467results about How to "Search results are accurate" patented technology

A search engine utilizes a bottom-up approach to index the content of a network with agent programs running on each host computer instead of relying on a top-down approach as used by conventional search engines. A central catalog of object references is compiled on the central site or sites from the meta data reported from each web site. One or more brochure files may also be created and stored on each web site to provide conceptual or non-key-word data about the site, such as target demographics and categorization information. This conceptual information is then utilized in constructing the central catalog so that more accurate search results may be generated for search queries applied to the catalog.

The present invention generally relates to the methods and systems for developing, specifying, and assigning descriptive information relating to the contents of a file (i.e., metadata). User interface controls on a computer screen implement a dynamically changing display which responds to user input by presenting new categories of choices. Additional controls allow optimization of the process of specifying and assigning descriptive metadata.

A computer-implemented method is provided for ranking files from an Internet search. In one embodiment, the method comprises assigning a score to each file based on at least one of the following factors: recency, editorial popularity, clickthru popularity, favorites metadata, or favorites collaborative filtering. The files may be organized based on the assigned scores to provide users with more accurate search results.

An implicit links enhancement system and method for search engines that generates implicit links obtained from mining user access logs to facilitate enhanced local searching of web sites and intranets. The implicit links search enhancement system and method includes extracting implicit links by mining users' access patterns and then using a modified link analysis algorithm to re-rank search results obtained from traditional search engines. More specifically, the implicit links search enhancement method includes extracting implicit links from a user access log, generating an implicit links graph from the extracted implicit links, and computing page rankings using the implicit links graph. The implicit links are extracted from the log using a two-item sequential pattern mining technique. Search results obtained from a search engine are re-ranked based on an implicit links analysis performed using an updated implicit links graph, a modified re-ranking formula, and at least one re-ranking technique.

The invention discloses a multi-task layered image retrieval method based on a depth self-coding convolution neural network. The method is characterized by mainly comprising a multi-task end-to-end convolution neural network for deep learning and training recognition, a rapid visual segmentation detection and positioning method of a region-of-interest secondary screening module based on an RPN network, a coarse search of a full-graph sparse hash code, an area sensing semantic feature and matrix h accurate comparison and search based on the maximum response, and a region-of-interest selectivitycomparison algorithm. According to the method, the end-to-end training can be achieved, the interest region with higher quality can be automatically selected, the automation degree and the intelligent level of search by images can be effectively improved, and the image retrieval requirements of the big data age can be met by using little storage space at a high search speed.

An implicit links enhancement system and method for search engines that generates implicit links obtained from mining user access logs to facilitate enhanced local searching of web sites and intranets. The implicit links search enhancement system and method includes extracting implicit links by mining users' access patterns and then using a modified link analysis algorithm to re-rank search results obtained from traditional search engines. More specifically, the implicit links search enhancement method includes extracting implicit links from a user access log, generating an implicit links graph from the extracted implicit links, and computing page rankings using the implicit links graph. The implicit links are extracted from the log using a two-item sequential pattern mining technique. Search results obtained from a search engine are re-ranked based on an implicit links analysis performed using an updated implicit links graph, a modified re-ranking formula, and at least one re-ranking technique.

Disclosed is a method for characterizing the content of a digital image. A component of the method is the homogeneity measure, which calculates a value or values based on the local similarity of portions of a sample area. A further component is a means of optimally adjusting the shape and offset of the local sample area to improve the resulting measure. The homogeneity measure may be applied at a higher level to segment an image, by grouping together larger portions of an image whose sub-components are locally similar. A strongly homogeneous region is a contiguous region composed of elements exhibiting local homogeneity better than that of surrounding elements, and may be applied to identify the scale and other characteristics of a portion of an image. This may be used to facilitate algorithms that separate lighting and texturing information, and in identifying the composition of an image.

The invention provides a query method and system and relates to the field of computer technology. The method comprises receiving a query word string and analyzing the speeches that the words in the query word string belong to, wherein the speeches comprises entity words, property words, property values corresponding to the property words, relative words, and relative values corresponding to the relative words; according to the speeches that the words in the query word string belong to, determining a query structure correspondingly matched with the query word string and query objects in the query structure, match-analyzing the query structure in a structuralized knowledge library to obtain object values corresponding to the query objects in the query structure; separately displaying the object values in a result area corresponding to the query word string. The query method and system can help a user to query required answers rapidly to achieve high searching efficiency; by avoiding the shortcoming of immature semantic parsing in the prior art, can precisely perform question query and answering; can also effectively answer uncommon questions.

The invention provides a search method and a search device. The search method comprises the following steps of: receiving query information, wherein the query information comprises at least one query keyword; determining first expansion information, which is associated with the query information, in a preset knowledge base, wherein the preset knowledge base comprises different pieces of expansion information; according to the query information and the first expansion information, calling a third-party search engine for searching; acquiring and outputting a search result; and determining first matching information, which is associated with the query information and the first expansion information, in the search result, and adding the first matching information which serves as the expansion information into the preset knowledge base. By the embodiment of the invention, search efficiency can be improved, and the search result is more accurate.

The invention provides a voice search method, a voice search device and a voice search system. The voice search method comprises the following steps: receiving a voice search request from a user; performing intent resolution on the voice search request so as to acquire the search intent of the user; acquiring a search result according to the search intent of the user; generating a voice dialogue result of the voice search request according to the search result and the search intent of the user and supplying the voice dialogue result to the user. According to the voice search method provided by the invention, simple voice broadcast for the search result is not performed, but a dialogue form corresponds to the input voice search request, so that the device and the system are more intelligent. Furthermore, the search intent of the user can be analyzed, the corresponding search result is acquired so as to generate the voice dialogue result, user requirements can be met more accurately, and the result is more accurate, individualized and humanized.

The invention relates to a data searching method and system based on semantic analysis, and belongs to the technical field of information searching. The method comprises the steps that keywords input by a user are divided into words with different parts of speech through a word dividing library, for results after the words are divided, geographical positions to be searched can be determined according to a geographical noun library if the parts of the speech are geographical names, otherwise matching is carried out by entering a phrase word library, query is further carried out based on keywords corresponding to the matched record, and words input by the user are regarded as the keywords of the inquire if matching is failed. The data searching method and system based on the semantic analysis carries out the semantic analysis based on a large number of behaviors of the user, can carry out good analysis on non-concrete nouns, verb-object phrases, modification phases and other vocabularies with generalization properties, can automatically select new words needing to be put into a library at the same time, is few in searching repeated process, high in searching efficiency and relatively precise in researching result and can meet demands of clients.

A system and method of facilitating search by color processes an image of an item to obtain its color space parameter values and / or color names based on the color space parameter values, and adds the color space parameter values and / or color names as color attribute of the item. When a user inputs a color as a search criterion, the method searches the color attributes of items, and returns a list of items matching the color selected by the user.

The invention relates to correlation techniques of search algorithm, in particular to a universal intelligent search method capable of understanding natural semantics and being widely applied in different fields such as shopping, music and books. The method mainly solves the problems that the present key word search algorithm is only that key word segments are matched with target contents based on an index, and the matching result is in a mass when semantic keywords are connected, and aims to realize the technical purposes as follows: a natural semantic search sentence can be initiatively analyzed and understood, is divided into semantic bodies rather than key words; the index matching of key words is replaced by the matching of bodies and sentence types, and finally the effect of understanding natural semantics is realized; and further, the method provides a more convenient and accurate search method for natural language search, in particular to voice interaction search.

The present invention discloses a search system based on web page extraction and comprises a web page download unit for downloading web pages and a result storage unit for storing search results; the invention also comprises a template storage unit and a web page extraction unit, wherein, the template storage unit is used for storing one or more than one templates which are used for recording the properties of preset web pages; the web page extraction unit uses the content of the web page which is downloaded by the web page download unit and matched with the template as a search result. The invention also discloses a corresponding search method based on the web page extraction. Through matching the properties of the downloaded web page and the preset web page, thereby the invention achieves more accurate search results.

The invention discloses a situational searching method. The situational searching method comprises the following steps of: acquiring a query string which is required to be input by a user from the current operation interface of the user, and pre-processing the obtained query string; according to the contents of the current operation interface of the user, performing contextual analysis and situational calculation on the pre-processed query string; and according to a contextual analysis result and a situational calculation result, recognizing the retrieval intention of the user, and according to the recognized retrieval intention of the user, performing situational searching, and returning a retrieval result. Meanwhile, the invention also discloses a situational searching system. By the invention, the retrieval intention of the user can be completely understood, and seamless connection between the current operation and searching of the user is realized.

Methods, systems and apparatus, including computer program products, for expanding search queries. One method includes receiving a search query, selecting a synonym of a term in the search query based on a context of occurrence of the term in the received search query, the synonym having been derived from statistical machine translation of the term, and expanding the received search query with the synonym and using the expanded search query to search a collection of documents. Alternatively, another method includes receiving a request to search a corpus of documents, the request specifying a search query, using statistical machine translation to translate the specified search query into an expanded search query, the specified search query and the expanded search query being in the same natural language, and in response to the request, using the expanded search query to search a collection of documents.

The invention discloses a hierarchical structure applied to a distributed search engine. The hierarchical structure comprises a physical layer, an abstract layer, an application layer and a presentation layer. The invention also discloses a distributed search engine system, which comprises a Web server, an agent node, a query agent pool, an abstract adapter and a plurality of working nodes, wherein the query agent pool consists of a plurality of query nodes. The invention also discloses an implement method for the distributed search engine system, which comprises the following steps of: S1, registration of a query node; S2, registration of working nodes; S3, state update of the nodes; and S4, distribution and retrieval of a query request. The hierarchical structure has the advantages of good performance, high reliability, diversification, specialization, strong applicability and the like.

A video search system based on image recognition and matching comprises a server side and a user side. The server side comprises a video collecting module, a video character feature collecting module, a video character information analysis module, a video preprocessing module, a video content analysis module, a video feature extraction module, a server database module, an image feature extraction module, a server control module and a user feedback information analysis and process module. The user side comprises a user search information uploading module, a video search result displaying module and a search result user feedback module. The invention provides a method for searching for videos through images. Good user experience is achieved; by uploading one image, a user can obtain video resource information related to the content of the image; the returned search result is accurate, the information amount is large, and mass video resources on the Internet can be fully utilized.

The invention relates to the artificial intelligence technology field and particularly relates to an image text matching model training method, a bidirectional search method and a related device. Thetraining method comprises steps that global and local representations of an image sample and a text sample are extracted, a pre-constructed matching model is trained, the global and local representations of the image sample and the text sample are respectively mapped to the specified semantic space by the matching model, and similarity of the global representation and similarity of the local representation are calculated; according to the default weight of the similarity of the global representation and the default weight of the similarity of the local representation, the image and text matching degree is determined through employing a weighted summation mode. The training method is advantaged in that the acquired matching degree is based on considering both detail characteristics of an image and global characteristics, and the acquired matching degree is more accurate and comprehensive.

The embodiment of the invention discloses a method for generating an application type label, wherein the method includes steps of acquiring a first class information of a first application data source, acquiring a second class information of a second application data source; integrating the first class information and the second class information to be a third class information; acquiring application information of the first application data source or the second application data source; matching the application description information with the class label in the third class information, and marking the class label in the third class information for the sample application mark; according to the application information, extracting corresponding feature vectors corresponding to the sample application label, and applying the class label corresponding to the label and the corresponding feature vector training classifier according to the sample; marking the class label in the third class information for the target application mark according to the training classifier. The method is the class label generated by the target application, thus the result is more accurate in searching for the application.

The invention discloses a training data generation method and device and a searching method and device based on a neural network model. The training data generation method comprises the following steps: obtaining history search data of a user, wherein the history search data comprises history search words and history search results corresponding to the history search words; obtaining history search and click behaviors of the user, and carrying out classification on the history search results corresponding to the history search words according to the history search and click behaviors to generate labels of the history search results; and according to a preset strategy, the labels of the history search results, the history search words and the history search results, generating training data. The method does not need participation of workers, thereby realizing an automation function; and the training data is generated according to the labels of the history search results, and the training data can be more suitable for the neural network model, so that searching can be carried out through the neural network model to obtain a more accurate search result, and search accuracy is improved.

The invention discloses a multi-feature fusion human face image searching method. The method comprises the following steps of: acquiring an image and text information surrounding the image from the Internet, storing the image and the text information respectively into an original image library and a text library; filtering non-human face images and irregular human face images from the original image library, saving regular human face images and storing the human face images into an human face image library; extracting the bottom layer features of the human face images in the human face image library; reducing the dimensions of the bottom layer features and storing the bottom layer features into a feature library; establishing feature indexes for the bottom layer features and storing the feature indexes into a feature index file; establishing text indexes for the text information in the text library and storing the text indexes into a text index file; receiving a text and an image which are submitted by a user; and indexing the text from the text index file, searching the image from the feature index file according to the searching result of the text, and feeding back the searching result of the image to the user. The method has the characteristics of high searching precision, high reliability and high efficiency.

The invention discloses a system and a method for parallel video copy detection. The method includes the steps of 1, selecting key-frames of a query video and a reference video by a parallel method and extracting MIFT features of the key-frames; 2, clustering extracted feature data of the reference video by a parallel hierarchical clustering method; 3, quantizing the features of the query video and the reference video by a quantitative method according to clustered results; 4, establishing indexes of quantized data of the reference video; and 5, retrieving by the parallel method, preliminarily searching in the indexes to obtain an alternative video by utilizing quantized data of the query video and then computing space consistency and time consistency to finally confirm a copy video. According to the system for parallel video copy detection, a parallel mechanism is adopted on the basis of fast retrieval of visual vocabulary bag model BOF (beginning of file), and accordingly detection efficiency of the system is improved greatly.

The invention discloses a cloud search platform. The cloud search platform comprises a user authority system and a second presentation layer. The cloud search platform is characterized by further comprising a basic data layer, a data persistence layer, a JAVA application support layer and a first persistence layer which sequentially perform bidirectional data communication, wherein the user authority system performs bidirectional data communication with the basic data layer, the data persistence layer, the JAVA application support layer and the first persistence layer respectively; and the second persistence layer performs bidirectional data communication with the basic data layer. The cloud search platform provided by the invention has the characteristics of all-directional search, high reliability, high memory space, low delay and the like; and a cloud search platform which is more advanced, easy to use, and fast in response and analysis speeds is provided for enterprises or governments.

The invention discloses a method and a device for mobile search. The method comprises the following steps of: recovering a search request containing one or more query key words; calculating the score of each search type domain, wherein the score is the score of any of the following items or the comprehensive score of the items: the similarity between the search request and the search type domain, a public search rate, corresponding to the search type domain, of the search request and an individual user interest score of the search type domain, and the public search rate is the number of public searches or the times of the clicking of a public search result; and according to the scores of the various search type domains, selecting one or more search type domains to be searched for the query key words. By using the method and the device, individual and accurate search results can be provided for a user.