46results about How to "Optimize search results" patented technology

The invention belongs to internet massage processing field, characterized in that: common inquiry set which is cared by user, according to one ore more search engine logs, is screened out firstly using the inquiry of corresponding information of user number; then calculating user touching ratio corresponding to user touching page of inquiry in common inquiry set, if multi-search engine logs is used, user touching ratio are united; user touching pages are effectively screened out based on user touching ratio, the ralated inquiry and corresponding result page address are stored related database; at last, the result from user information and the result by searching by search engine are united effectively when user puts in the demand of inquiry, and are reordered to return to user. The method has the advantage that computer is completed automatically and efficiently submits search engine performance in real time and objectively.

The invention relates to the field of Chinese information retrieval (IR), in particular to an intelligent retrieval method based on domain ontology (DO) and an intelligent retrieval system including the method. The system comprises an ontology inference module used for analyzing natural query language input by users, an index processing module used for creating an index database, a query processing module used for performing special query, and a result optimizing ordering module used for processing query results. In addition, the system also comprises a domain library, a data repository, and the index database, which are built based on a certain domain. The intelligent retrieval system and the intelligent retrieval method based on the domain ontology make full use of concepts in the domain library and interrelation thereof, can correctly understand the demand of users, optimize retrieval results, return professional domain information for the users more accurately and comprehensively, and significantly improve the property of the information retrieval in the field of professional technology.

The invention discloses methods and systems for recommending network information and creating a network source index. The method for recommending network information comprises the following steps: obtaining a resource access log when a user accesses network sources; splitting the network resources in the resource access log of the user into network resource access pairs, wherein each network resource access pair contains information of two network resources in the resource access log; collecting the network resource pairs of each user, and obtaining a two-dimensional relation of the network resources; and recommending related network source for the user according to the two-dimensional relation of the network sources when recommending is required for the user. According to the invention, the range of information obtained by the user can be expanded, the utilization factor of information can be improved, and the difficulty for the user to obtain useful information is reduced.

The invention discloses an image retrieval method. The method comprises the steps that an image input by a user is acquired, and the image is input into a trained similarity hash function network to generate a similarity hash code; the generated similarity hash code and each hash code in a hash code library of a to-be-retrieved image is subjected to bitwise exclusive OR operation to generate a plurality of Hamming distances, and multiple image samples corresponding to the Hamming distances less than or equal to a hash code with a set threshold serve as retrieval results. The image retrieval method can solve the technical problems of low retrieval accuracy, poor method generalization capability, excessive classification, poor adaptability to sample data and large time consumption of existing content-based image similarity hash.

A system and method for extracting the site information, that it to browse and filter the page data of WWW is composed of a finder for finding home page data and sending the resultant to page searching file, a data extractor for extracting the contents from said file to create an extracted file, and a memory device for storing the finding condition, filtering condition, and said page searching file. The finding data come from different sites can be displayed by said system.

The embodiment of the invention provides a semantic information retrieval system and method based on domain ontology, and relates to the field of intelligent information retrieval. The semantic information retrieval system and method based on the domain ontology can achieve individual recommendation of user retrieval requirements, optimize the retrieval result and improve the precision ratio. The method comprises the steps that the system carries out analysis processing on an initial inquiry request input by a user, and a final inquiry request is obtained; a network information file meeting the conditions is matched out from a resource base according to the final total inquiry request; analysis and semantic processing are carried out on the network information file with the help of a domain ontology base, and useless and unrelated information is filtered out; individual processing is carried out according to a user interest model, the information meeting the user individual requirements is found out, correlation sequencing is carried out on the retrieval results by combining semantic correlation degree and interest correlation degree, and the retrieval results subjected to sequencing processing are fed back to the user.

The invention discloses an information retrieval-oriented information map generation method and dynamic updating method. A document weight is calculated for a user retrieval result, and data visualization is performed according to the document weight so as to generate an information map; and a user retrieval demand is optimized through interactive operation, and the information map is dynamically adjusted and updated. The information map generation method specifically comprises the steps of establishing a target document set and constructing a hierarchical knowledge concept generation system; preprocessing the target document set; inputting a retrieval demand by a user, performing calculation to obtain a domain feature and a scale feature of the retrieval demand, generating an initial retrieval result and performing sorting; and generating the information map by the initial retrieval result according to the sorting through a data visualization method, and performing display. According to the information retrieval-oriented information map generation method and dynamic updating method, the user can retrieve the information and display the retrieval result more visually and flexibly, and the problems in information retrieval-oriented information map generation and dynamic updating are solved.

The invention discloses a device and a method for searching documents, aiming at the problem that the conventional topic model cannot automatically identify topics. The device of the invention comprises a topic identification module, a random walking module and a searching module. The method comprises the following steps: identifying the topics from a document set by using a probabilistic topic model, and acquiring topic distribution of the documents according to the identified topics; calculating a random walking order score of a topic level for each document; and calculating the relevant score of the document relevant to the searching keywords according to the searching keywords and the topics, and combining the random walking order importance score and the relevant score of the topic level to acquire a searching result.

The invention provides an interest point recommendation method and device, and relates to the technical field of electronic maps. The method includes the steps: acquiring search habit messages of users; determining click rate of each target interest point message corresponding to each search word to be processed according to the search habit information; updating sequences of the target interest point messages in search results of the corresponding search words to be processed according to the click rate of each target interest point message to obtain a collating sequence, and recommending the target interest point messages to the users according to the collating sequence. The search habit messages comprise search words of the users and the target interest point messages, and the target interest point messages are interest point messages selected by the users from the search results of the search words. The method relieves the technical problems that differences among search results of interest points and user expectations still exist in the prior art.

The embodiment of the invention provides a searching method and device, and the method comprises the steps: determining a label semantic space vector corresponding to a retrieval keyword, wherein thelabel semantic space vector is used for describing semantics of the retrieval keyword; calculating tag semantic space characteristics of the retrieval keyword according to the tag semantic space vector, wherein the tag semantic space characteristics are used for describing the similarity between the retrieval keyword and a retrieval result; and according to the tag semantic space features, sortingthe search results corresponding to the retrieval keywords. According to the embodiment of the invention, the accuracy of the search result can be effectively improved, and the fuzzy search result ofthe search term is optimized.

The invention relates to an approximate video retrieval method based on improvement of editing distance, and belongs to the computer video processing and mode identification field. The approximate video retrieval method comprises the steps: converting a video into an image frame sequence, and using a method based on scores to calculate the similarity of two frames, thus reducing the information loss during the similarity calculating process by using a word bag model; using the improved editing distance to calculate the similarity among video sequences, wherein the query range during the retrieval process is effectively reduced by calculating the relative editing distance similarity among frame sequences; and integrating with a sequence similarity scoring method based on a dynamic planning mechanism, thus effectively reducing the error results and improving the retrieval accuracy of the method.

The invention provides a remote sensing image multi-label retrieval method and system based on a full convolutional neural network. multi-label image retrieval is realized by considering multi-category information of remote sensing images, and the method comprises the following steps: inputting a retrieval image library, and dividing the retrieval image library into a training set and a verification set; Constructing a full convolutional neural network model FCN, and performing network training by using the training set; Performing multi-class label prediction on each image in the verificationset by using FCN to obtain a segmentation result; Carrying out up-sampling on each convolutional layer feature map; Extracting local features of each image in the verification set to obtain feature vectors for retrieval; And finally, carrying out coarse-to-fine two-step retrieval based on the extracted multi-scale features and the multi-label information. According to the method, the full convolutional neural network is used for learning multi-scale local features of the image, multi-label information hidden in the image is fully mined, and compared with an existing remote sensing image retrieval method based on a single label, the accuracy of image retrieval is effectively improved.

The invention provides a full-text database accurate and efficient retrieval method for perfecting subject terms. The method comprises the steps of firstly, creating a long-tail subject term lexicon,designing a long-tail subject term selection algorithm, selecting a user to optimize a search expression, and clearing the search intention; designing a subject term synonymy expansion algorithm basedon the word forest expansion version, and expanding the subject term into a synonymy set of the subject term; and designing an improved synonym similarity calculation method based on the Known synonym tree, and calculating the similarity between the expanded subject term set and the original subject term by adopting an improved algorithm; setting an appropriate similarity threshold value throughmultiple groups of experimental statistical analysis, and simplifying the subject term set according to the similarity critical value to obtain an improved subject term set; and finally, improving a retrieval result sorting algorithm. The recall ratio, the precision ratio and the retrieval efficiency of the method are greatly improved, and good performance is shown.

The invention discloses an interactive retrieval method and device, computer equipment and a storage medium, belongs to the crossing field of computer software and avionics systems, and is used for improving the retrieval efficiency of an electronic flight bag, guaranteeing the flight operation safety, enabling the flight operation of civil aviation aircrafts to be more modernized, ensuring flightsafety effectively, reducing the workload of pilots, improving economic benefits of airlines, and meanwhile, improving the core competitiveness of civil aviation airplanes. According to the main technical scheme, the method comprises the steps of S1, expanding a query statement input by a user in combination with a knowledge graph; s2, selecting a plurality of documents with higher similarity with the expanded query statement from a document library, and returning the documents to the user; and S3, adjusting the weight of each node in the knowledge graph according to user behaviors, and optimizing a retrieval result.

The invention discloses a multi-kernel hash learning-based large-scale medical image retrieval method. The method concretely comprises the steps of: constructing a kernel matrix through fusion of a plurality of different kernel functions; completely converting an image into hash codes by use of the learned hash function, and carrying out compression on the hash codes; solving distances of medical images through hamming distance measurement, sorting according to an ascending order, selecting m images with minimum distances to be returned to a user; optimizing the sorting of the retrieved images again by the user by use of a relevance feedback algorithm until the sorting satisfies the requirement of the user. The method is high in calculation efficiency, rapid in retrieval speed, small in memory space, high in retrieval precision, clear in steps and strong in pertinence, contributes to medical diagnosis of a doctor, reduces the workload of the doctor and improves the working efficiency.

The invention discloses a large-scale image high-speed retrieval method based on multi-view enhanced deep hash. The method comprises the following steps: step 1, acquiring image multi-view feature representation; step 2, calculating a view relation matrix; step 3, designing a loss function of the model; step 4, performing fusion and enhancement; step 5, training the built model on a large-scale image training data set; step 6, testing the trained model to generate a hash code, and then performing hash retrieval; step 7, carrying out an experiment to evaluate the index. According to the method,the influence of Hamming radius extension on the result is small; and along with the increase of the code length, the precision is kept stable.

The invention provides a field information retrieval method based on behaviors. The information retrieval method based on the behaviors can achieve optimization by means of department subordination relations of users and various interdependent user behaviors. The information retrieval method based on the behaviors includes the following steps that firstly, the department subordination relations of system users are obtained through analysis of an department subordination relations, namely which users belong to which departments, and activities performed in the process that the users use the information management system; secondly, the similarity degree of information is calculated through behavior data; thirdly, full-text retrieval is conducted on search terms inputted by the users, matching contents are obtained to the maximum to obtain full text matching results; fourthly, the full-text retrieval is filtered, the results are ranked according to interest reflected by the user behaviors, and the contents which the users are most interested in are ranked to the headmost position.

The invention provides a network page efficient and accurate deduplication system based on cloud computing, and aims to solve the problems that most of web pages searched by an existing search engineare static web pages, due to the existence of a large amount of transshipment and plagiarism, the main content of a large number of web pages is repeated, and for the search engine, the repeated web pages virtually increase the burden of index storage, and meanwhile, more retrieval time can be consumed; the webpage deduplication system based on the Hadoop cloud platform is designed and realized bycombining an open source framework, other modules of a search engine can be better connected by adopting a mode of detecting and judging duplicate in real time after a spider program captures a webpage; and in a massive webpage collection stage, the network page efficient and accurate deduplication system based on cloud computing can preprocess the web pages in advance, then web page similarity detection and discovery are carried out, repeated web pages or web pages with high similarity are removed, and therefore index quality is improved, retrieval results are optimized, and good search experience is provided for users.

The invention discloses a multi-dimensional data feature selection method combining a genetic algorithm and a dragonfly algorithm. The method comprises the following specific steps that traffic accident data is simply cleaned; the dragonfly algorithm is embedded into the genetic algorithm to intervene the crossover operation, and the optimal intersection position is found out through the dragonflyalgorithm to improve the optimization speed of the genetic algorithm. A dragonfly algorithm is embedded into a genetic algorithm to intervene mutation operation, and different gene position mutationprobabilities are set by judging whether food and natural enemy gene positions calculated by the dragonfly algorithm are selected or not, so that the convergence rate of the algorithm is increased. The data feature screening result is taken as the input of a multi-dimensional data feature selection method combining a genetic algorithm and a dragonfly algorithm, and the output result is the data feature selected by the algorithm. Experiments prove that the method has good performance for different classifiers, and it is verified that the feature selection method is effective and has robustness.

The invention provides a similar document retrieval auxiliary device and a similar document retrieval auxiliary method. The extent of the impact exerted on retrieval precision by an essential factor affecting similar document retrieval precision and information concerning countermeasures for improving retrieval precision are shown to a user, and therefore the retrieval work of the user can cyclically run in a highly-efficient manner, and the efficiency and quality of the retrieval work can be improved. Based on a right set of previous input documents and right answer documents, the essential factor is analyzed, and a correspondence relation is established between a value range of the essential factor and retrieval precision and is stored in a table. Through computer processing operation, newly input documents are subjected to the same essential analysis operation, retrieval precision corresponding to the value range which conforms with the essential factor value of the newly input documents is calculated based on comparison with the above table. Then, through computer processing, the retrieval precision and/or an average deviation value compared with the whole retrieval precision of the previous input documents is shown to the user. In a more ideal condition, information of countermeasures for improving the retrieval precision is shown to the user.

The invention provides a video index establishing method, comprising the following steps: A1, motion activity descriptors of all video frames which are contained in all video lenses are extracted, the motion activity descriptors which are corresponding to a plurality of video lenses and are obtained by extraction are trained, thus generating a corresponding rule set; B1, the rule set is queried according to the motion activity descriptors which are corresponding to the various video lenses, thus obtaining the corresponding type attributes of the lenses of the various video lenses; C1, a video index is established according to the type attributes of the lenses of the video lenses and the motion activity descriptors. The invention simultaneously provides a video index establishing device, an extraction method and a device of the motion activity descriptors, which can meet the delicate and the effective query requirements of a user and improve the user experience.