generating topic sentences of the texts. A data set conceptual model and content narrative mode characteristics are obtained by creating the domain knowledge map, and a deep learning model is used for conducting labeling and classifying training on text data, so that the topic sentences of the texts are generated, and knowledge-based query and statistics are achieved. The method has high application applicability and a good topic sentence generation effect on the limited domain data set.

In various embodiments, a method and an apparatus for extracting topic sentences of webpages are provided. The method comprises: obtaining candidate webpages, and a pre-built machine learning model, each candidate webpage contains multiple preselected candidate topic sentences, and each candidate topic sentence includes several word segments; determining word feature values that indicate importance levels of the word segments in each candidate webpage respectively, and inputting the word feature values to the machine learning model to obtain an importance value for each word segment; for each candidate webpage, determining a partial order value for each candidate topic sentence according to the importance values of the word segments included in the candidate topic sentence; and for each candidate webpage, selecting one of the plurality of candidate topic sentences that is associated with a partial order value larger than a preset threshold value as a target topic sentence of the candidate webpage.

The embodiment of the invention discloses a method and a device for determining the similarity between an inquiry sentence and a webpage, a terminal and a server. The method comprises the following steps: translating a target inquiry sentence into at least one candidate sentence having similar semantics through a pre-created phrase translation model; determining the similarity between the target inquiry sentence and a webpage topic sentence according to the similarity between the at least one candidate sentence and the webpage topic sentence, wherein the webpage topic sentence is a webpage title or a sentence for describing major webpage content obtained by resolving webpage content based on a set algorithm. By adopting the technical scheme provided by the embodiment, the webpage recall rate of any inquiry sentence by a search engine can be increased, the search engine can return webpage sets with small differences specific to inquiry sentences having different representation forms but similar semantics, and the user satisfaction of an inquiry result is improved.

Methods for generating jokes include coupling server(s) with database(s) having words stored therein; receiving text at the server(s) from an external source communicatively coupled with the server(s) through a telecommunications network; in response to receiving the text at the server(s): selecting one or more topic keywords of the topic sentence using the server(s); generating one or more punch words with the server(s) using words stored in the database related to the topic keyword(s); adding bridges to the punch word(s), using the server(s), to generate one or more jokes; communicating a signal to a first computing device through the telecommunications network using the server(s); and in response to receiving the signal at the first computing device, displaying or speaking the joke(s) using the first computing device. Systems for generating jokes include networked computer components configured to carry out the methods. The methods/systems may also be used for recognizing jokes.

An embodiment of the invention discloses a user query statement syntactic structure determining method and device. The method includes identifying the alignment relationship between the segmentation of the user query statements and the segmentation of preset webpage topic sentences; according to the alignment relationship and the syntactic structures of the webpage topic sentences, establishing the syntactic structures of the user query statements. According to the technical scheme, the syntactic structures of the user query statements can be captured in a manner of facilitating the subsequent processing, especially facilitating the subsequent measuring of correlation processing between the user query statements and webpage topic sentences to be matched.

Embodiments of the invention provide a text processing method, electronic equipment and a storage medium. The method comprises the steps of obtaining a to-be-processed text and a label corresponding to the text; determining a plurality of candidate event sentences from event sentences contained in the text, the event sentences being sentences containing event trigger words and named entities; respectively calculating the similarity between the plurality of candidate event sentences and the label; and taking the candidate event sentences of which the similarity meets a preset similarity condition in the plurality of candidate event sentences as topic sentences corresponding to the text. In this way, the determined topic sentences are high in accuracy, and the text can be described simply and accurately.

The invention discloses an intelligent information extraction method and system. According to the method, after an NMF method is adopted to decompose a document feature matrix to obtain K clustering topics, clustering sentences in the document to obtain a set of a plurality of sentences; taking the first k topic sentence sets with the highest semantic similarity with the question text, using the question text again to respectively retrieve the k topic sentence sets, returning m most relevant sentences from each sentence set to form a corresponding document, and combining the obtained k documents into a long document; and performing answer extraction on the long document and the question text by adopting an MRC model combining a bidirectional attention flow model and a question text tensorinitialized Point Net model. According to the scheme provided by the invention, an improved MRC algorithm is adopted, and question text information is utilized in the stages of document theme aggregation and filtering, sentence retrieval recall and sorting and answer extraction; the OOV problem existing in the prior art is effectively solved, the data labeling cost is low, and the calculation efficiency and accuracy are both achieved.

The invention discloses a method and a system for a reading robot to perform automatic questioning and answering. The system comprises a question type analysis module, a word similarity matching module, a paragraph topic sentence and author viewpoint sentence extraction module and an answer sentence extraction module, and firstly proposes a method of analyzing question types according to general questioning and answering characteristics and then performing linear interpolation, thereby effectively filling in the blank of analyzing the question types only according to question sentences; according to the method, some answer candidate sentences which are not matched with question keywords and frameworks can be recalled by extracting paragraph theme sentences and author viewpoint sentences; according to the invention, intelligent answering for questioning and answering purposes in reading understanding can be realized, and the answering rate of automatic questioning and answering is effectively improved; the method provided by the invention can also be applied to natural language processing tasks such as text implication, automatic abstract, complex problem solving and the like.

The invention discloses a text classification method and device, and relates to the technical field of computers. A specific embodiment of the text classification method comprises the steps of extracting sentence vectors of a to-be-classified text by using a word vector model to obtain a sentence vector set of the to-be-classified text; extracting a topic sentence vector of the to-be-classified text from the sentence vector set of the to-be-classified text; utilizing a topic sentence vector of the to-be-classified text; using a classifier model to predict the category of the to-be-classified text, the classifier model comprising a plurality of clustering center vectors, and predicting the category corresponding to the clustering center vector having the smallest distance with the topic sentence vector of the to-be-classified text as the category of the to-be-classified text. According to the embodiment, new events can be automatically classified, and the labor cost is reduced, and theclassification accuracy can be improved, and the problem that the category number is difficult to determine is solved.

The invention discloses an international organization science and technology term topic sentence extraction method based on a multi-machine translation engine, and relates to the technical field of information science and knowledge engineering. According to the method, on the basis of a cross-language term knowledge base, by calling various machine translation engines and adopting a natural language chunk processing technology, the technical steps of automatic term sentence recognition, topic statement chunk analysis, translation connection and fusion and the like of the cross-language science and technology text are designed. Therefore, term knowledge point rapid identification, subject sentence automatic detection and high-quality translation fusion generation are carried out on international organization science and technology knowledge, so that subject sentence extraction accuracy and fluency are guaranteed, finally science and technology information issued by an international organization is dynamically monitored, and the knowledge processing requirement of a user for cross-language professional science and technology knowledge is met.

The instruction discloses a method and device for splitting video strips. A method for video stripping, comprising: extracting audio from a target video to be stripped, and converting the audio into corresponding text; determining several topic sentences from the text; using each topic sentence as For the initial sentence of the text paragraph, the text is divided into several text paragraphs; based on the text paragraphs obtained after the division, the target video is correspondingly split into several video segments.

The embodiment of the invention provides a data processing method and device, an abstract display method and device, equipment and a storage medium. The method comprises the steps of obtaining to-be-processed text data, wherein the to-be-processed text data comprises a plurality of sentences; inputting the plurality of sentences into a target abstract generation model, and determining at least one topic sentence in the plurality of sentences; obtaining an abstract of the to-be-processed text data according to the at least one topic sentence; wherein the input of the target abstract generation model is a plurality of sample sentences in the training text data and context sentence segments corresponding to the sample sentences respectively, and the output of the target abstract generation model is at least one topic sentence corresponding to the training text data. According to the embodiment of the invention, the abstract comprising the context information can be generated based on the target abstract generation model with the capability of integrating the context information, and the generation quality of the abstract is improved.

The invention discloses a method for extracting a text topic sentence. The method comprises the following steps: constructing a writing structure model according to a text body, and calculating the weight of each constituent part of the model and the topic value degree of a corresponding text block; constructing an iterative model and calculating the value degree, and dynamically adjusting the text processing granularity; using the sentence mutual information as a relative weighted value,creating the decision tree based on weight proof to extract the text topic sentences, the invention can be applied to extraction of the text topic sentences with complex styles, changeable writing structures and uncertain capacity, and the purposes of accuracy, easy expansion and wide range are achieved.

The invention relates to an industrial internet patent identification method based on multi-instance learning, which utilizes a natural language processing technology to segment summary information in a patent into sentences, and utilizes a text subject sentence extraction algorithm based on a sentence relation graph to extract subject sentences in a summary, so that the calculation overhead can be effectively reduced. Meanwhile, by combining the topic sentences extracted from the title and the abstract, the patents are converted into sentence packages, each patent is regarded as a package, and each sentence in the package is regarded as an example. And finally, the category of the new sample is predicted by adopting a K Neighbor Networks (KNN) algorithm (K Neighbor Networks, KNN). By means of the method, the industrial internet patent recognition effect can be effectively improved, the manual review cost is greatly reduced, and the method has very important significance on patent retrieval.

The invention provides a composition off-topic detection method and system. The method comprises the following steps: extracting problem key features from topic sentence codes by a neural network model; extracting key answer features from the composition codes, and inputting the key question features and the key answer features to an attention mechanism layer. It is ensured that the features input to the attention mechanism layer are key features, interference caused by other useless features to data processing of the attention mechanism layer is reduced, and the accuracy of data processing of the attention mechanism layer is improved. On the basis, the attention mechanism layer determines features of subject sentences concerned by the composition and features of the composition concerned by the subject sentences, the features are selected from two directions, the comprehensiveness of the selected features is ensured, the neural network model can perform classification based on more comprehensive features, the accuracy of the classification result is improved, and therefore, the detection accuracy is improved.

The invention relates to an automatic microblog home page data recommendation method. The method includes the following steps: (1) a microblog list is screened out from mass microblogs; (2) microblog topic sentences are extracted, and according to the sizes of pictures, microblog topic sentences with corresponding lengths are extracted; (3) chosen pictures, the sizes of which are closest to the sizes of target pictures, are automatically cut. The automatic microblog home page data recommendation method can automatically recommend latest and hottest blog article pictures and outlines to home pages to satisfy the demand of users. An automatic statistic screening method is adopted to fill home page data, consequently, the freshness, extensiveness and update cycle are increased, and manpower and cost are reduced. A manual test shows that the quality of picture screen and cutting reaches 99.9 percent and the accuracy of microblog outline recommendation reaches more than 98 percent.