233 results about "Similarity computing" patented technology

The invention discloses a text similarity computing method. The method comprises the following steps of text representation and text similarity computing. The aim of text representation is that a text document of product description is converted into a vector for description. In the text similarity computing method, natural language processing technologies such as Chinese words segmentation, stop word removing, word frequency statistics and the like are used for converting all the description texts of products into vectors; the text similarity is computed by a method based on a Hamming distance, and the other advantage of the Hamming distance is that the computing speed is very high. Due to the fact that the method of statistical machine learning is used, so that the text similarity computing method is more stable and effective compared with a method based on rules.

The invention relates to a method and a device for combining entities in a knowledge map. The method comprises the following steps of generating a first-grade feature vector according to structural data corresponding to the entities in the knowledge map; generating a second-grade feature vector according to terms in documents corresponding to the entities; determining the similarity of the different entities according to the first-grade feature vector and the second-grade feature vector. According to the method and the device, the first-grade feature vector and the second-grade feature vector are constructed by entity IDs respectively, and the similarity of the entity IDs with the same name is calculated, and whether the entity IDs with the same name are the same object or not can be accurately judged, so that a condition that one object in the knowledge map has a plurality of entity IDs can be reduced, the content of the knowledge map is more accurate and the structure is more compact.

Content recommendation apparatus and methods using a tag cloud provide a content recommendation service via a network. The apparatus includes a content tag cloud generating module configured to generate a content tag cloud by analyzing a tag assigned to each content and accumulating frequencies per tag of each content. The apparatus also includes a user tag cloud generating module configured to generate a user tag cloud by accumulating frequencies per tag of contents used by a user. The apparatus further includes a similarity computing module and a recommending module. The similarity computing module is configured to compute a similarity between users using the user tag cloud, and the recommending module is configured to recommend content by computing a probability that a target user will use a specific content based on the computed similarity between users.

The invention discloses a time window based LDA microblog topic trend detection method and apparatus. The method comprises: extracting a topic word from a word set by utilizing an LDA model in each time window, and obtaining global topics; performing similarity computing on the global topics, and performing K-means clustering to obtain hot topics conforming to public opinion analysis; extracting feature words of each hot topic in each time window in sequence in combination with the hot topic through the LDA topic model; and in combination with results of the feature words, computing a popular value of the hot topic in each time window, and drawing a trend graph of the hot topic. The apparatus comprises a first acquisition module, a second acquisition module, an extraction module and a drawing module. According to the detection method and apparatus, the precision of microblog topic detection is improved, so that a trend index is more expressive, and a more accurate basis is provided for analyzing a hot topic trend.

The invention discloses a real-time video monitoring method integrating edge computing with cloud computing. The real-time video monitoring method comprises the following steps: constructing a sampletraining library, training a CNN convolution neutral network by using image information in the sample training library, constructing a CNN model structure as training chip content, and embedding the training chip in front edge equipment; extracting a monitoring video of the edge equipment in real time, decomposing an original video into multiple frame images; performing feature and similarity computing on multiple frame images through the CNN training chip; storing the extracted similarity picture and information to remote cloud-side equipment to backup and manage for a long time, providing for the supervision department to use, analyze and compute, pushing the message to related department to implement the alarm to the related department. The invention further discloses a real-time videomonitoring system. The method disclosed by the invention has the advantages that the CNN image identification research technology is used for processing the picture in real time at the edge equipmentside, so that the mass storage and computing pressure of the cloud server is solved.

An object of the present invention is to provide a document processing device and document processing method that can provide a search result satisfactory to a user with respect to WWW documents in which a number of links among WWW documents is low and a number of accesses by users is low. An access pattern collection unit 101 generates an access user vector uj of one WWW document Dj and an access user vector uje of another document Dje. A user similarity computing unit 105 computes a document similarity sim (uj, uje) which indicates a user similarity between the WWW document Dj and WWW document Dje. A keyword vector smoothing unit 106 acquires a smoothed keyword weight vector w′j by correcting a keyword weight vector wj in one document, using the computed document similarity sim (uj, uje). An rearranging unit 110 calculates an evaluation value B_SCORE for input information for searching, based on the smoothed keyword weight vector w′j.

The invention discloses an indoor positioning method and system for a mobile robot based on a geomagnetic sequence. The indoor positioning method of the mobile robot based on the geomagnetic sequence includes a geomagnetic sequence fingerprint collection method of the mobile robot and an improved particle filter positioning algorithm; the system includes a mobile robot end and a cloud server end, and the mobile robot end is mainly responsible for the collection of the geomagnetic sequence fingerprint , motion control and collection of motion information, application to the server for positioning services and visualization of positioning information; the cloud server is mainly responsible for storing geomagnetic sequence fingerprints, processing geomagnetic sequence fingerprint data, calculating the similarity of geomagnetic sequences, and providing data for multiple mobile robots at the same time. Provide location services. The invention is less disturbed by the external environment, does not need other additional infrastructure, is simple to use and operates, and is suitable for the positioning of a mobile robot in an indoor one-dimensional passage corridor. The actual running effect proves that this method has good positioning accuracy for mobile robots.

The invention discloses a virtual medical expert based on a medical record database. The virtual medical expert comprises a medical data input device, a private statement similarity computing module, a diagnosis experience similarity computing module, a medical image similarity computing module, a comprehensive similarity computing module as well as a diagnosis and treatment scheme generation module, wherein the private statement similarity computing module, the diagnosis experience similarity computing module and the medical image similarity computing module are respectively used for computing indexes of the private statement similarity, examination result similarity and medical image similarity between a patient to be diagnosed and treated as well as diagnosed and treated patients recorded in the medical record database; the comprehensive similarity computing module is used for integrating the indexes of the private statement similarity, examination result similarity and medical image similarity with expert scoring-based clinical judgment importance indexes of the private statements of the patients, the examination results and the medical images, thereby computing the comprehensive similarity between the patient to be diagnosed and treated as well as the diagnosed and treated patients recorded in the medical record database; the diagnosis and treatment scheme generation module is used for selecting the highest-comprehensive-similarity medical record from all medical records of the diagnosed and treated patients in the medical record database, and outputting a corresponding diagnosis and treatment scheme. Through adoption of the virtual medical expert, the clinical values of electrical medical record information can be sufficiently exerted, and diagnosis and treatment suggestions can be offered to doctors or users.

The invention discloses a text similarity computing method for computing similarity between two texts. The data of at least two objects can be extracted from each text, and the object indicates the feature capable of reflecting the text semantic. The method comprises the following steps: determining shared objects of two texts, wherein the number of the shared object is at least two; computing theHamming distance of each shared object between two texts; determining the similarity between two texts according to at least one of the follows when the Hamming distance of each of at least two shared objects satisfies a first preset condition: word vector similarity of predetermined objects in at least two shared objects, the Hamming distance, and the splicing character string similarity. Through the above-mentioned scheme, the efficiency and the accuracy of the text similarity computation can be improved.

A face detecting system using a symmetric axis, the system including: an image inputting unit for inputting a still image thereto; an image converting unit for reducing the inputted image to a first certain size of image for a real-time process by using a pyramidal resolution reduction way; a scanning unit for scanning the reduced image by applying a second certain size of symmetric mask; a region dividing unit for dividing a region of the scanned mask into a left region and a right region with respect to the symmetric axis; a symmetric-similarity computing unit for computing a symmetric similarity of a local region (Symmetric_measure_local) at the two divided regions and then computing a symmetric similarity of a global region (Symmetric_measure_global) to compute a symmetric similarity of a face region; and a face region extracting unit for extracting a region having a symmetric axis at which the computed symmetric similarity is above a predetermined value, as the face region.

The invention discloses a friend recommendation method of a mobile social network. The friend recommendation method comprises the following steps that firstly, the dispersing geographic position similarity between any two users is calculated; secondly, the online time similarity of any two users is calculated; thirdly, the interest similarity of any two users is calculated; fourthly, the difference distance between any two users is calculated; fifthly, a target user is selected, and friend candidates are recommended for the target user according to the difference distance between the target user and other users. According to the friend recommendation method, the difference distance between the users of the mobile social network is calculated from a plurality of dimensionalities according to the dispersing geographic position similarity, the online time similarity and the interest similarity, the same groups are divided according to the difference distance, and the other users with the high similarity with the target user in space, time and interest can be accurately recommended for the target user to become the friend candidates of the target user; the friend recommendation method can be applied to a friend recommendation system, and individualized friend recommendation content can be pushed for the target user.

The invention relates to analysis of social network user recommendation and information service recommendation technologies, and more particularly to a geographical social network based user similarity computation method, belonging to the field of public opinion monitoring. The method mainly comprises following steps of space time semantic meaning extracting, user space time access model establishing and user similarity computing. The method has the beneficial effects that space time semantic meaning division is more comprehensive and the problem caused by data sparsity is solved; due to consideration of impact of both location function semantic meaning and geographical factor on user similarity, user portrait modeling is more comprehensive; and social network user similarity computation is carried out by combination of physical space time vicinity of traces and semantic meaning space time similarity, so that accuracy is higher, the social network user groups are divided, and the use classification and interest type determination effect are better.

The invention discloses a short text classification method based on semantic graphs, which is characterized by including the steps: A, constructing a semantic graph model for each piece of text information, and combining all semantic graph models; B, comparing similarity level among different texts according to the semantic graph models and by means of a similarity computing method; and C, according to the text similarity level, classifying the texts by the aid of a text semantic graph classifier. The short text classification method based on the semantic graphs has the advantages that semantic connotations of documents can be highlighted to a maximum degree by using the graph models to represent the texts, latent semantic information and theme features in the texts can be accurately described to a great extent by the aid of the TSG (text semantic graph) models constructed by the method, and the TSG classification method can be more reliable and efficient in use as compared with other classification methods by means of the feature, so that human cost is greatly reduced, artificial arrangement of the text information is avoided to a great extent, and the text information is automatically organized by a computer.

To provide a document extracting device, a document extracting program, and a document extracting method, having a low cost and a small amount of computation required to extract documents, a document extracting device includes similarity computing device computing all degrees of similarity between a plurality of documents to be candidates for extraction, and document extracting device to extract a combination of documents whose sum of the degrees of similarity between the documents computed by the similarity computing device is the smallest when any number of documents are extracted from among a group of the documents. As a result, since the cost required for works of giving keywords to the respective documents is not required to extract the documents, and even when the number of documents is increased, the amount of computation required to extract the documents is not increased extremely.

The present invention discloses a video summarization system and the method thereof. A similarity computing apparatus computes the similarity between each frame to obtain multiple similarity values. A key frame extracting apparatus chooses the key frames from the frames wherein the sum of the similarity values between the key frames is a minimum. A feature space mapping apparatus converts the sentences into multiple corresponding sentence vectors and computes the distance between each sentence vector to obtain multiple distance values. A clustering apparatus divides the sentences into multiple clusters according to the distance values and the importance of the sentences, and also applies a splitting step to split the cluster with the highest importance into multiple new clusters. A key sentence extracting apparatus chooses multiple key sentence from the clusters, wherein the sum of the importance of the key sentences is the maximum.

To provide a document extracting device, a document extracting program, and a document extracting method, having a low cost and a small amount of computation required to extract documents, a document extracting device includes similarity computing device computing all degrees of similarity between a plurality of documents to be candidates for extraction, and document extracting device to extract a combination of documents whose sum of the degrees of similarity between the documents computed by the similarity computing device is the smallest when any number of documents are extracted from among a group of the documents. As a result, since the cost required for works of giving keywords to the respective documents is not required to extract the documents, and even when the number of documents is increased, the amount of computation required to extract the documents is not increased extremely.

The invention discloses a method and a system for predicting the association relationship between a disease and LncRNA. The method comprises the steps: the LncRNA-miRNA association relationship and the miRNA-disease association relationship are obtained from a known database, and a LncRNA-miRNA-disease interaction network is constructed according to the two relationship; a disease hyper-expressionprofile and a LncRNA hyper-expression profile are constructed based on the LncRNA-miRNA-disease interaction network; the prediction model of the disease and LncRNA association relationship is trainedaccording to the disease hyper-expression profile and the LncRNA hyper-expression profile by using LncRNA similarity computing and disease similarity computing based on the RBF neural network; and the LncRNA-disease association pairs of the candidate samples are predicted by using the prediction model. The most promising LncRNA disease association for further experimental verification is provided, the potential disease-related LncRNA can be effectively mined from the mass biological data, the cost and the expense of the biological experiment can be reduced and the research progress in the bioinformatics field can be accelerated.

The invention discloses an improved text similarity solving algorithm based on semantic analysis. The algorithm comprises the steps of performing word segmentation and stop word removing processing on two texts; computing weights of the words in the texts based on an improved information theory method; acquiring weights of positions and properties of the words according to the positions and the properties of the words; constructing a target function shown in the description of the extracted text words according to the abovementioned three factors; and at last respectively reducing dimensions of the two feature words according to a semantic similarity, thus acquiring two feature word vectors, and then solving the text similarity sim (W1, W2) between the texts (W1, W2) according to a Pearson correlation coefficient. Compared with traditional text similarity computing method, the algorithm provided by the invention has higher accuracy, wider application range and higher application value, can accurately compute contribution degrees of the different words to a text thought and solve the problems of polysemy and synonym, is more accordant with an empirical value, and meanwhile provides a good theoretical basis for subsequent text clustering.

The invention relates to a text similarity computing method and system based on an improved LDA topic model. The method comprises the following steps: acquiring a plurality of text sets in a WMF_LDA topic model; performing similarity computing on words in a preprocessed word set by virtue of a word2vec word vector model and generating a plurality of word similarity values; generating a field topicword set according to similarity between the words; obtaining a probability distribution of a document subjected to word semantic combination on different topics by virtue of the LDA topic model; anddetermining topic distribution similarity between any two texts, thus the text similarity is obtained. The method provided by the invention has the advantages that firstly screening is performed, sothat quantity of the words in the topic word set is reduced; and unified mapping is performed on synonyms and words in the same field, and then probability distributions of the texts are obtained through modeling, so that computation dimensionality is low in a process of computing the similarity of two texts, wasted space is reduced, and the problems that a word level is focused on excessively anddifferent types of texts can not be fully excavated and utilized are solved.

The invention provides a similarity computing method and a similarity computing device on the basis of semanteme. The similarity computing method includes acquiring sentences S1 and S2 to be compared; subjecting the sentences S1 and S2 to word segmentation; mapping obtained words with semantic mapping after word segmentation to be uniformized expression; and computing similarity Sim (S1 and S2) between the S1 and the S2 which are processed in the last step. By mapping the words with semantic mapping to be uniformized expression and integrating with computing of the similarity, similarity between the sentences are displayed semantically other than literally, and accuracy of similarity between computing sentences is improved.

A card issuing system that performs personal authentication efficiently and with high accuracy. The card issuing system includes: a personal card reading unit operable to read a face image and personal information from a personal card; an image capturing unit operable to take a face image of a user; a similarity computing unit operable to compute a degree of similarity between the face image read by the personal card reading unit and the face image taken by the image capturing unit; and a display unit operable to display the face image read by the personal card reading unit, the face image taken by the image capturing unit, and the degree of similarity computed by the similarity computing unit side-by-side.

The invention provides a phishing website detection method excavated based on a network group. The phishing website detection method includes constructing a webpage set related to a given iffy webpage by using a crawler, then obtaining an iffy website and a potential target website thereof, furthermore calculating website feature signatures, and judging whether the iffy website is a phishing website or not through similarity calculations of the website feature signatures, if yes, returning to the target website thereof. The purpose of the phishing website detection method is detecting whether the website which the iffy webpage is located in is the phishing website so as to find the target website of the phishing website when detection is finished. The process of phishing detection is the process of finding the target website, and due to the fact that the iffy website is only calculated with the related potential target website for similarity, comparisons of the iffy website with irrelevant legal websites in internet are reduced, phishing detection efficiency is improved, resources are saved effectively, and work efficiency is improved.

Embodiments of the invention provide a user behavior similarity computing method and device. The method comprises the steps of collecting a behavior characteristic value corresponding to each of a plurality of behavior characteristics of first class of users and second class of users; screening the plurality of behavior characteristics to obtain a target behavior characteristic set; building a first generalized linear model according to the target behavior characteristic set, computing a first maximum likelihood estimated value of the first generalized linear model by use of an optimization method, and obtaining an estimated parameter corresponding to the first maximum likelihood estimated value; and computing the behavior similarity of a to-be-tested user and the first class of users by use of the estimated parameter and the behavior characteristic value corresponding to each behavior characteristic in a target behavior characteristic set corresponding to the to-be-tested user. According to the method, the similarity of behaviors of different users is analyzed by full use of behavior characteristics of a lot of users, so that the use ratio of collected behavior characteristics of a lot of users is improved.

The invention relates to a video quality evaluation system comprising a video quality objective evaluation system and a video quality comprehensive evaluation system. The video quality objective evaluation system comprises a peak value signal-to-noise ratio computing unit, an edge structure similarity computing unit, a sharpness computing unit, a space-time threshold computing unit, and an objective evaluation data processing unit. The objective evaluation data processing unit comprises an original data evaluation matrix constructing unit, a standard matrix constructing unit, and a weight computing unit. The video quality comprehensive evaluation system comprises an subjective result receiving module, an objective result receiving module, a second original data evaluation matrix constructing unit, a second standard matrix constructing unit, and a second weight computing unit. The video quality evaluation system effectively improves the accuracy of video evaluation results and achieves high-efficient evaluation.

The invention discloses a method and a device for computing similarity of webpages. The method comprises the following steps of obtaining webpage information of two to-be-compared webpages; respectively extracting content information, structure information and picture information from the webpage information of the two webpages; respectively extracting content feature vectors from the content information of the two webpages and computing the similarity of the content feature vectors; respectively extracting structure feature vectors from the structure information of the two webpages and computing the similarity of the structure feature vectors; respectively extracting picture feature vectors from the picture information of the two webpages and computing the similarity of the picture feature vectors; and computing to obtain the final similarity of the two webpages according to the similarity of the content feature vectors, the similarity of the structure feature vectors and the similarity of the picture feature vectors. According to the method and the device for computing the similarity of the webpages, which are provided by the invention, the similarity computing effect is better.

The invention discloses a method and a device for visualizing text set similarity, which belong to the field of text information processing and information visualizing. The method comprises the following steps of: preprocessing text sets; calculating the text set similarity; and visualizing the text set similarity. The device comprises a preprocessing block, a text similarity computing module and a visualizing module. According to the method, a text set similarity computing model is built to measure the text set similarity, a similarity result is displayed in a graphical mode, so that a beneficial effect that users can directly and quickly understand the text set content and similarity information is realized.

The embodiment of the invention provides a semantic similarity computing method, a search result processing method and a search result processing device. The semantic similarity computing method comprises the following steps: obtaining a first text string and a second text string; carrying out word segmentation on the first text string and the second text string respectively to generate word segmentation results; generating and reserving a plurality of semantic layers from the word segmentation obtained by the first text string and the second text string according to the word segmentation results; carrying out dependency similarity computing on each semantic layer of the first text string with all semantic layers of the second text string respectively to obtain N*N dependency similarity values; computing the semantic similarity values of the first text string and the second text string according to the computed N*N dependency similarity values. Through the semantic similarity computing method, the search result processing method and the search result processing device provided by the embodiment of the invention, the accuracy of semantic similarity computing between the text strings can be improved.