Data retrieval method and device, computer equipment and storage medium
Through hash processing and hash table query of cross-modal financial databases, multimodal retrieval of financial data assets is realized, the problem of low intelligence in the existing technology is solved, and the accuracy and intelligence of data retrieval is improved. .
Patent Information
- Application Number
- CN202411164578.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-08-22
AI Technical Summary
The existing data retrieval methods cannot meet the multimodal retrieval needs of financial data assets, and are less intelligent and cannot effectively query between different types of data assets.
Using a cross-modal financial database, the target features of the data to be queried through the feature extraction network are extracted, the target hash function is obtained, the hash value calculation is performed, the matching data is queried using the hash table, and the target data is determined according to the preset filtering rules, and the target data is determined. user.
It realizes fast and accurate query of multimodal data, improves the intelligence and accuracy of data retrieval, and meets the users' multimodal search needs.
Smart Images

Figure CN119046326A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the fields of data processing technology and financial technology, and in particular to data retrieval methods, devices, computer equipment and storage media. Background Art
[0002] With the advent of the big data era, for the financial industry, various financial enterprises have accumulated a large amount of financial data assets, and also precipitated a large amount of free assets, which include images, texts, audio and video, etc. With the continuous increase of information in the financial field, the acquisition and processing of financial data assets has become increasingly important.
[0003] At present, financial enterprises usually use tools such as MySQL databases to perform data retrieval and processing of financial data assets. However, as the amount and existence of financial data increase, the single-modal retrieval method based on MySQL database cannot meet the growing data query requirements. For MySQL database, when using MySQL database for data retrieval, only a simple search can be performed in a certain type of data assets, that is, only different types of data assets can be searched separately in the same category. For example, for a query request input by a user with a text to be queried, only text retrieval results related to the text to be queried will be output, and image retrieval results related to the text to be queried cannot be output. Therefore, the existing data retrieval method cannot meet the needs of multi-modal retrieval when searching data assets, and the intelligence of data retrieval processing is low. Summary of the invention
[0004] The purpose of the embodiments of the present application is to propose a data retrieval method, apparatus, computer equipment and storage medium to solve the technical problems that the existing data retrieval methods cannot meet the needs of multimodal retrieval when retrieving data assets and the intelligence of data retrieval processing is low.
[0005] In order to solve the above technical problems, the present application provides a data retrieval method, which adopts the following technical solution:
[0006] Receiving a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data;
[0007] Extracting target features of the data to be queried based on a preset feature extraction network;
[0008] Obtaining a target hash function corresponding to the cross-modal financial database;
[0009] Performing hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value;
[0010] Querying a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and querying a second hash value that meets a preset similarity condition with the target hash value from the hash table;
[0011] Querying the second target data matching the second hash value from the cross-modal financial database, and determining the third target data from the second target data according to a preset screening rule;
[0012] First target data matching the first hash value is queried from the cross-modal financial database, and the first target data and the third target data are returned to the user.
[0013] Furthermore, before the step of obtaining the target hash function corresponding to the cross-modal financial database, the method further includes:
[0014] Obtain a pre-collected test dataset;
[0015] Obtain a preset number of candidate hash functions;
[0016] Performing a test process on each of the candidate hash functions based on the test data set, and calculating the computational efficiency, collision rate and retrieval precision value of each of the candidate hash functions respectively;
[0017] Calculating a processing score of each candidate hash function based on the computational efficiency, collision rate, and retrieval precision value of each candidate hash function;
[0018] A hash function with the highest processing score is selected from all the candidate hash functions and used as the target hash function.
[0019] Furthermore, the step of calculating the processing score of each candidate hash function based on the computational efficiency, collision rate and retrieval precision value of each candidate hash function specifically includes:
[0020] Determine the preset weight setting algorithm;
[0021] Determine a first weight corresponding to the computational efficiency, determine a second weight corresponding to the collision rate, and determine a third weight corresponding to the retrieval precision value based on the weight setting algorithm;
[0022] Call the preset processing score statistics formula;
[0023] Based on the processing score statistical formula, the first weight, the second weight, the third weight, and the computational efficiency, collision rate and retrieval precision value of each candidate hash function are calculated and processed to obtain a processing score of each candidate hash function.
[0024] Furthermore, the step of searching the hash table for a second hash value that meets a preset similarity condition with the target hash value specifically includes:
[0025] Determine a target bucket corresponding to the target hash value from the hash table;
[0026] Performing similarity comparison processing on the target hash value and all the specified hash values contained in the target bucket respectively to obtain corresponding similarity comparison results;
[0027] Based on the similarity comparison result, a third hash value having a similarity comparison result greater than a preset similarity comparison threshold is screened out from all the specified hash values;
[0028] The third Hash value is used as the second Hash value.
[0029] Furthermore, the step of determining the third target data from the second target data according to a preset screening rule specifically includes:
[0030] Performing feature extraction on each of the second target data to obtain a plurality of specified features;
[0031] Mapping each of the designated features to a preset common subspace to obtain a plurality of first data, and mapping the target feature to the common subspace to obtain second data;
[0032] Based on a preset similarity measurement algorithm, respectively calculate the similarity between each of the first data and the second data;
[0033] Filtering out a specified similarity greater than a preset similarity threshold from all the similarities;
[0034] Acquire third data corresponding to the specified similarity from all the first data;
[0035] The matching data corresponding to the third data is screened out from the second target data to obtain the third target data.
[0036] Furthermore, the step of returning the first target data and the third target data to the user specifically includes:
[0037] generating sorting information of the first target data and the third target data;
[0038] Call the preset display list;
[0039] Based on the sorting information, filling the first target data and the third target data into the display list to obtain a target display list;
[0040] The target display list is returned to the user.
[0041] Furthermore, after the step of searching the hash table for a second hash value that meets a preset similarity condition with the target hash value, the method further includes:
[0042] Get the number of elements in the hash table;
[0043] Obtaining a capacity threshold corresponding to the hash table;
[0044] Calculating the difference between the capacity threshold and the number of elements;
[0045] Determine whether the difference is less than a preset value;
[0046] If so, the hash table is expanded based on a preset expansion base.
[0047] In order to solve the above technical problems, the embodiment of the present application further provides a data retrieval device, which adopts the following technical solution:
[0048] A receiving module, configured to receive a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data;
[0049] An extraction module, used for extracting target features of the data to be queried based on a preset feature extraction network;
[0050] A first acquisition module, configured to acquire a target hash function corresponding to the cross-modal financial database;
[0051] A first calculation module, configured to perform hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value;
[0052] A query module, configured to query a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and query a second hash value that meets a preset similarity condition with the target hash value from the hash table;
[0053] A first processing module, configured to query the cross-modal financial database for second target data matching the second hash value, and determine third target data from the second target data according to a preset screening rule;
[0054] The second processing module is configured to query the first target data matching the first hash value from the cross-modal financial database, and return the first target data and the third target data to the user.
[0055] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:
[0056] Receiving a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data;
[0057] Extracting target features of the data to be queried based on a preset feature extraction network;
[0058] Obtaining a target hash function corresponding to the cross-modal financial database;
[0059] Performing hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value;
[0060] Querying a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and querying a second hash value that meets a preset similarity condition with the target hash value from the hash table;
[0061] Querying the second target data matching the second hash value from the cross-modal financial database, and determining the third target data from the second target data according to a preset screening rule;
[0062] First target data matching the first hash value is queried from the cross-modal financial database, and the first target data and the third target data are returned to the user.
[0063] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:
[0064] Receiving a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data;
[0065] Extracting target features of the data to be queried based on a preset feature extraction network;
[0066] Obtaining a target hash function corresponding to the cross-modal financial database;
[0067] Performing hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value;
[0068] Querying a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and querying a second hash value that meets a preset similarity condition with the target hash value from the hash table;
[0069] Querying the second target data matching the second hash value from the cross-modal financial database, and determining the third target data from the second target data according to a preset screening rule;
[0070] First target data matching the first hash value is queried from the cross-modal financial database, and the first target data and the third target data are returned to the user.
[0071] Compared with the prior art, the embodiments of the present application have the following beneficial effects:
[0072] The present application first receives a data query request for a pre-built cross-modal financial database triggered by a user; then extracts the target features of the data to be queried based on a preset feature extraction network; then obtains a target hash function corresponding to the cross-modal financial database, and performs hash value calculation processing on the target features based on the target hash function to obtain a corresponding target hash value; subsequently, a first hash value identical to the target hash value is queried from a hash table corresponding to the cross-modal financial database, and a second hash value that meets a preset similarity condition with the target hash value is queried from the hash table; further, second target data matching the second hash value is queried from the cross-modal financial database, and third target data is determined from the second target data according to a preset screening rule; finally, first target data matching the first hash value is queried from the cross-modal financial database, and the first target data and the third target data are returned to the user. The present application uses a cross-modal financial database constructed based on a target hash function and a hash table to perform data retrieval processing on a target feature corresponding to the data to be queried contained in a data query request triggered by a user, so that the first target data and the third target data matching the target feature can be quickly and accurately queried and returned to the user to complete the data retrieval processing. Based on the cross-modal financial database, the user's multi-modal retrieval needs are effectively met, the intelligence of the data retrieval processing is improved, and the accuracy of the data retrieval processing is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS
[0073] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0074] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;
[0075] Figure 2 A flowchart of an embodiment of a data retrieval method according to the present application;
[0076] Figure 3 is a schematic structural diagram of an embodiment of a data retrieval device according to the present application;
[0077] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION
[0078] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.
[0079] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0080] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.
[0081] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.
[0082] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.
[0083] The terminal devices 101, 102, and 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, and the like.
[0084] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .
[0085] It should be noted that the data retrieval method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the data retrieval device is generally arranged in the server / terminal device.
[0086] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.
[0087] Continue to refer Figure 2 , shows a flow chart of an embodiment of the data retrieval method according to the present application. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted. The data retrieval method provided in the embodiment of the present application can be applied to any scenario where data retrieval is required, and the data retrieval method can be applied to products in these scenarios, for example, data retrieval in the financial insurance field. The data retrieval method comprises the following steps:
[0088] Step S201, receiving a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data.
[0089] In this embodiment, the data retrieval method is executed on the electronic device (eg Figure 1 The server / terminal device shown in the figure) can obtain the data to be queried through a wired connection or a wireless connection. It should be noted that the above-mentioned wireless connection method may include but is not limited to 3G / 4G / 5G connection, Wi-Fi connection, Bluetooth connection, Wi MAX connection, Zigbee connection, UWB (Ultra Wi DeBand) connection, and other wireless connection methods currently known or to be developed in the future. The above-mentioned cross-modal financial database is a database for effective retrieval and analysis of different modal data in the cross-modal financial database based on cross-modal retrieval technology (DCMH, Deep Cross-Modal Hash ing) and hash coding technology. The construction process of the cross-modal financial database includes: data collection and preprocessing: financial data in the financial field is collected through web crawlers, API interfaces, etc., including multi-modal data of types such as text data, image data, and audio data. The collected data is pre-processed, including data cleaning, format conversion, word segmentation, and other operations. Data storage and indexing: The pre-processed multi-modal data is stored in the financial database, and a corresponding index structure is established. According to the data type, a suitable database storage solution is adopted, such as relational database, graph database, etc. Cross-modal retrieval financial database, or cross-modal retrieval model: In the training process of the cross-modal retrieval model, image data is extracted through the ResNet38 network to obtain image features, and text data is extracted through the BERT model to obtain text features. In the process of training the model, hash representation is used for cross-modal retrieval, that is, the image features and text features obtained by using a predetermined target hash function are mapped to the same hash space, and a corresponding hash table is established to store hash values and corresponding database entries. The key of the hash table is the hash value, and the value is the corresponding database entry or index. Specifically, contrastive learning loss is used for model training to bring the potential representation of paired text and images closer, and to pull unpaired text and images farther apart. The cosine distance between data pairs is adaptively reduced and enlarged, the feature representation of the same sample pair is brought closer, and the distance between different sample pairs is pulled farther apart, so that the semantic information of different modalities can be enhanced. At the same time, alignment loss is used for model training, that is, the potential representation of paired data of different modalities should be consistent, thereby eliminating the heterogeneity between data of different modalities. Finally, a trained cross-modal retrieval model that meets the needs is obtained, that is, the construction of a cross-modal financial database is completed. Among them, in the business scenario of financial insurance data retrieval, financial data may include transaction data, payment data, business data, etc.
[0090] Step S202: extracting target features of the data to be queried based on a preset feature extraction network.
[0091] In this embodiment, if the data to be queried is image data, the corresponding feature extraction network may adopt a ResNet38 network. If the data to be queried is text data, the corresponding feature extraction network may adopt a BERT model.
[0092] Step S203: Obtain a target hash function corresponding to the cross-modal financial database.
[0093] In this embodiment, the method for constructing the above-mentioned target hash function will be further described in detail in subsequent specific embodiments, and will not be elaborated on here.
[0094] Step S204: performing hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value.
[0095] In this embodiment, the target hash function is used to perform hash value calculation processing on the target feature to generate a target hash value corresponding to the target feature.
[0096] Step S205, querying a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and querying a second hash value that meets a preset similarity condition with the target hash value from the hash table.
[0097] In this embodiment, the target hash value is used to perform data matching on the hash table corresponding to the cross-modal financial database, so as to query the first hash value that is the same as the target hash value from the hash table. The specific implementation process of querying the second hash value that meets the preset similarity condition with the target hash value from the hash table will be further described in detail in the subsequent specific embodiments of this application, and will not be elaborated on here.
[0098] Step S206: query the cross-modal financial database for second target data that matches the second hash value, and determine third target data from the second target data according to a preset screening rule.
[0099] In this embodiment, the cross-modal financial database is queried by using the second hash value to query the second target data matching the second hash value from the cross-modal financial database. The second target data has the same target hash value as the above-mentioned data to be queried. In addition, the specific implementation process of determining the third target data from the second target data according to the preset screening rule will be further described in detail in the subsequent specific embodiments of this application, and will not be elaborated on here.
[0100] Step S207: query the cross-modal financial database for first target data that matches the first hash value, and return the first target data and the third target data to the user.
[0101] In this embodiment, the cross-modal financial database can be queried by using the first hash value to query the first target data matching the first hash value from the cross-modal financial database. Among them, through the use of the above-mentioned cross-modal financial database, it is possible to use one modality (such as image data) to retrieve semantically related data queries in another modality (such as text data). In addition, the specific implementation process of returning the first target data and the third target data to the user will be further described in detail in the subsequent specific embodiments of this application, and will not be elaborated on here. In addition, if the data to be queried is image type data, the corresponding output first target data and third target data are text type data. If the data to be queried is text type data, the corresponding output first target data and third target data are image type data.
[0102] The present application uses a cross-modal financial database constructed based on a target hash function and a hash table to perform data retrieval processing on a target feature corresponding to the data to be queried contained in a data query request triggered by a user, so that the first target data and the third target data matching the target feature can be quickly and accurately queried and returned to the user to complete the data retrieval processing. Based on the cross-modal financial database, the user's multi-modal retrieval needs are effectively met, the intelligence of the data retrieval processing is improved, and the accuracy of the data retrieval processing is guaranteed.
[0103] In some optional implementations, before step S203, the electronic device may further perform the following steps:
[0104] Get a pre-collected test dataset.
[0105] In this embodiment, the test data set is feature data collected in advance for performing performance tests on candidate hash functions to evaluate the performance of different hash functions in terms of calculation speed, collision rate, retrieval accuracy, etc.
[0106] Get a preset number of candidate hash functions.
[0107] In this embodiment, there is no specific limitation on the selection of the above-mentioned preset number, and it can be set according to actual business usage requirements. Specifically, after obtaining multiple initial hash functions, the initial hash functions can be screened according to uniformity, that is, a hash function that can evenly map the input data to the hash value range is selected from multiple initial hash functions and used as a subsequent hash function. The above-mentioned initial hash functions may include at least MD5, SHA-1, SHA-2, bcrypt, Scrypt, SHA-3, and HMAC. By preliminarily screening the initial hash functions to obtain candidate hash functions, it helps to reduce hash conflicts and improve retrieval efficiency, and subsequently only the candidate hash functions need to be tested instead of all the initial hash functions, which can effectively reduce the workload of test processing, thereby improving the generation efficiency of the target hash function.
[0108] The candidate hash functions are tested based on the test data set, and the computational efficiency, collision rate and retrieval accuracy of the candidate hash functions are calculated respectively.
[0109] In this embodiment, each candidate hash function is tested using the test data set to evaluate the performance data of each candidate hash function in terms of calculation speed, collision rate and retrieval accuracy, and then the calculation efficiency, collision rate and retrieval accuracy value of each candidate hash function on the test data are calculated.
[0110] Based on the computational efficiency, collision rate and retrieval precision value of each candidate hash function, a processing score of each candidate hash function is calculated.
[0111] In this embodiment, the specific implementation process of calculating the processing score of each candidate hash function based on the computational efficiency, collision rate and retrieval accuracy value of each candidate hash function will be further described in detail in subsequent specific embodiments of the present application and will not be elaborated here.
[0112] A hash function with the highest processing score is selected from all the candidate hash functions and used as the target hash function.
[0113] In this embodiment, the target processing score with the highest value is screened out from all the calculated processing scores, and then a hash function matching the target processing score is screened out from all the candidate hash functions and used as the target hash function.
[0114] The present application obtains a pre-collected test data set; then obtains a preset number of candidate hash functions; then tests and processes each of the candidate hash functions based on the test data set, and respectively calculates the computational efficiency, collision rate and retrieval precision value of each of the candidate hash functions; subsequently, based on the computational efficiency, collision rate and retrieval precision value of each of the candidate hash functions, calculates the processing score of each of the candidate hash functions; finally, the hash function with the highest processing score is screened out from all the candidate hash functions and used as the target hash function. The present application tests and evaluates the candidate hash functions by using the test data set, and calculates the processing score of each of the candidate hash functions according to the computational efficiency, collision rate and retrieval precision value obtained, and then screens out the hash function with the highest processing score from the candidate hash functions according to the obtained test results and uses it as the target hash function. Since the target hash function is the hash function with the best comprehensive score combining the computational efficiency, collision rate and retrieval precision value, the subsequent use of the target hash function to apply to the computational processing of the cross-modal financial database can effectively improve the computational performance, and effectively improve the accuracy of the obtained target hash function.
[0115] In some optional implementations of this embodiment, the processing score of each candidate hash function is calculated based on the computational efficiency, collision rate and retrieval precision value of each candidate hash function, including the following steps:
[0116] Determines the preset weight setting algorithm.
[0117] In this embodiment, there is no specific limitation on the selection of the weight setting algorithm. For example, any one of the various algorithms such as hierarchical analysis method, entropy method, independence weight method, information weight method, etc. can be used, or two algorithms can be randomly selected from the various algorithms and combined as the above-mentioned weight setting algorithm.
[0118] Based on the weight setting algorithm, a first weight corresponding to the calculation efficiency is determined, a second weight corresponding to the collision rate is determined, and a third weight corresponding to the retrieval precision value is determined.
[0119] In this embodiment, the weight setting algorithm can be used to perform weight generation processing on the computing efficiency, collision rate and retrieval accuracy value respectively to generate a first weight corresponding to the computing efficiency, a second weight corresponding to the collision rate, and a third weight corresponding to the retrieval accuracy value.
[0120] Call the preset processing score statistics formula.
[0121] In this embodiment, the above processing score statistical formula specifically includes: CS=u*d+v*(1 / e)+w*f, CS is the processing score, d is the calculation efficiency, u is the first weight, e is the collision rate, v is the second weight, f is the retrieval accuracy value, and w is the third weight.
[0122] Based on the processing score statistical formula, the first weight, the second weight, the third weight, and the computational efficiency, collision rate and retrieval precision value of each candidate hash function are calculated and processed to obtain a processing score of each candidate hash function.
[0123] In this embodiment, the processing score of the corresponding candidate hash function can be obtained by substituting the first weight, the second weight, the third weight and the computational efficiency, collision rate and retrieval accuracy value of the candidate hash function into the corresponding positions of the processing score statistical formula for calculation.
[0124] The present application determines a preset weight setting algorithm; then determines a first weight corresponding to the computational efficiency, a second weight corresponding to the collision rate, and a third weight corresponding to the retrieval precision value based on the weight setting algorithm; then calls a preset processing score statistical formula; subsequently, based on the processing score statistical formula, the first weight, the second weight, the third weight, and the computational efficiency, collision rate and retrieval precision value of each candidate hash function are calculated and processed to obtain the processing score of each candidate hash function. The present application can effectively ensure the accuracy and rationality of the generated first weight corresponding to the computational efficiency, the second weight corresponding to the collision rate, and the third weight corresponding to the retrieval precision value by using the weight setting algorithm. Furthermore, on the basis of the obtained first weight, second weight and third weight, by using the processing score statistical formula to calculate the computational efficiency, collision rate and retrieval precision value of each candidate hash function, the processing score of each candidate hash function can be generated quickly and accurately, thereby improving the generation efficiency of the processing score and ensuring the data accuracy of the generated processing score.
[0125] In some optional implementations, the step of searching the hash table for a second hash value that meets a preset similarity condition with the target hash value in step S205 includes the following steps:
[0126] A target bucket corresponding to the target hash value is determined from the hash table.
[0127] In this embodiment, when constructing a hash table corresponding to the cross-modal financial database, the hash table is expanded into multiple buckets in advance, and each bucket stores a hash value within a certain range. The target bucket refers to the bucket storing the target hash value.
[0128] The target hash value is respectively compared with all the specified hash values contained in the target bucket to obtain corresponding similarity comparison results.
[0129] In this embodiment, the target hash value can be compared with all the specified hash values contained in the target bucket by using a similarity comparison method or a local sensitive hashing method, thereby obtaining a corresponding similarity comparison result. Among them, the local sensitive hashing method can be preferably used for similarity comparison processing. Local sensitive hashing (LSH) is an algorithm for approximate search of high-dimensional data. It can map similar elements in the original data space to similar positions in the hash table, thereby achieving efficient approximate search.
[0130] Based on the similarity comparison result, a third Hash value whose similarity comparison result is greater than a preset similarity comparison threshold is screened out from all the specified Hash values.
[0131] In this embodiment, there is no specific limitation on the value of the similarity comparison threshold, which can be set according to actual business usage requirements.
[0132] The third Hash value is used as the second Hash value.
[0133] The present application determines the target bucket corresponding to the target hash value from the hash table; then performs similarity comparison processing on the target hash value and all the specified hash values contained in the target bucket to obtain the corresponding similarity comparison result; then based on the similarity comparison result, a third hash value whose similarity comparison result is greater than a preset similarity comparison threshold is screened out from all the specified hash values; and the third hash value is subsequently used as the second hash value. The present application performs similarity comparison processing on the target hash value and all the specified hash values contained in the target bucket where the target hash value is located, and according to the obtained similarity comparison result, the second hash value that meets the preset similarity conditions with the target hash value is quickly and accurately screened out from the specified hash values, thereby ensuring the data accuracy of the obtained second hash value.
[0134] In some optional implementations, the step of determining the third target data from the second target data according to a preset screening rule in step S206 includes the following steps:
[0135] Feature extraction is performed on each of the second target data to obtain a plurality of specified features.
[0136] In this embodiment, the feature extraction network mentioned above can be used to extract features from each of the second target data, thereby obtaining a corresponding plurality of specified features.
[0137] Each of the designated features is mapped to a preset common subspace to obtain a plurality of first data, and the target feature is mapped to the common subspace to obtain second data.
[0138] In this embodiment, the common subspace refers to mapping features of different modalities (such as text, image, audio, etc.) to a common subspace. This common subspace is a special space with a dimension less than or equal to the original feature space, which allows comparison of data from different modalities.
[0139] Based on a preset similarity measurement algorithm, the similarity between each of the first data and the second data is calculated respectively.
[0140] In this embodiment, there is no limitation on the selection of the above similarity measurement algorithm. For example, any one of the algorithms such as Euclidean distance, cosine similarity, and Pearson correlation coefficient may be used.
[0141] A designated similarity greater than a preset similarity threshold is screened out from all the similarities.
[0142] In this embodiment, there is no specific limitation on the value of the similarity threshold, which can be set according to actual usage requirements, for example, it can be set to 0.9.
[0143] The third data corresponding to the specified similarity is acquired from all the first data.
[0144] In this embodiment, the third data refers to data in all the first data that have a corresponding relationship with the specified similarity.
[0145] The matching data corresponding to the third data is screened out from the second target data to obtain the third target data.
[0146] In this embodiment, the third target data refers to data in the second target data that has a matching relationship with the third data.
[0147] The present application obtains a plurality of designated features by extracting features from each of the second target data; then maps each of the designated features to a preset common subspace to obtain a plurality of first data, and maps the target features to the common subspace to obtain the second data; then, based on a preset similarity measurement algorithm, respectively calculates the similarity between each of the first data and the second data; subsequently, selects a designated similarity greater than a preset similarity threshold from all the similarities; further obtains the third data corresponding to the designated similarity from all the first data; and finally selects the matching data corresponding to the third data from the second target data to obtain the third target data. After extracting features from each of the second target data to obtain a plurality of designated features, the present application obtains the first data and the second data by mapping the designated features and the target features based on the use of a common subspace. Then, based on the use of a similarity measurement algorithm and a similarity threshold, the final third target data is quickly and accurately selected from the second target data, thereby ensuring the data accuracy of the obtained third target data.
[0148] In some optional implementations of this embodiment, the returning the first target data and the third target data to the user in step S207 includes the following steps:
[0149] Sorting information of the first target data and the third target data is generated.
[0150] In this embodiment, the content of the above-mentioned sorting information includes: sorting the sorting information of the first target data as sorting first, then obtaining the similarities respectively corresponding to the target features of each third target data and the target data to be queried, and then sorting all the third target data in descending order according to the similarities of each third target data, and sequentially obtaining the sorting information of each third target data. Specifically, the sorting information of the data with the highest similarity among all the third target data is sorting second, the sorting information of the data with the second highest similarity among all the third target data is sorting third, and so on, until the generation of sorting information for all the third target data is completed.
[0151] Call the preset display list.
[0152] In this embodiment, the above display list is a pre-built list template for feeding back data query requests of relevant users. The specific content structure of the display list is not limited and can be set according to actual data feedback requirements.
[0153] Based on the sorting information, the first target data and the third target data are filled into the display list to obtain a target display list.
[0154] In this embodiment, the numbers corresponding to the first target data and the third target data can be determined from the display list according to the sorting information, and then the first target data and the third target data can be filled into the display list accordingly according to the numbers, thereby obtaining a target display list.
[0155] The target display list is returned to the user.
[0156] In this embodiment, by returning the first target data and the third target data that match the query data input by the user to the user in a sorted target display list, the user can query the data results most relevant to the query data from the target display list, which is beneficial to improving retrieval efficiency and reducing the time users spend browsing irrelevant information.
[0157] The present application generates the sorting information of the first target data and the third target data; then calls the preset display list; then based on the sorting information, fills the first target data and the third target data into the display list to obtain the target display list; and subsequently returns the target display list to the user. The present application generates the sorting information of the first target data and the third target data, and intelligently uses the sorting information to generate a target display list corresponding to the first target data and the third target data, and then returns the target display list to the user, so that the user can query the data results most relevant to the data to be queried from the target display list, which is conducive to improving retrieval efficiency and reducing the time users spend browsing irrelevant information.
[0158] In some optional implementations of this embodiment, after step S205, the electronic device may further perform the following steps:
[0159] Get the number of elements in the hash table.
[0160] In this embodiment, the number of elements in the hash table may be calculated by counting all elements stored in the hash table.
[0161] Obtain a capacity threshold corresponding to the hash table.
[0162] In this embodiment, the capacity threshold refers to a value corresponding to the maximum element capacity that the hash table can bear.
[0163] The difference between the capacity threshold and the number of elements is calculated.
[0164] In this embodiment, the capacity threshold and the number of elements are subjected to a difference process to obtain the difference between the two.
[0165] Determine whether the difference is less than a preset value.
[0166] In this embodiment, there is no specific limitation on the value of the preset value, which can be set according to the actual structural design requirements of the hash table. The content of the structural design requirements may include: if the number of elements in the hash table is close to the capacity of the hash table, the problem of too high conflict rate may occur.
[0167] If so, the hash table is expanded based on a preset expansion base.
[0168] In this embodiment, the selection of the above expansion base is not specifically limited, for example, 1.5, 2, etc. can be used. After the expansion of the hash table is completed using the expansion base, the original elements in the hash table will be further remapped to the new expanded hash table. In addition, when the data in the cross-modal financial database changes, the hash table and the corresponding hash value will be automatically updated to ensure the accuracy of the retrieval results.
[0169] The present application obtains the number of elements in the hash table; then obtains the capacity threshold corresponding to the hash table; then calculates the difference between the capacity threshold and the number of elements; subsequently determines whether the difference is less than a preset value; if so, expands the hash table based on a preset expansion base. The present application intelligently compares the number of elements in the hash table with the capacity threshold corresponding to the hash table, and when it detects that the difference between the capacity threshold and the number of elements is less than a preset value, it automatically expands the hash table based on a preset expansion base, thereby effectively avoiding the problem of excessively high conflict rates in the hash table, improving the intelligence of the structural maintenance of the hash table, and helping to ensure that the hash table has better performance operation.
[0170] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.
[0171] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned first target data, the above-mentioned first target data can also be stored in a node of a blockchain.
[0172] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.
[0173] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
[0174] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.
[0175] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).
[0176] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.
[0177] Further references Figure 3 , as a response to the above Figure 2 In order to realize the method shown in the figure, the present application provides an embodiment of a data retrieval device, and the embodiment of the device is Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.
[0178] like Figure 3 As shown, the data retrieval device 300 described in this embodiment includes: a receiving module 301, an extraction module 302, a first acquisition module 303, a first calculation module 304, a query module 305, a first processing module 306 and a second processing module 307. Among them:
[0179] The receiving module 301 is used to receive a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data;
[0180] An extraction module 302 is used to extract target features of the data to be queried based on a preset feature extraction network;
[0181] A first acquisition module 303, configured to acquire a target hash function corresponding to the cross-modal financial database;
[0182] A first calculation module 304 is used to perform hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value;
[0183] A query module 305 is used to query a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and query a second hash value that meets a preset similarity condition with the target hash value from the hash table;
[0184] A first processing module 306, configured to query the cross-modal financial database for second target data matching the second hash value, and determine third target data from the second target data according to a preset screening rule;
[0185] The second processing module 307 is configured to query the cross-modal financial database for first target data matching the first hash value, and return the first target data and the third target data to the user.
[0186] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the data retrieval method in the aforementioned implementation mode, and will not be described in detail here.
[0187] In some optional implementations of this embodiment, the data retrieval device further includes:
[0188] A second acquisition module is used to acquire a pre-collected test data set;
[0189] A third acquisition module is used to acquire a preset number of candidate hash functions;
[0190] A testing module, used to test each of the candidate hash functions based on the test data set, and respectively calculate the computational efficiency, collision rate and retrieval accuracy value of each of the candidate hash functions;
[0191] A second calculation module, used to calculate the processing score of each candidate hash function based on the calculation efficiency, collision rate and retrieval precision value of each candidate hash function;
[0192] The screening module is used to screen out the hash function with the highest processing score from all the candidate hash functions and use it as the target hash function.
[0193] In some optional implementations of this embodiment, the second calculation module includes:
[0194] A first determination submodule, used to determine a preset weight setting algorithm;
[0195] A second determination submodule, configured to determine, based on the weight setting algorithm, a first weight corresponding to the computational efficiency, a second weight corresponding to the collision rate, and a third weight corresponding to the retrieval precision value;
[0196] The first calling submodule is used to call a preset processing score statistical formula;
[0197] The first calculation submodule is used to calculate the first weight, the second weight, the third weight, and the calculation efficiency, collision rate and retrieval accuracy value of each candidate hash function based on the processing score statistical formula to obtain the processing score of each candidate hash function.
[0198] In some optional implementations of this embodiment, the query module 305 includes:
[0199] A third determination submodule, configured to determine a target bucket corresponding to the target hash value from the hash table;
[0200] A comparison submodule, used to perform similarity comparison processing on the target hash value and all the specified hash values contained in the target bucket respectively, to obtain corresponding similarity comparison results;
[0201] A first screening submodule, configured to screen out third hash values having similarity comparison results greater than a preset similarity comparison threshold from all the specified hash values based on the similarity comparison result;
[0202] The fourth determining submodule is configured to use the third Hash value as the second Hash value.
[0203] In some optional implementations of this embodiment, the first processing module 306 includes:
[0204] An extraction submodule, used for performing feature extraction on each of the second target data to obtain a plurality of specified features;
[0205] A mapping submodule, used for mapping each of the specified features to a preset common subspace to obtain a plurality of first data, and mapping the target feature to the common subspace to obtain second data;
[0206] A second calculation submodule, used for respectively calculating the similarity between each of the first data and the second data based on a preset similarity measurement algorithm;
[0207] A second screening submodule, used for screening out a specified similarity greater than a preset similarity threshold from all the similarities;
[0208] An acquisition submodule, used for acquiring third data corresponding to the specified similarity from all the first data;
[0209] The third screening submodule is used to screen out matching data corresponding to the third data from the second target data to obtain the third target data.
[0210] In some optional implementations of this embodiment, the second processing module 307 includes:
[0211] A generating submodule, used for generating sorting information of the first target data and the third target data;
[0212] The second calling submodule is used to call a preset display list;
[0213] a filling submodule, configured to fill the first target data and the third target data into the display list based on the sorting information to obtain a target display list;
[0214] The return submodule is used to return the target display list to the user.
[0215] In some optional implementations of this embodiment, the data retrieval device further includes:
[0216] A fourth acquisition module, used to obtain the number of elements in the hash table;
[0217] A fifth acquisition module, used to acquire a capacity threshold corresponding to the hash table;
[0218] A third calculation module, used to calculate the difference between the capacity threshold and the number of elements;
[0219] A judging module, used to judge whether the difference is less than a preset value;
[0220] An expansion module is used to expand the hash table based on a preset expansion base.
[0221] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.
[0222] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital processors (DSPs), embedded devices, etc.
[0223] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.
[0224] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Med ia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of a data retrieval method, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.
[0225] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run the computer-readable instructions stored in the memory 41 or process data, such as the computer-readable instructions for running the data retrieval method.
[0226] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.
[0227] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the data retrieval method as described above.
[0228] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.
[0229] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.
Claims
1. A data retrieval method, characterized in that: The steps include: Receiving a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data; Extracting target features of the data to be queried based on a preset feature extraction network; Obtaining a target hash function corresponding to the cross-modal financial database; Performing hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value; Querying a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and querying a second hash value that meets a preset similarity condition with the target hash value from the hash table; Querying the second target data matching the second hash value from the cross-modal financial database, and determining the third target data from the second target data according to a preset screening rule; First target data matching the first hash value is queried from the cross-modal financial database, and the first target data and the third target data are returned to the user.
2. The data retrieval method according to claim 1, characterized in that: Before the step of obtaining the target hash function corresponding to the cross-modal financial database, the method further includes: Obtain a pre-collected test dataset; Obtain a preset number of candidate hash functions; Performing a test process on each of the candidate hash functions based on the test data set, and calculating the computational efficiency, collision rate and retrieval precision value of each of the candidate hash functions respectively; Calculating a processing score of each candidate hash function based on the computational efficiency, collision rate, and retrieval precision value of each candidate hash function; A hash function with the highest processing score is selected from all the candidate hash functions and used as the target hash function.
3. The data retrieval method according to claim 2, characterized in that: The step of calculating the processing score of each candidate hash function based on the computational efficiency, collision rate and retrieval precision value of each candidate hash function specifically includes: Determine a preset weight setting algorithm; Determine a first weight corresponding to the computational efficiency, determine a second weight corresponding to the collision rate, and determine a third weight corresponding to the retrieval precision value based on the weight setting algorithm; Call the preset processing score statistics formula; Based on the processing score statistical formula, the first weight, the second weight, the third weight, and the computational efficiency, collision rate and retrieval precision value of each candidate hash function are calculated and processed to obtain a processing score of each candidate hash function.
4. The data retrieval method according to claim 1, characterized in that: The step of searching the hash table for a second hash value that meets a preset similarity condition with the target hash value specifically includes: Determine a target bucket corresponding to the target hash value from the hash table; Performing similarity comparison processing on the target hash value and all the specified hash values contained in the target bucket respectively to obtain corresponding similarity comparison results; Based on the similarity comparison result, a third hash value having a similarity comparison result greater than a preset similarity comparison threshold is screened out from all the specified hash values; The third Hash value is used as the second Hash value.
5. The data retrieval method according to claim 1, characterized in that: The step of determining the third target data from the second target data according to the preset screening rule specifically includes: Performing feature extraction on each of the second target data to obtain a plurality of specified features; Mapping each of the designated features to a preset common subspace to obtain a plurality of first data, and mapping the target feature to the common subspace to obtain second data; Based on a preset similarity measurement algorithm, respectively calculate the similarity between each of the first data and the second data; Filtering out a specified similarity greater than a preset similarity threshold from all the similarities; Acquire third data corresponding to the specified similarity from all the first data; The matching data corresponding to the third data is screened out from the second target data to obtain the third target data.
6. The data retrieval method according to claim 1, characterized in that: The step of returning the first target data and the third target data to the user specifically includes: generating sorting information of the first target data and the third target data; Call the preset display list; Based on the sorting information, filling the first target data and the third target data into the display list to obtain a target display list; The target display list is returned to the user.
7. The data retrieval method according to claim 1, characterized in that: After the step of searching the hash table for a second hash value that meets a preset condition of being close to the target hash value, the method further includes: Get the number of elements in the hash table; Obtaining a capacity threshold corresponding to the hash table; Calculating the difference between the capacity threshold and the number of elements; Determine whether the difference is less than a preset value; If so, the hash table is expanded based on a preset expansion base.
8. A data retrieval device, characterized in that: include: A receiving module, configured to receive a data query request for a pre-built cross-modal financial database triggered by a user; wherein the data query request carries data to be queried, and the data to be queried includes image data / text data; An extraction module, used for extracting target features of the data to be queried based on a preset feature extraction network; A first acquisition module, configured to acquire a target hash function corresponding to the cross-modal financial database; A first calculation module, configured to perform hash value calculation processing on the target feature based on the target hash function to obtain a corresponding target hash value; A query module, configured to query a first hash value that is the same as the target hash value from a hash table corresponding to the cross-modal financial database, and query a second hash value that meets a preset similarity condition with the target hash value from the hash table; A first processing module, configured to query the cross-modal financial database for second target data matching the second hash value, and determine third target data from the second target data according to a preset screening rule; The second processing module is configured to query the first target data matching the first hash value from the cross-modal financial database, and return the first target data and the third target data to the user.
9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the data retrieval method according to any one of claims 1 to 7 when executing the computer-readable instructions.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the data retrieval method according to any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Hash search method for reducing hash collision
CN104504038A
Text retrieval method and device, equipment and program product
CN109684442A
Method, system and device for storing digital assets based on block chain and storage medium
CN113159952A
Data retrieval method and device based on privacy protection, computer equipment and medium
CN115017107A
Cited By
Target user retrieval method and device, medium and equipment
CN120561148A
