Digital archive intelligent retrieval method and system for storage device

Abstract

The invention discloses a digital archive intelligent retrieval method and system for a storage device, and relates to the technical field of information, and the method comprises the following steps: inputting information by a terminal user, preprocessing to obtain an associated retrieval string, and performing multi-dimensional retrieval in a digital archive by using the associated retrieval string to obtain a plurality of associated information sets; historical retrieval records of a terminal user are obtained, the browsing tendency of the terminal user is analyzed, and a retrieval associated information set is obtained by screening multiple associated information sets and fed back to the terminal user. The technical problem that personalized and targeted retrieval results cannot be accurately and efficiently provided for users in the face of massive digital archives in the existing retrieval technology is solved, the users can quickly and effectively obtain valuable personalized information by accurately matching user requirements with the retrieval results, the functions of the digital archives are fully played, and the user experience is improved. The technical effect of improving the efficiency and satisfaction degree of obtaining the digital archive information by the user is achieved.

Description

Technical Field

[0001] The present application relates to the field of information technology, and in particular to a method and system for intelligent retrieval of digital archives for storage devices. Background Art

[0002] In the digital archive management scenario, with the explosive growth of information, the scale and complexity of digital archives are increasing day by day, and efficient and accurate retrieval and utilization of digital archives have become a crucial link. Traditional digital archive retrieval methods often have many limitations, such as relying only on simple keyword matching, failing to fully understand the user's true intentions, insufficient consideration of the individualization of user needs, lack of analysis and utilization of user historical retrieval behavior, and difficulty in providing comprehensive and accurate retrieval results that meet user expectations. When faced with massive digital archives, the retrieval efficiency is low and the accuracy is poor, which cannot well meet the ever-changing and increasingly complex retrieval needs of users.

[0003] At present, there is a technical problem in the relevant technologies that the retrieval technology cannot accurately and efficiently provide users with personalized and targeted retrieval results when faced with massive digital archives. Summary of the invention

[0004] The present application provides a digital archive intelligent retrieval method and system for storage devices, which uses terminal user input information, obtains related search strings after preprocessing, and uses them to multi-dimensionally search in digital archives to obtain multiple related information sets. The terminal user's historical search records are obtained and their browsing tendencies are analyzed, and multiple related information sets are screened based on them to obtain search related information sets, which are fed back to the terminal user, so that users can quickly and effectively obtain valuable personalized information by accurately matching user needs with search results, giving full play to the role of digital archives, and achieving the technical effect of improving the efficiency and satisfaction of users in obtaining digital archive information.

[0005] The present application provides a digital archive intelligent retrieval method for a storage device, comprising:

[0006] Interactively obtain input information from a terminal user; preprocess the input information to obtain a related search string; use the related search string to perform multi-dimensional data retrieval in a digital archive to obtain multiple related information sets; interactively obtain the terminal user's historical search records, and obtain the terminal user's browsing tendency based on the historical search record analysis; perform information screening on the multiple related information sets according to the browsing tendency to obtain a search-related information set; and feed back the search-related information set to the terminal user.

[0007] The present application also provides a digital archive intelligent retrieval system for a storage device, comprising:

[0008] An input information acquisition module, the input information acquisition module is used to interactively obtain input information of a terminal user; an associated search string acquisition module, the associated search string acquisition module is used to preprocess the input information to obtain an associated search string; a multidimensional data retrieval module, the multidimensional data retrieval module is used to use the associated search string to perform multidimensional data retrieval in a digital archive to obtain multiple associated information sets; a browsing tendency acquisition module, the browsing tendency acquisition module is used to interactively obtain the historical search records of the terminal user, and obtain the browsing tendency of the terminal user based on the historical search record analysis; an information screening module, the information screening module is used to perform information screening on the multiple associated information sets according to the browsing tendency to obtain a search associated information set; a search associated information set feedback module, the search associated information set feedback module is used to feed back the search associated information set to the terminal user.

[0009] The present application also provides a storage device, which is a Blu-ray memory. An intelligent retrieval engine is stored on the storage device. When the intelligent retrieval engine is executed by a processor, a digital archive intelligent retrieval method for the storage device is implemented.

[0010] The proposed method and system for intelligent retrieval of digital archives for storage devices is firstly provided by the terminal user, and the related retrieval string is obtained through preprocessing, and then multiple related information sets are obtained by multi-dimensional retrieval in the digital archive. The historical retrieval records of the terminal user are obtained and their browsing tendency is analyzed, and multiple related information sets are screened accordingly to obtain the retrieval related information set, which is fed back to the terminal user. By accurately matching the user's needs with the retrieval results, the user can quickly and effectively obtain valuable personalized information, give full play to the role of digital archives, and achieve the technical effect of improving the efficiency and satisfaction of users in obtaining digital archive information. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0012] Figure 1 A schematic diagram of a flow chart of a digital archive intelligent retrieval method for a storage device provided in an embodiment of the present application;

[0013] Figure 2 A schematic diagram of the structure of a digital archive intelligent retrieval system for a storage device provided in an embodiment of the present application.

[0014] Explanation of reference numerals: input information acquisition module 10 , associated search string acquisition module 20 , multi-dimensional data retrieval module 30 , browsing tendency acquisition module 40 , information screening module 50 , retrieval associated information set feedback module 60 . DETAILED DESCRIPTION

[0015] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0016] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0017] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0018] The present application embodiment provides a method for intelligently retrieving digital archives for a storage device, such as Figure 1 As shown, the method includes:

[0019] Step S100, interactively obtain input information from the terminal user. Specifically, the system provides the terminal user with an input interface that is easy to operate and understand. The interface can be in the form of a web page, an application program interface, or a voice input interface, etc. The user can enter text descriptions, keywords, and sentences through the keyboard, or click to select preset options to express his or her search needs. If it is voice input, the system will start the voice recognition function to convert the user's voice into processable text information. During the user's input process, the system will monitor the input status in real time to ensure the completeness and accurate reception of the information. For example, for the input text, it will check whether there are grammatical errors or incomplete expressions, and give prompts when necessary. Once the user completes the input and confirms the submission, the system will immediately obtain the input information and pass it to the subsequent processing link to prepare for the subsequent preprocessing and retrieval work.

[0020] Step S200, preprocessing the input information to obtain an associated search string. Specifically, after receiving the input information from the terminal user, the system performs a preprocessing operation, cleans the input information, removes possible noise data, such as special characters, meaningless spaces, etc., performs text error correction, checks for spelling errors or grammatical errors in the input information, and attempts to automatically correct them, uses natural language processing technology to segment the text, splits the input information into meaningful word units, extracts keywords, and selects keywords that can represent the core content of the input information from the segmented words through word frequency statistics, part-of-speech analysis, etc., generates an associated search string based on keywords and related semantic rules, involving the combination, expansion or conversion of keywords to form a search string that more accurately reflects the user's search intention, and the system will continuously optimize and adjust the processing steps to ensure that the associated search string finally obtained can best meet the user's real needs and provide strong support for subsequent accurate retrieval.

[0021] In one possible implementation, the input information is preprocessed to obtain an associated search string, and step S200 further includes step S210, synchronizing the input information to a pre-built search intent analysis model to perform search intent analysis and obtain conventional associated search information. Specifically, it is necessary to pre-build a model specifically for search intent analysis. During the construction process, word embedding technology is used to convert words in the input text information into vector representations for subsequent processing. A convolutional neural network or a recurrent neural network is used to capture the sequence features and contextual information of the input text. An attention mechanism is introduced to focus on key parts of the input text, and different weights are assigned to different words or phrases to more accurately understand the user's intent. The extracted features are comprehensively analyzed through a fully connected layer, so that when the input information is received, the user's possible search intent can be inferred, and a series of conventional associated search information is generated based on this.

[0022] Step S220, based on the historical search records, recommendation result feedback analysis is performed to obtain search restrictions. Specifically, the user's historical search records are obtained, data cleaning and preprocessing are performed on the records to remove noise and outliers, keywords in the historical search records are extracted using algorithms such as word frequency-inverse document frequency, similar historical search behaviors are clustered into one category through clustering analysis methods such as hierarchical clustering, and the clustering results are analyzed to find out the patterns that users frequently avoid or are not interested in. According to the analysis results, rules such as browsing time thresholds and click count thresholds are set to determine which behavior patterns are considered undesirable, thereby forming search restrictions.

[0023] Step S230, the associated search string is screened out from the conventional associated search information according to the search restriction. Specifically, the conventional associated search information obtained previously is compared and matched with the determined search restriction, and the conventional associated search information is screened and filtered one by one according to the threshold and rules set in the search restriction, and those information that do not meet the search restriction are removed, leaving the part that meets the requirements, and finally forming a more accurate associated search string that meets the user's needs. The associated search string fully considers the user's search intention while also complying with the search restriction, and can provide an accurate basis for subsequent retrieval and information screening in a more targeted manner.

[0024] In a possible implementation, the input information is synchronized to a pre-built search intent analysis model to perform search intent analysis and obtain conventional associated search information. Step S210 further includes: Step S211, pre-building a first search expansion branch, wherein the first search expansion branch is generated based on a cognate word expansion algorithm. Specifically, the principles and rules of the cognate word expansion algorithm are clarified, for example, determining how to identify the root of the input word, and expanding related cognate words based on the root, and programming implementation is performed according to the rules and principles. During the implementation process, it is necessary to establish a vocabulary to store common roots and their corresponding cognate words, and through continuous testing and optimization of the code, ensure that the first search expansion branch can accurately and quickly generate related cognate word expansion results based on the input word.

[0025] Step S212, pre-constructing a second search expansion branch, wherein the second search expansion branch is constructed and generated based on a keyword matching algorithm. Specifically, the logic of the keyword matching algorithm is first designed, including determining the characteristics of the keyword, the matching mode and the similarity calculation method, etc., using the corresponding programming language and data structure to implement the algorithm, it is necessary to establish a keyword database, and optimize the efficiency of the matching process, and through repeated testing and adjustment, ensure that the second search expansion branch can effectively find matching keywords according to the input information.

[0026] Step S213, pre-constructing the third search extension branch, wherein the third search extension branch is constructed and generated based on the knowledge graph retrieval algorithm. Specifically, first construct the knowledge graph, organize various entities, relationships and attributes in the form of a graph, design a retrieval algorithm based on the knowledge graph, determine how to perform path search, relationship reasoning and information extraction in the graph, use appropriate technologies and frameworks to implement the algorithm, and effectively integrate it with the knowledge graph. After sufficient testing and optimization, ensure that the third search extension branch can obtain valuable extended information from the knowledge graph.

[0027] Step S214, connect the first search expansion branch, the second search expansion branch and the third search expansion branch in parallel to obtain the search intent analysis model. Specifically, after completing the separate construction of the three search expansion branches, design a mechanism to connect them in parallel, which involves creating a unified interface so that input information can be passed to the three branches at the same time and the results returned by them can be integrated. Through the programming implementation of data communication and coordinated control, it is ensured that the three branches can work together, and the overall model after parallel connection is comprehensively tested and verified to ensure that it can accurately and comprehensively analyze the search intent.

[0028] In a possible implementation, based on the historical search record, recommendation result feedback analysis is performed to obtain search restrictions, and step S220 further includes step S221, wherein the historical search record includes multiple historical recommendation information and multiple historical browsing information of multiple historical search behaviors. Specifically, the system collects and stores multiple historical recommendation information received by the user in multiple past search behaviors, and the recommendation information is various documents, web links, pictures, etc., and records multiple historical browsing information of the user for these recommendation information, including browsing time, depth, repeated browsing, etc.

[0029] Step S222: Reversely extract multiple historical unviewed information from the multiple historical recommended information according to the multiple historical browsing information. Specifically, by analyzing the multiple historical browsing information, determine the recommended information that the user has actually browsed, remove the browsed part from all the multiple historical recommended information, and reversely extract multiple historical unviewed information that the user has not browsed.

[0030] Step S223, extract keywords from the multiple historical unbrowsed information to obtain multiple taboo keyword groups. Specifically, natural language processing technology and keyword extraction algorithms are used to analyze the text content of the multiple historical unbrowsed information, extract keywords that can represent the theme or core content of the unbrowsed information, and form multiple taboo keyword groups.

[0031] Step S224: preset a taboo frequency threshold. Specifically, based on experience, data analysis or general rules of user behavior, a taboo frequency threshold is set in advance, and the threshold is used to determine which keywords are considered to be significantly uninteresting or avoided by users.

[0032] Step S225, after aggregating the multiple taboo keyword groups, extract K taboo keywords that meet the taboo frequency threshold to form the search restriction. Specifically, the extracted multiple taboo keyword groups are integrated and summarized, and according to the preset taboo frequency threshold, K taboo keywords whose occurrence frequency reaches or exceeds the threshold are screened from the aggregated keyword set. The keywords together constitute the search restriction, which is used to avoid the appearance of related content in subsequent searches and recommendations. The whole process is to analyze the browsing situation of multiple search recommendation results under multiple search behaviors of the user's history, infer the subject name of the unbrowsed information, and then aggregate to obtain a keyword set that the user does not like, so as to form an effective search restriction.

[0033] Step S300, use the associated search string to perform multi-dimensional data search in the digital archive to obtain multiple associated information sets. Specifically, the multi-dimensional data search function of the digital archive is activated. The multi-dimensional search is not only based on simple keyword matching, but also searches the digital archive from multiple dimensions and angles. For example, multiple attributes such as text content, creation time, author, and category are searched at the same time. The system compares and matches the associated search string with various data in the digital archive. In this process, advanced search algorithms and data structures are used to improve the efficiency and accuracy of the search. As the search proceeds, the system will gradually collect various information related to the associated search string. The information comes from different data sources, different file formats, and different storage locations. All relevant information that meets the conditions is integrated together to form multiple associated information sets. Each associated information set contains a group of digital archive data related to the associated search string in different aspects.

[0034] In a possible implementation, the associated search string is used to perform multi-dimensional data retrieval in the digital archive to obtain multiple associated information sets. Step S300 further includes step S310, and the information structure of the digital archive includes text data, picture data, video data and audio data. Specifically, the text data includes various documents, such as reports, papers, novels, news, etc. The text data is collected and stored in the form of electronic documents, and its format may be txt, doc, pdf, etc. In terms of picture data, it covers a variety of types such as photos, illustrations, charts, and design drawings. The formats of pictures are rich and varied, and the common ones are jpg, png, gif, etc., which exist in the digital archive in the form of image files. Video data includes movies, TV series, documentaries, short videos, teaching videos, etc. The video formats are mp4, avi, mkv, etc., which are saved in the form of a large number of data files. Audio data includes music, audio books, lecture recordings, radio programs, etc. Common audio formats include mp3, wav, flac, etc., which are also stored in specific audio file formats. When storing these different types of information, a suitable database management system and storage architecture will be used to ensure that all types of information can be effectively organized, indexed and accessed. In order to ensure the security and integrity of the information, backup, encryption and other technical means will also be used. Through the comprehensive collection and orderly management of text, image, video and audio materials, a complete digital archive information structure is formed, providing a rich resource base for subsequent retrieval and utilization.

[0035] In a possible implementation, the associated search string is used to perform multi-dimensional data search in the digital archive to obtain multiple associated information sets, and step S300 further includes step S320, performing permutation and combination enumeration on the associated search string to obtain multiple search phrases, wherein the number of search terms in each search phrase is greater than 3. Specifically, the associated search string obtained through the previous processing is obtained, and various possible permutations and combinations are performed on the words in the associated search string. During the combination process, it is ensured that the number of search terms contained in each generated search phrase is greater than 3, and multiple different search phrases are obtained by exhaustively enumerating all possible combinations.

[0036] Step S330, using an inverted index to quickly locate and extract multiple groups of documents including the multiple search phrases in the text data. Specifically, using a pre-established inverted index structure, the inverted index infers the subject names of unbrowsed information based on the browsing conditions of multiple search recommendation results under multiple search behaviors in the user's history, and then obtains a set of keywords that the user does not like through subject name aggregation to construct, and taking multiple search phrases as input, the inverted index can quickly locate the text position containing the phrase, and extract the corresponding multiple groups of documents from the position.

[0037] Step S340: Aggregate the multiple sets of documents to obtain a first associated information set. Specifically, gather the multiple sets of documents extracted, sort and merge the documents, remove duplicate parts, and form a unified set, namely, the first associated information set.

[0038] Step S350, and so on, use the multiple search phrases to perform multi-dimensional data search in the digital archive to obtain the multiple related information sets. Specifically, in addition to searching in text materials, multiple search phrases are also used to perform similar search operations in other dimensions of the digital archive, such as picture materials, video materials, and audio materials. According to the same process, related information is located, extracted, and aggregated, thereby obtaining multiple related information sets of different dimensions. This process is repeated until the search and information collection are completed in all required dimensions, and finally multiple complete related information sets are obtained.

[0039] Step S400, interactively obtain the historical search records of the terminal user, and obtain the browsing tendency of the terminal user based on the analysis of the historical search records. Specifically, the system interacts with the terminal user to obtain the user's historical search records. The system will request relevant permissions from the user to ensure that these historical records can be accessed and used legally and in compliance with regulations. After obtaining the historical search records, the system will conduct an in-depth analysis of the records. The content of the analysis includes but is not limited to the keywords searched by the user each time, the time and frequency of the search, the search results viewed, the time spent on each result, whether there are repeated searches for the same or similar topics, etc. For the user's browsing behavior, the system will focus on the specific content that the user clicks to view, the length of time the user views it, whether there are operations such as collection or sharing, and consider new browsing permissions. If the user sets specific browsing permissions for certain search results, such as only being able to browse them themselves, If the information is visible to only certain users or specific groups, the system will process the search results with special permissions accordingly during the analysis process, and partially hide or specially mark them according to the permissions. By comprehensively analyzing these historical search records and browsing behaviors, the system can infer the user's interest preferences, key areas of concern, and browsing habits, thereby deriving the user's browsing tendencies. For example, if the user frequently searches for and browses science and technology information for a long time, the system can determine that the user has a high interest and tendency in the field of science and technology; if the user often searches in a specific time period, it may indicate that the user has a stronger information demand in that time period. Based on the browsing tendencies derived from the analysis, the system can provide users with search services and information recommendations that are more in line with their needs and preferences.

[0040] In a possible implementation, the historical search records of the terminal user are interactively obtained, and the browsing tendency of the terminal user is obtained based on the analysis of the historical search records. Step S400 further includes step S410, which decomposes and reorganizes the multiple historical browsing information based on the information structure of the digital archive to obtain historical text browsing information, historical picture browsing information, historical video browsing information, and historical audio browsing information. Specifically, multiple historical browsing information of the terminal user is obtained, and the browsing information is decomposed and reorganized according to the information structure classification of the digital archive. For example, according to the type of information, the browsing information is divided into text-related, picture-related, video-related, and audio-related. Through the operation, the historical text browsing information, historical picture browsing information, historical video browsing information, and historical audio browsing information are obtained respectively.

[0041] Step S420, calculate the proportion of the historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information to obtain a first information allocation strategy. Specifically, count the number or duration of the historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information, calculate the proportion of each type of browsing information in the total browsing information, and determine the first information allocation strategy based on the proportion, which reflects the user's relative tendency to browse different types of information.

[0042] Step S430, interactively obtain the historical collection information of the terminal user, and calculate the proportion of the historical collection information based on the information composition of the digital archive to obtain the second information allocation strategy. Specifically, the system interacts with the terminal user to obtain the user's historical collection information, and also classifies the historical collection information into types such as text, picture, video and audio according to the information composition of the digital archive, calculates the proportion of each type of collection information, and thus obtains the second information allocation strategy, which reflects the user's preference for different types of information in the collection behavior.

[0043] Step S440, interactively obtain multi-dimensional browsing time constraints. Specifically, through interaction with the user, the restriction conditions on browsing time are obtained. The multi-dimensional browsing time constraints are determined by inferring the topic names of the unbrowsed information based on the multiple browsing situations of multiple search recommendation results under the user's historical multiple search behaviors, and then obtaining the keyword set that the user does not like through topic name aggregation.

[0044] Step S450, based on the multi-dimensional browsing time constraint, the browsing deviation of the historical text browsing information, the historical picture browsing information, the historical video browsing information and the historical audio browsing information is calculated to obtain a third information allocation strategy. Specifically, the multi-dimensional browsing time constraint is applied to the historical browsing information, the difference between the actual time of each type of browsing information and the constraint time is compared, the browsing deviation is calculated, and based on the deviation data, the third information allocation strategy is determined to reflect the deviation and tendency of the user in browsing time.

[0045] Step S460, average calculation is performed on the first information allocation strategy, the second information allocation strategy and the third information allocation strategy to obtain the browsing tendency. Specifically, the relevant data of the first, second and third information allocation strategies obtained above are integrated, and their average values ​​are calculated to obtain a comprehensive browsing tendency, which can more comprehensively and accurately reflect the user's overall preference and behavior pattern in digital archive browsing, and provide an important basis for subsequent information recommendation and retrieval services.

[0046] Step S500: performing information screening on the plurality of related information sets according to the browsing tendency to obtain a search related information set. Specifically, the system obtains the user's browsing tendency obtained from the previous analysis. The browsing tendency is calculated by integrating multiple factors such as the user's historical browsing information, favorite information, browsing time, etc., and can more accurately reflect the user's interest preferences and information demand direction. The system obtains multiple related information sets previously obtained through multi-dimensional data retrieval. The information set contains relevant information of various types and topics. The system uses the user's browsing tendency as the standard and basis for screening. For each related information set, the system will analyze the information content in detail. If the information content is highly relevant to the user's browsing tendency, such as meeting the topics and types that the user often browses or having similarities with the information collected by the user, then the information is considered to meet the screening conditions. If the information is significantly different from the user's browsing tendency, or is a type and topic that the user has shown no interest in in historical browsing, then the information will be excluded from the screening results. During the screening process, the system will continuously compare and evaluate the specific content of each related information set to ensure that the screened information fits the user's interests and needs to the greatest extent possible. Through screening, the final result is an optimized and selected retrieval related information set that can better meet the user's expectations and needs.

[0047] Step S600, the retrieval-related information set is fed back to the terminal user. Specifically, the system completes the sorting and optimization of the retrieval-related information set to ensure the accuracy and effectiveness of the information. The system sends the sorted retrieval-related information set to the terminal user through the pre-set communication channel and data transmission method. If it is in a web application, the system will load the information on the corresponding page and display it to the user through the user's browser. The layout of the page will be carefully designed to present the retrieval-related information in a clear and intuitive way, and will be arranged according to relevance, importance or other user-preferred sorting methods. For mobile applications, the information will be displayed to the user through in-application message push or a specific interface. In order to improve the user experience, it will be provided. Provide concise and clear titles and summaries to help users quickly understand the main content of the information. If feedback is provided via email, the system will generate an email with standard format and clear content, and will elaborate on the key content of the retrieval-related information set in the body of the email, and may attach relevant links or attachments for users to further view and download. While providing feedback, the system will also record relevant data of this feedback, such as the time of feedback, the status of user receipt, etc., for subsequent tracking and optimization of services. The system will accurately, promptly and effectively deliver the carefully screened and organized retrieval-related information sets to end users in a variety of possible ways to help users obtain the information they need.

[0048] The embodiment of the present application uses the terminal user input information, obtains the associated search string after preprocessing, and uses it to multi-dimensionally search in the digital archive to obtain multiple associated information sets. The terminal user's historical search records are obtained and their browsing tendencies are analyzed, and multiple associated information sets are screened accordingly to obtain the search associated information set, which is fed back to the terminal user, so that the user can quickly and effectively obtain valuable personalized information by accurately matching the user's needs with the search results, giving full play to the role of the digital archive, and achieving the technical effect of improving the efficiency and satisfaction of users in obtaining digital archive information.

[0049] In the above, refer to Figure 1 A digital archive intelligent retrieval method for a storage device according to an embodiment of the present invention is described in detail. Figure 2 A digital archive intelligent retrieval system for a storage device according to an embodiment of the present invention is described.

[0050] According to an embodiment of the present invention, a digital archive intelligent retrieval system for storage devices is used to solve the technical problem that the existing retrieval technology cannot accurately and efficiently provide users with personalized and targeted retrieval results when facing massive digital archives. By accurately matching user needs with retrieval results, users can quickly and effectively obtain valuable personalized information, give full play to the role of digital archives, and achieve the technical effect of improving the efficiency and satisfaction of users in obtaining digital archive information. A digital archive intelligent retrieval system for storage devices includes: an input information acquisition module 10, an associated search string acquisition module 20, a multi-dimensional data retrieval module 30, a browsing tendency acquisition module 40, an information screening module 50, and a retrieval associated information set feedback module 60.

[0051] The input information acquisition module 10 is used to interactively obtain input information from a terminal user.

[0052] The associated search string acquisition module 20 is used to pre-process the input information to obtain an associated search string.

[0053] The multi-dimensional data retrieval module 30 is used to perform multi-dimensional data retrieval in the digital archive using the associated search string to obtain multiple associated information sets.

[0054] The browsing tendency acquisition module 40 is used to interactively acquire the historical search records of the terminal user, and acquire the browsing tendency of the terminal user based on the analysis of the historical search records.

[0055] The information screening module 50 is used to screen the multiple related information sets according to the browsing tendency to obtain a search related information set.

[0056] The search-related information set feedback module 60 is used to feed back the search-related information set to the terminal user.

[0057] The specific configuration of the associated search string acquisition module 20 will be described in detail below. As described above, the input information is preprocessed to obtain the associated search string, and the associated search string acquisition module 20 further includes: a conventional associated search information acquisition unit, the conventional associated search information acquisition unit is used to synchronize the input information to a pre-built search intent analysis model to perform search intent analysis and obtain conventional associated search information; a feedback analysis unit, the feedback analysis unit is used to perform recommendation result feedback analysis based on the historical search records and obtain search restrictions; an associated search string screening unit, the associated search string screening unit is used to filter out the associated search string from the conventional associated search information according to the search restrictions.

[0058] Among them, the input information is synchronized to a pre-built retrieval intent analysis model to perform retrieval intent analysis to obtain general related retrieval information, and the general related retrieval information acquisition unit further includes: a first retrieval expansion branch construction subunit, the first retrieval expansion branch construction subunit is used to pre-build a first retrieval expansion branch, wherein the first retrieval expansion branch is constructed and generated based on a cognate expansion algorithm; a second retrieval expansion branch construction subunit, the second retrieval expansion branch construction subunit is used to pre-build a second retrieval expansion branch, wherein the second retrieval expansion branch is constructed and generated based on a keyword matching algorithm; a third retrieval expansion branch construction subunit, the third retrieval expansion branch construction subunit is used to pre-build a third retrieval expansion branch, wherein the third retrieval expansion branch is constructed and generated based on a knowledge graph retrieval algorithm; a retrieval intent analysis model acquisition subunit, the retrieval intent analysis model acquisition subunit is used to connect the first retrieval expansion branch, the second retrieval expansion branch and the third retrieval expansion branch in parallel to obtain the retrieval intent analysis model.

[0059] Among them, based on the historical search records, recommendation result feedback analysis is performed to obtain search restrictions, and the feedback analysis unit further includes: a historical search record composition subunit, the historical search record composition subunit is used for the historical search records to include multiple historical recommendation information and multiple historical browsing information of multiple historical search behaviors; a reverse extraction subunit, the reverse extraction subunit is used to reversely extract from the multiple historical recommendation information according to the multiple historical browsing information to obtain multiple historical unbrowsed information; a keyword extraction subunit, the keyword extraction subunit is used to perform keyword extraction on the multiple historical unbrowsed information to obtain multiple taboo keyword groups; a taboo frequency threshold preset subunit, the taboo frequency threshold preset subunit is used to preset the taboo frequency threshold; a search restriction composition subunit, the search restriction composition subunit is used to extract K taboo keywords that meet the taboo frequency threshold to constitute the search restriction after aggregating the multiple taboo keyword groups.

[0060] The specific configuration of the multi-dimensional data retrieval module 30 will be described in detail below. As described above, the multi-dimensional data retrieval is performed in the digital archive using the associated search string to obtain multiple associated information sets. The multi-dimensional data retrieval module 30 further includes: an information composition unit for the digital archive, and the information composition unit for the digital archive is used to form the information of the digital archive including text data, picture data, video data and audio data.

[0061] Among them, the multi-dimensional data retrieval module 30 further includes: a search phrase acquisition unit, the search phrase acquisition unit is used to perform permutations, combinations and enumerations on the associated search strings to obtain multiple search phrases, wherein the number of search terms in each search phrase is greater than 3; a document extraction unit, the document extraction unit is used to use an inverted index to quickly locate and extract multiple groups of documents including the multiple search phrases in the text data; a first associated information set acquisition unit, the first associated information set acquisition unit is used to aggregate the multiple groups of documents to obtain a first associated information set; multiple associated information set acquisition units, the multiple associated information set acquisition units are used to perform multi-dimensional data retrieval in the digital archive using the multiple search phrases and obtain the multiple associated information sets.

[0062] The specific configuration of the browsing tendency acquisition module 40 will be described in detail below. As described above, the historical search records of the terminal user are interactively obtained, and the browsing tendency of the terminal user is obtained based on the analysis of the historical search records. The browsing tendency acquisition module 40 further includes: a decomposition and reorganization unit, which is used to decompose and reorganize the multiple historical browsing information based on the information structure of the digital archive to obtain historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information; a first information allocation strategy acquisition unit, which is used to calculate the proportion of the historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information to obtain a first information allocation strategy; a proportion calculation unit, which is used to calculate the proportion of the historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information to obtain a first information allocation strategy. Used to interactively obtain the historical collection information of the terminal user, and calculate the proportion of the historical collection information based on the information composition of the digital archive to obtain a second information distribution strategy; a multidimensional browsing time constraint acquisition unit, the multidimensional browsing time constraint acquisition unit is used to interactively obtain multidimensional browsing time constraints; a deviation calculation unit, the deviation calculation unit is used to calculate the browsing deviation of the historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information according to the multidimensional browsing time constraints to obtain a third information distribution strategy; a mean calculation unit, the mean calculation unit is used to calculate the mean of the first information distribution strategy, the second information distribution strategy and the third information distribution strategy to obtain the browsing tendency.

[0063] The digital archive intelligent retrieval system for storage devices provided by the embodiment of the present invention can execute the digital archive intelligent retrieval method for storage devices provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0064] In one embodiment, a storage device is provided. The storage device is a Blu-ray memory. An intelligent retrieval engine is stored on the storage device. When the intelligent retrieval engine is executed by a processor, a digital archive intelligent retrieval method for the storage device is implemented.

[0065] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0066] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A digital archive intelligent retrieval method for a storage device, characterized in that: The method comprises: Interactively obtain input information from end users; Preprocessing the input information to obtain a related search string; Using the associated search string to perform multi-dimensional data search in the digital archive to obtain multiple associated information sets; Interactively obtain the terminal user's historical search records, and obtain the terminal user's browsing tendency based on the historical search records; Perform information screening on the plurality of related information sets according to the browsing tendency to obtain a search related information set; The search-related information set is fed back to the terminal user.

2. A digital archive intelligent retrieval method for a storage device as claimed in claim 1, characterized in that: Preprocessing the input information to obtain a related search string, the method further comprising: Synchronize the input information to a pre-built search intent analysis model to perform search intent analysis and obtain conventional related search information; Performing recommendation result feedback analysis based on the historical search records to obtain search restrictions; The associated search string is filtered out from the regular associated search information according to the search restriction.

3. A digital archive intelligent retrieval method for a storage device as claimed in claim 2, characterized in that: The method further comprises: Pre-constructing a first search expansion branch, wherein the first search expansion branch is constructed and generated based on a cognate word expansion algorithm; Pre-constructing a second search expansion branch, wherein the second search expansion branch is generated based on a keyword matching algorithm; Pre-constructing a third search expansion branch, wherein the third search expansion branch is generated based on a knowledge graph retrieval algorithm; The first retrieval expansion branch, the second retrieval expansion branch and the third retrieval expansion branch are connected in parallel to obtain the retrieval intent analysis model.

4. The method according to claim 2, characterized in that Based on the historical search records, recommendation result feedback analysis is performed to obtain search restrictions, and the method further includes: The historical search record includes multiple historical recommendation information and multiple historical browsing information of multiple historical search behaviors; Reversely extracting from the plurality of historical recommendation information to obtain a plurality of historical unbrowsed information according to the plurality of historical browsed information; Extracting keywords from the plurality of historical unbrowsed information to obtain a plurality of taboo keyword groups; Preset taboo frequency threshold; After aggregating the plurality of taboo keyword groups, K taboo keywords satisfying the taboo frequency threshold are extracted to form the search restriction.

5. The method according to claim 1, characterized in that The information structure of the digital archive includes text data, picture data, video data and audio data.

6. The method according to claim 5, characterized in that The method further comprises: using the associated search string to perform multi-dimensional data search in the digital archive to obtain multiple associated information sets. Performing permutations, combinations and enumerations on the associated search strings to obtain multiple search phrases, wherein the number of search terms in each search phrase is greater than 3; Using an inverted index to quickly locate and extract multiple groups of documents including the multiple search phrases in the text data; Aggregating the multiple groups of documents to obtain a first associated information set; By analogy, the multiple search phrases are used to perform multi-dimensional data search in the digital archive to obtain the multiple related information sets.

7. The method according to claim 4, characterized in that Interactively obtaining the historical search records of the terminal user, and obtaining the browsing tendency of the terminal user based on the historical search records analysis, the method further includes: Decomposing and reorganizing the plurality of historical browsing information based on the information structure of the digital archive to obtain historical text browsing information, historical picture browsing information, historical video browsing information and historical audio browsing information; Calculating the proportion of the historical text browsing information, the historical picture browsing information, the historical video browsing information, and the historical audio browsing information to obtain a first information allocation strategy; Interactively obtain the historical collection information of the terminal user, and calculate the proportion of the historical collection information based on the information composition of the digital archive to obtain a second information distribution strategy; Interactively obtain multi-dimensional browsing time constraints; According to the multi-dimensional browsing time constraint, the browsing deviation of the historical text browsing information, the historical picture browsing information, the historical video browsing information and the historical audio browsing information is calculated to obtain a third information allocation strategy; The browsing tendency is obtained by calculating the mean of the first information allocation strategy, the second information allocation strategy and the third information allocation strategy.

8. A digital archive intelligent retrieval system for storage devices, characterized in that: The system is used to implement a digital archive intelligent retrieval method for a storage device according to any one of claims 1 to 7, and the system comprises: An input information acquisition module, the input information acquisition module is used to interactively obtain input information of a terminal user; An associated search string acquisition module, the associated search string acquisition module is used to pre-process the input information to obtain an associated search string; A multi-dimensional data retrieval module, the multi-dimensional data retrieval module is used to use the associated search string to perform multi-dimensional data retrieval in the digital archive to obtain multiple associated information sets; A browsing tendency acquisition module, the browsing tendency acquisition module is used to interactively obtain the historical search records of the terminal user, and obtain the browsing tendency of the terminal user based on the historical search records analysis; An information screening module, the information screening module is used to perform information screening on the multiple related information sets according to the browsing tendency to obtain a search related information set; A retrieval-related information set feedback module is used to feed back the retrieval-related information set to the terminal user.

9. A storage device, characterized in that: The storage device is a Blu-ray memory, and an intelligent search engine is stored on the storage device. When the intelligent search engine is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images