46results about How to "Meet search requirements" patented technology

The invention discloses a hot word recommending method and device with combination of multiple data sources. The method comprises the following steps: screening hot data of the multiple data sources; selecting search logs of one or more users; intersecting the screened hot data and the search logs to generate a recommended hot word list; selecting a certain quantity of recommended hot words from the recommended hot word list for display. On the basis of the method and the device, on one hand, the recommended hot words are associated with historical search record of the user; on the other hand, the recommended hot words are associated with the search trend of the current society, the integrated recommended hot words are more accurate, effective and reliable, the search requirements of the user are met, and the search experience of the user is improved.

The invention provides a vehicle image accurate retrieval method based on big data. The method comprises the steps that (1) the vehicle target and the vehicle front windshield area are detected by using a detection method based on the YOLO V2 algorithm; (2) a vehicle target attribute identification model based on a deep convolutional neural network is constructed and trained so as to identify thevehicle target attributes; (3) the ORB features and the feature descriptors of the front windshield area Ia of the suspected vehicle and the front windshield area Ib of a vehicle to be retrieved are extracted; (4) the Hamming distance between the ORB features of the image Ia and the image Ib is calculated, and the matching point pairs between Ia and Ib are found through the fast nearest neighbor method for violent matching; (5) the wrongly matched feature point pairs of the step (4) are eliminated by using the grid-based motion statistical method GMS; and (6) the ORB feature similarity betweenthe front windshield area of the suspected vehicle and the front windshield area of the vehicle to be retrieved is calculated. The problem of fast and accurate vehicle search in the massive traffic images can be solved on the whole.

The invention discloses an ontological semantic extension and collaborative filtering weight-fused process knowledge retrieval method. According to the method, weighted fusion is carried out on ontological semantic extension and vector space model-based knowledge matching method and a user attribute collaborative filtering matching method, and finally user personalized retrieval results accordingwith real retrieval intentions of users are obtained through forgetting curve-based weight control. According to the method, advantages of ontological knowledge retrieval and collaborative filtering are combined, dynamic and static attributes of users are sufficiently considered, and related information is integrated, so that retrieval results which more accord with user retrieval requirements andhave user personalization can be obtained and then sharing and reusing of war industry knowledges are improved.

The invention discloses a processing method and a processing device for searching recommendation words. The processing method comprises the following steps: obtaining keywords input in an searching bar; matching the keywords with local historical searching results to obtain first matched results; sending the keywords to a server end and receiving second matched results, which are matched with the keywords and returned by the server end; integrating the first matched results with the second matched results to obtain searching recommendation word data; and displaying the searching recommendation word in a drop-down box of the searching bar. According to the technical scheme provided by the invention, keywords which are input into the searching bar are matched locally and on the server end, on the one hand, the keywords can be associated with the historical searching habits of a user, on the other hand, the keywords can be associated with a searching trend of the current society, so that the searching recommendation word data which are relatively accurate, effective and reliable, and meets the searching needs of a user can be obtained, and the user can conveniently and freely select, and the searching efficiency and the searching experience of the user are improved.

The invention discloses a word meaning extracting method based on search interactive information and user search intention. The method comprises the following steps in order: recording historic interactive information of each user; using the query keywords in the first K1 rank of the user query frequency to construct a query keyword vector; using corresponding click results obtained after querying the query keywords in the first K2 ranks of the query frequency of each user to construct a corresponding click vector, and setting items corresponding to the query keywords in the first K2 ranks ofthe query frequency of each user in the query keyword vector as 1, and setting other items as zero; extracting a meaning item number of anyone of high-frequency query keywords; clustering the users; and computing the preference rank of the meaning item corresponding to each query keyword of users in the same class. The method can avoid the cost problem caused by manual marking; meanwhile, the problems that the derivative free method has poor effect and is limited by the field. A personalized search service can be provided for a single user according to the analysis result obtained by the method provided by the invention.

The invention discloses a method and device for information searching and belongs to the technical field of Internet. The method comprises the steps that an interactive statement used for the information searching is received; a target semantic attribute template is determined from a semantic attribute template set, wherein the target semantic attribute template has a one-to-one correspondence relation with semantic attributes of multiple continuous characters and words in the interactive statement; a phrase composed of the multiple continuous characters and words matched with the target semantic attribute template is acquired and taken as a single query condition of the interactive statement; and the searching is conducted according to the single query condition. According to the invention, the phrase composed of the multiple continuous characters and words matched with the target semantic attribute template is acquired and taken as the single query condition used for the information searching; semantic information denoted by the single query condition is complete and can be taken as a complete query limiting condition, so that searching results can satisfy a user's search demands to the hilt; and the problem of low search accuracy caused by information loss can be avoided.

The invention discloses a method for rapidly establishing a full-text retrieval tool for common files, belonging to the field of retrieval tools. The method specifically comprises the steps that: (1) a document analysis module reads all the file analysis HTTP requests and sends the requests to a Chinese words segmentation module; (2) the Chinese words segmentation module segments attribute content in the received HTTP requests; (3) a full-text retrieval establishment module customizes a retrieval service type; (4) a retrieval module after analyzing a retrieval command performs corresponding operation and finishes establishment of the retrieval tool; (5) after a user submits search terms, the retrieval module performs word segmentation treatment on the search terms, generates an inquiry request and inquires in an index library and takes on the inquired result to the user. The method for rapidly establishing the full-text retrieval tool for common files realizes establishment of search engine dedicated to personnel and enterprises, personal retrieval requirements can be realized only by taking relatively little time and effort, and a plenty of internal files can be managed easily.

A data processing method includes: generating a candidate search word set for a search word; determining a candidate search term for correcting the search term from the candidate search term set according to an appearance probability of any candidate search term in the candidate search term set and a semantic similarity between the candidate search term and the search term, wherein an occurrence probability of the candidate search term is calculated according to a context relationship of the candidate search term. In this way, the user's retrieval experience is improved.

The invention provides a book retrieval method and device and a readable storage medium. The method comprises the following steps of acquiring retrieval text by performing text recognition on retrieval voice input by a user; performing semantic analysis on the retrieval text to obtain a semantic word segmentation vector; processing the semantic word segmentation vector according to a trained question template model to obtain a question label corresponding to the retrieval text; retrieving the retrieval text according to a preset book atlas and the problem label to acquire a retrieval result and report the result to the user, and therefore the accuracy of the result obtained through retrieval is improved, and the real retrieval requirement of the user is met.

The invention discloses a sorting algorithm for standard literature retrieval, relates to the technical field of keyword retrieval. The sorting algorithm includes: influencing the boost scoring by setting the query configuration; packaging the method for realizing the edismax and the map function; carrying out quantitative regularization processing on a plurality of fields such as query title names, question records and text; setting the scoring weight for the text relevancy of each field, endowing the fields with different weight levels according to accurate and fuzzy matching; after two times of data regularization processing, carrying out reverse order sorting through the packaged map function, and the result is fed back, so that the accuracy in the standard literature retrieval processis improved.

The embodiment of the invention discloses a label reconstruction method and device. The method comprises the following steps: obtaining a target product label; carrying out label reconstruction on thetarget product label by using a deep neural network model component, and generating a reconstruction label of the target product label, wherein the deep neural network model component is obtained bytraining according to a corresponding relation between a plurality of historical original tags and a plurality of historical reconstruction tags. By utilizing the embodiment of the invention, the product label can be quickly reconstructed, and the display speed of the product reconstruction label is accelerated.

The invention provides a commodity searching method. The method comprises the steps that search terms for commodities are received, search results corresponding to the search terms are obtained, descriptors of the search results are matched with the search terms in a normalized word list, wherein the normalized word list is constructed based on a regular network representing inter-term relevance;and the search results are output. The invention further provides a commodity searching device, computer equipment and a computer readable storage medium.

The invention relates to a similar or repeated video retrieval method and device, the device comprises an application program module, a task scheduling module, a gene extraction module and a gene database module, the method comprises the following steps: firstly extracting video genes from all videos in a video database, and inserting the video genes into a Milvus vector database; during retrieval, and the gene of the video to be retrieved is also extracted, a retrieval request is initiated, and the gene database returns a retrieval result which specifically comprises video gene storage and similar video retrieval. The invention has the advantages that 1) compared with the prior art, the retrieval accuracy is higher, and the situation of missing detection can be reduced; (2) retrieval results are sorted according to the number of similar pictures, and the retrieval requirements of users are better met; 3) the retrieval result contains the overlap ratio of the to-be-retrieved video and the target video, so that the user can judge whether the target video is a repeated video or not; and 4) the invention does not depend on video text description information, and only needs to provide a video file.

The embodiment of the invention discloses an information search method and a device, electronic equipment and a storage medium, and the method comprises the steps: receiving a search request which comprises a search keyword; searching based on the search keyword to obtain related information of at least one candidate object; determining a relationship between the at least one candidate object based on the related information of the at least one candidate object; based on the relationship between the at least one candidate object, obtaining a sorting value of the at least one candidate object;and sorting the at least one candidate object based on the at least one candidate object sorting value, and outputting a search result. According to the method of the invention, the search results better meet the search requirements of most users in sorting, so that the search efficiency is improved.

The invention discloses a search method based on an electronic book, a server, a client and a system. The method comprises the following steps: receiving a target search term sent by a client; determining at least one co-occurrence search word corresponding to the target search word; aiming at any co-occurrence search word, obtaining a book word associated with the co-occurrence search word; and sending the book words to the client, so that the book words are displayed on the client. By adopting the scheme, the search word with strong semantic association with the target search word can be obtained, the search result obtained based on the co-occurrence search word not only can meet the search requirement of the user, but also can provide diversified search results for the user, and the hit rate of the search results is improved; in addition, according to the scheme, the book words associated with the co-occurrence search words are sent to the client side, the book search path of the user can be shortened by displaying the book words in the client side, and the book search efficiency is improved.

The invention relates to a matching degree computing method of sortable precise attributes of one-chain-type services. In a traditional producing and manufacturing electric business platform, parameters of part of attributes of the services/products are precise. It is needed that the parameters of requirements and the attributes are absolutely unified or compatible. Some products or services capable of meeting user requirements are eliminated from an alternative sample set. According to the values of the parameters of the attributes, the attributes of the alternative services/products are ranked so that a one-way sequence can be formed, and a similarity calculation formula between nodes and a target node on a chain is provided. According to the similarity calculation formula, the matching degrees between the required attributes and the attributes of the services/products can be calculated, then, a weighing mode is used, and the overall matching degree between the requirements and the services/products is obtained. The method is high in flexibility degree, the intrinsic requirement of a user is met better, and the matching degree computing method is efficient, fast and close to the requirements.

The invention provides a method and device for providing related people information in a search result. The method comprises the following steps: receiving a search request for searching a first personnel name, and searching according to the search request to obtain a first search result corresponding to the first personnel name; determining one or multiple second personnel names with the highest relevancy with the first personnel name according to one or multiple hotspot search data, and acquiring people information corresponding to the second people names; displaying a search result page including a first interface and a second interface, wherein the first interface is used for displaying the first search result corresponding to the first people name, and the second interface is used for displaying people information corresponding to the second people names. According to the technical scheme, when searching based on the people names, a user can see the related people information corresponding to the people names with the highest relevancy with the currently searched people name at one glance on the search result page, so that the use experience of the user for a search engine is enhanced.