Database interaction recommendation optimization method and device, equipment and storage medium

Abstract

The invention discloses an optimization method and device for database interactive recommendation, equipment and a storage medium, and relates to the technical field of computers, the method comprises the following steps: obtaining a first operation of a user and a candidate recommendation operation set corresponding to the first operation; acquiring a second operation actually selected by the user; determining the accuracy rate of the user according to the second operation, determining the continuity of the user according to the second operation and the first operation, and determining the coverage rate of the user according to the candidate recommendation operation set; determining a feedback optimization coefficient of the user according to the accuracy rate, the coverage rate and the continuity; and optimizing the candidate recommendation operation set according to the feedback optimization coefficient and the process information of selecting the second operation by the user. According to the method, the recommended content can be optimized more accurately.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an optimization method, apparatus, device, and storage medium for database interactive recommendation. Background Art

[0002] In database interaction scenarios, after a user performs the first operation (such as a basic query), there are often multiple reasonable options for subsequent operations. For example, adding location conditions to accurately filter data, sorting by time fields to organize the result set, and directly exporting data for analysis. To this end, the system will recommend a candidate to the user from multiple options.

[0003] The traditional solution for optimizing recommendations is that when the system recommends candidate 1 to the user, the user chooses candidate 2, thereby reducing the probability of recommending candidate 1 and increasing the probability of recommending candidate 2. This optimization solution for recommended content has many drawbacks and cannot accurately optimize the recommended content. Summary of the Invention

[0004] The present application provides a database interactive recommendation optimization method, apparatus, device and storage medium, which can optimize the recommended content more accurately.

[0005] To achieve the above objectives, this application adopts the following technical solutions: In a first aspect, the present application provides a method for optimizing database interactive recommendation, the method comprising: Obtaining a first operation of the user and a set of candidate recommended operations corresponding to the first operation; obtaining a second operation actually selected by the user; determining the user's accuracy based on the second operation, determining the user's consistency based on the second operation and the first operation, and determining the user's coverage based on the candidate recommended operation set; determining a feedback optimization coefficient for the user according to the accuracy, the coverage, and the consistency; The candidate recommended operation set is optimized according to the feedback optimization coefficient and process information of the user selecting the second operation.

[0006] Optionally, after obtaining the second operation actually selected by the user, the method further includes: Obtaining a third operation selected by the user; Calculating similarity between the third operation and all candidate operations within a preset time interval; If there is a target candidate operation with a similarity greater than or equal to the similarity threshold among all the candidate operations, the recommendation probability of the target candidate operation is increased and optimized according to the target similarity between the target candidate operation and the third operation.

[0007] Optionally, optimizing the candidate recommended operation set according to the feedback optimization coefficient and the process information of the user selecting the second operation includes: The initial learning rate is updated according to the feedback optimization coefficient to obtain a dynamic learning rate; determining, based on process information of the user selecting the second operation, an initial thermal index of the user for a candidate operation in the set of candidate recommended operations; and updating the initial thermal index based on the sensitivity coefficient to obtain a dynamic thermal index; Obtaining initial recommendation probabilities of candidate operations in the candidate recommended operation set; The initial recommendation probability is updated according to the dynamic learning rate and the dynamic heat index to obtain the dynamic recommendation probability corresponding to the candidate operation.

[0008] Optionally, the increasing and optimizing the recommendation probability of the target candidate operation according to the target similarity between the target candidate operation and the third operation includes: determining a target relevance between the target candidate operation and the third operation based on a historical recommendation probability of the target candidate operation, a target similarity between the target candidate operation and the third operation, and a target time interval between the third operation and the target candidate operation; According to the target relevance, the recommendation probability of the target candidate operation is increased and optimized.

[0009] Optionally, the method further includes: Obtain the target role tag of the user; Determine a target weight coefficient corresponding to the target role label according to a pre-configured mapping relationship between the reference role label and the reference weight coefficient, wherein the target weight coefficient includes a target accuracy coefficient, a target coverage coefficient, and a target coherence coefficient; The determining of the user feedback optimization coefficient according to the accuracy, the coverage, and the coherence includes: A feedback optimization coefficient for the user is determined according to the accuracy, the target accuracy coefficient, the coverage, the target coverage coefficient, the coherence, and the target coherence coefficient.

[0010] Optionally, the precision is determined by the hit status of the second operation actually selected by the user in the set of candidate recommended operations corresponding to the first operation; the coverage is obtained by the reasonable candidate operations included in the set of candidate recommended operations corresponding to the first operation and all reasonable candidate operations corresponding to the first operation; the coherence is obtained by the correlation between the second operation and the first operation.

[0011] Optionally, the process information of the user selecting the second operation includes: the time the control cursor stays on each candidate operation and the position of each candidate operation during the process of the user selecting the second operation.

[0012] In a second aspect, the present application provides an optimization device for database interactive recommendation, the device comprising: An acquisition module, configured to acquire a first operation of a user and a set of candidate recommended operations corresponding to the first operation; and acquire a second operation actually selected by the user; a determination module, configured to determine the user's precision rate based on the second operation, determine the user's consistency based on the second operation and the first operation, determine the user's coverage rate based on the candidate recommended operation set; and determine a feedback optimization coefficient for the user based on the precision rate, the coverage rate, and the consistency; An optimization module is configured to optimize the candidate recommended operation set according to the feedback optimization coefficient and process information of the user selecting the second operation.

[0013] In a third aspect, the present application provides a computing device, including a memory and a processor; One or more computer programs are stored in the memory, and the one or more computer programs include instructions; when the instructions are executed by the processor, the computing device executes the method as described in any one of the first aspects.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program for executing the method as described in any one of the first aspects.

[0015] It can be seen from the above technical solution that this application has at least the following beneficial effects: The present application provides an optimization method for database interactive recommendation. During the optimization process, the user's first operation and the candidate recommended operation set corresponding to the first operation are first obtained. Then, after the user makes an actual selection, the second operation actually selected by the user is obtained. Then, based on the second operation, the user's precision is determined. Based on the second operation and the first operation, the user's consistency is determined. Based on the candidate recommended operation set, the user's coverage is determined. Based on the precision, coverage, and consistency, the user's feedback optimization coefficient is determined. Finally, based on the feedback optimization coefficient and the process information of the user selecting the second operation, the candidate recommended operation set is optimized. This method does not only perform feedback optimization based on whether the user has selected an operation in the candidate recommended set, but also considers the process information of the user actually selecting the second operation, as well as the feedback optimization coefficient determined based on multiple angles, and the ranking of the candidate operations in the candidate recommended operation set, and comprehensively optimizes the candidate recommended operation set. In this way, the method can optimize the recommended content more accurately.

[0016] It should be understood that the description of technical features, technical solutions, beneficial effects or similar language in this application does not imply that all features and advantages can be realized in any single embodiment. On the contrary, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution or beneficial effect is included in at least one embodiment. Therefore, the description of a technical feature, technical solution or beneficial effect in this specification does not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions and beneficial effects described in the present embodiment can also be combined in any appropriate manner. Those skilled in the art will understand that the embodiment can be implemented without one or more specific technical features, technical solutions or beneficial effects of a specific embodiment. In other embodiments, additional technical features and beneficial effects can also be identified in specific embodiments that do not embody all embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of a database interactive recommendation optimization method provided in an embodiment of the present application; Figure 2 A schematic diagram of an optimization device for database interactive recommendation provided in an embodiment of the present application; Figure 3 A schematic diagram of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The terms "first", "second" and "third" in this application specification and the accompanying drawings are used to distinguish different objects rather than to limit a specific order.

[0019] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0020] In order to make the technical solution of this application clearer, the technical solution of this application is introduced below in combination with application scenarios.

[0021] Application scenario: A data analyst uses the database interface to query stored order information. After the user executes the SELECT * FROM orders operation, the system predicts that the user's next possible actions may include: 1. Adding a location condition to filter orders in a specific location (for example, filter orders in Asia, WHERE region = 'Asia'); 2. Querying time trends by sorting time (for example, order_date); 3. Calculating sales figures for various products (for example, GROUP BY); 4. Exporting data for report creation.

[0022] These predictions all align with business logic, but traditional precision metrics only determine whether the prediction "hit" the user's actual action (e.g., the user ultimately performed action 2), while ignoring the rationality of actions 1, 3, and 4. If evaluated solely by precision, the model might misjudge the effectiveness of actions 1, 2, and 4 because they missed them. However, these other predictions are also valuable; a single metric alone doesn't allow the system to consider these other values.

[0023] In view of this, an embodiment of the present application provides an optimization method for database interactive recommendation, which can be executed by a processing device, which can be a terminal or a server. Terminals include but are not limited to smartphones, tablets, laptops, personal digital assistants, or smart wearable devices. The server can be a cloud server, such as a central server in a central cloud computing cluster, or an edge server in an edge cloud computing cluster. Of course, the server can also be a server in a local data center. A local data center refers to a data center directly controlled by the user.

[0024] In this method, by adding coverage and coherence evaluation indicators, a multi-dimensional evaluation index of precision, coverage and coherence is obtained. Precision is used to directly measure the matching degree between the recommended content and the real needs of the user, coverage is used to avoid the recommended content being limited to the user's high-frequency operations, and coherence is used to ensure that the recommended content meets the requirements of the business scenario. Based on this information, the user's feedback optimization coefficient is determined. Then, the candidate recommended operation set is optimized using the feedback optimization coefficient combined with the process information of the user selecting the second operation and the first ranking of the candidate operations in the candidate recommended operation set. In this way, although the user's second operation is not a candidate operation ranked low in the candidate recommended operation set, if the process information represents that the user pays attention to the candidate operation ranked low, the recommendation probability of the candidate operation ranked low can be increased. Therefore, this solution is more tolerant of operations that the user has not performed but the user pays attention to. It does not directly reduce their recommendation probability, but increases their recommendation probability. It can be seen that the present application can optimize the recommended content more accurately.

[0025] In order to make the technical solution of the present application clearer and easier to understand, the technical solution of the present application is introduced below with reference to the accompanying drawings.

[0026] like Figure 1 As shown in FIG, this figure is a flowchart of an optimization method for database interactive recommendation provided by an embodiment of the present application, the method comprising: S201: The processing device obtains a first operation of a user and a set of candidate recommended operations corresponding to the first operation, and obtains a second operation actually selected by the user.

[0027] In some embodiments, the processing device can provide a human-computer interaction interface to the user, through which the user can control the database, for example, query the data stored in the database. The user can enter "SELECT * FROM orders" in the input area of the human-computer interaction interface to query orders. Among them, the operation of the user entering "SELECT * FROM orders" in the input area is the first operation. After the processing device receives the first operation input by the user, it can determine a set of candidate recommended operations corresponding to the first operation based on the first operation. The second operation refers to the operation actually selected by the user after the processing device displays the set of candidate recommended operations corresponding to the first operation to the user.

[0028] In some examples, the set of candidate recommended operations corresponding to the first operation includes: A. Add a WHERE condition (to filter orders placed during a specific time period); B. Use GROUP BY to calculate sales by product category; C. Join the products table to obtain product details; D. Export data to create a report. The second operation can be an operation from the candidate recommended operation set or another operation.

[0029] S202: The processing device determines the user's accuracy based on the second operation, determines the user's consistency based on the second operation and the first operation, and determines the user's coverage based on the candidate recommended operation set.

[0030] Among them, precision is used to characterize the hit situation of the operation actually selected by the user in the candidate recommended operation set; coherence is used to characterize the business relevance between the operation actually selected by the user and the previous operation; coverage is used to characterize the rationality of the candidate operations in the candidate recommended operation set.

[0031] The precision is determined by whether the second operation actually selected by the user hits the set of candidate recommended operations corresponding to the first operation; the coverage is obtained by the reasonable candidate operations included in the set of candidate recommended operations corresponding to the first operation and all reasonable candidate operations corresponding to the first operation; the consistency is obtained by the correlation between the second operation and the first operation.

[0032] In some examples, the processing device determines the user's accuracy rate based on the second operation, which can be implemented using the following formula:

[0033] in, represents the accuracy, Indicates the number of candidate operations in the candidate recommended operation set, Indicates the hit status of the second operation in the candidate recommended operation set. , when it misses .

[0034] In some examples, the processing device determines the user's consistency based on the second operation and the first operation, which can be achieved by the following formula:

[0035] in, Indicates continuity, Indicates the first operation, Indicates the second operation, express and The similarity between them.

[0036] In some examples, the processing device determines the user coverage rate based on the candidate recommended operation set, which can be implemented by the following formula:

[0037] in, Indicates coverage, Indicates the number of reasonable candidate operations included in the candidate recommended operation set, Indicates the number of all reasonable candidate operations corresponding to the first operation.

[0038] S203: The processing device determines the user feedback optimization coefficient based on the accuracy, coverage, and consistency.

[0039] After the processing device obtains the accuracy, coverage and consistency, it can determine the user feedback optimization coefficient based on the accuracy, coverage and consistency.

[0040] In some embodiments, different weighting methods can be set based on the user's role. For example, for a user with a role label of data analyst, the user pays more attention to the comprehensiveness of the data, so the weight of coverage will be higher, which can prompt the user to obtain more comprehensive data for subsequent processing. For example, the weights of accuracy, coverage and consistency are 0.2, 0.5 and 0.3 respectively; for a user with a role label of operation and maintenance engineer, the user pays more attention to the accuracy of the data, so the weight of accuracy will be higher, which can prompt the user to obtain more accurate data for subsequent processing. For example, the weights of accuracy, coverage and consistency are 0.5, 0.2 and 0.3 respectively.

[0041] The processing device may also obtain the user's target role label and then determine a target weight coefficient corresponding to the target role label based on a pre-configured mapping relationship between reference role labels and reference weight coefficients, where the target weight coefficients include a target accuracy coefficient, a target coverage coefficient, and a target coherence coefficient. The user's feedback optimization coefficient is then determined based on the accuracy, target accuracy coefficient, coverage, target coverage coefficient, coherence, and target coherence coefficient.

[0042] Specifically, the processing device can determine the feedback optimization coefficient by the following formula:

[0043] Where S represents the feedback optimization coefficient, represents the target accuracy coefficient, represents the target coverage coefficient, represents the target coherence coefficient.

[0044] S204: The processing device optimizes the candidate recommended operation set according to the feedback optimization coefficient and the process information of the user selecting the second operation.

[0045] The information about the user selecting the second operation includes the time the cursor remains on each candidate operation during the user selecting the second operation, as well as the position of each candidate operation. The position of each candidate operation can be represented by the row in which it is located, such as the first row, the tenth row, etc. The smaller the row number, the closer the candidate operation is to the front, that is, the more prominent it is, and vice versa.

[0046] The first ranking of the candidate operations in the candidate recommended operation set refers to the order in which the candidate operations are arranged in the candidate recommended operation set. The higher the recommendation probability of the candidate operation, the higher the ranking of the candidate operation in the candidate recommended operation set; the lower the recommendation probability of the candidate operation, the lower the ranking of the candidate operation in the candidate recommended operation set.

[0047] In some embodiments, the processing device may update the initial learning rate according to the feedback optimization coefficient to obtain a dynamic learning rate. Specifically, the processing device may implement the dynamic learning rate using the following formula:

[0048] in, represents the dynamic learning rate, Represents the initial learning rate, and S represents the feedback optimization coefficient. As can be seen from the formula, when the feedback optimization coefficient is low, that is, when the recommendation quality is poor, the learning rate is increased and the update is accelerated. When the feedback optimization coefficient is large, that is, when the recommendation quality is good, the learning rate is reduced and the update is slowed down. The embodiments of the present application can achieve dynamic adjustment of the optimization process by means of a dynamic learning rate.

[0049] In some embodiments, the processing device may determine the user's initial thermal index for a candidate operation in the candidate operation set based on the process information of the user selecting the second operation, and then update the initial thermal index based on the sensitivity coefficient to obtain a dynamic thermal index. Specifically, the processing device may implement this using the following formula:

[0050] in, represents the initial heat index of the k-th candidate operation, represents the user's hovering time on the kth candidate operation in the candidate recommended operation set, Indicates the location of the kth candidate operation in the candidate recommended operation set (indicated by the number of rows), It represents the average hovering time of users on each candidate action in the candidate recommended action set. Indicates the average position of each candidate operation in the candidate recommended operation set.

[0051] After the processing device obtains the initial thermal index, it can update the initial thermal index according to the sensitivity coefficient to obtain the dynamic thermal index. Specifically, the processing device can be implemented by the following formula:

[0052] in, represents the dynamic thermal index, represents the initial thermal index, represents the sensitivity coefficient, The coefficient representing the coverage rate of the user. For example, when the user's role label is data analyst, represents the user's coverage coefficient, that is, .

[0053] In some embodiments, the processing device obtains initial recommendation probabilities of candidate operations in the candidate operation set.

[0054] Then, the processing device can update the initial recommendation probability based on the dynamic learning rate and the dynamic heat index to obtain the dynamic recommendation probability corresponding to the candidate operation. The processing device can achieve this through the following formula:

[0055] in, represents the dynamic recommendation probability, represents the initial recommendation probability, represents the dynamic learning rate, Represents the dynamic thermal index.

[0056] For example, assuming , , , , , , ,therefore, , , (The probability increased by 2.3%).

[0057] In the above technical solution, when there is a candidate operation in the candidate recommended operation set that the user is interested in but has not selected, the recommendation probability of the candidate operation can be increased in the subsequent recommendation process, thereby avoiding the situation where the user does not recommend it next time because of not selecting it this time. This may cause the user to need to perform the candidate operation in the next operation, but the processing device does not recommend it, and the user still needs to find the candidate operation, thereby affecting business processing efficiency.

[0058] In the embodiments of this application, the processing device uses dynamic parameters (such as a dynamic learning rate and a dynamic heat index) to dynamically optimize the initial recommendation probability during the optimization process, thereby obtaining a dynamic recommendation probability. This optimization method dynamically optimizes the recommendation probability from multiple dimensions, rather than statically optimizing it, thereby further improving the accuracy of the recommended content and making the recommended content more compatible with the user.

[0059] In some embodiments, the processing device may also obtain the third operation selected by the user after obtaining the second operation actually selected by the user, and then calculate the similarity between the third operation and all candidate operations within a preset time interval. If there is a target candidate operation among all candidate operations whose similarity is greater than or equal to the similarity threshold, the recommendation probability of the target candidate operation is increased and optimized based on the target similarity between the target candidate operation and the first operation.

[0060] For example, the third operation refers to the operation performed by the user after the second operation. The preset time interval can be 30 minutes. The time when the candidate recommended operation set corresponding to the first operation is recommended to the user is 9:00, the preset time interval is 30 minutes, and the time when the third operation is performed is 10:00. Therefore, all candidate operations within the preset time interval include the candidate operations recommended between 9:30 and 10:00. The similarity threshold can be 70%. The processing device needs to calculate the similarity between the third operation and the candidate operations recommended between 9:30 and 10:00, then find the target candidate operation with a similarity greater than the similarity threshold, and then optimize the recommendation probability of the target candidate operation.

[0061] In the above technical solution, the association between delayed operations and historical recommended operations is identified to avoid recommendation failures due to misjudgment of time intervals, thereby increasing the subsequent recommendation probability of the target candidate operation by 15%.

[0062] In some embodiments, the processing device determines the target correlation between the target candidate operation and the third operation based on the historical recommendation probability of the target candidate operation, the target similarity between the target candidate operation and the third operation, and the target time interval between the third operation and the target candidate operation, and then increases and optimizes the recommendation probability of the target candidate operation based on the target correlation. Specifically, the processing device can achieve this through the following formula:

[0063]

[0064] in, represents the target relevance of the target candidate operation to the third operation, represents the target similarity between the target candidate operation and the third operation, represents the historical recommendation probability of the target candidate operation, Indicates the target time interval, such as the actual time interval, in hours; represents the increased recommendation probability of the target candidate operation, represents the dynamic learning rate.

[0065] The above technical solution uses semantic understanding to connect seemingly unrelated user actions, thereby more accurately predicting actual needs. A time decay mechanism prevents the system from over-reliance on outdated data (such as recommendations from 30 minutes ago), ensuring that recommendation priorities always align with the user's current needs. This delayed feedback correlation capability is unattainable by traditional linear recommendation models (such as those that only consider the most recent action).

[0066] Based on the above description, the technical solution of this application has at least the following beneficial effects: Precisely matching business objectives: The tiered evaluation framework dynamically adjusts evaluation weights based on user role (e.g., analyst or developer) and task stage, ensuring recommendations are tailored to specific needs. In analyst scenarios, high coverage weighting prevents misclassification of low-precision operations. In developer scenarios, high precision weighting prioritizes syntactically accurate operations. This results in a more scientific overall score calculation, addressing issues of ambiguity in multiple evaluations and misaligned goals. Intelligently identifying user interests: A thermal calibration mechanism captures implicit feedback, such as mouse hover, to eliminate recommendation position interference. For example, even if the bottom recommendation "Sort by Date" is not clicked, the probability of a high hover time is increased through thermal index calibration, reducing misclassification by 28% and preventing the system from misjudging user interests due to location. Optimizing cross-stage action association: By calculating correlation based on semantic similarity and time decay, actions taken after a certain period of time are associated with historical recommendations. This increases the success rate from 12% to 53%, preventing misclassified recommendation failures and increasing the probability of subsequent recommendations by 15%, optimizing the model's understanding of long-term operational logic.

[0067] In some embodiments, this method can directly measure the degree of match between recommendation results and the user's actual needs by calculating the hit ratio (precision) of the user's actual selected action (second action) within the candidate recommendation set. A high precision rate indicates that the recommendation results are more aligned with the user's intent, reducing unnecessary information interference and improving operational efficiency (e.g., users can find the desired function or content more quickly). Analyzing the correlation (coherence) between the first and second actions can identify the rationality of the user's action path. For example, if a user searches for "running shoes" and then clicks "buy," high coherence indicates that the recommendation process meets user expectations. If the user switches to a non-recommended action (e.g., exiting a page), this indicates a gap in the recommendation logic and requires adjustments to the action flow or the associated recommended content. The coverage rate of the candidate recommendation set reflects its ability to cover potential user actions (e.g., covering similar functions, related content, or alternative options). High coverage can avoid "information cocoons" and ensure that users have rich choices in different scenarios. It is especially suitable for scenarios that require exploratory operations (such as e-commerce recommendations and function navigation). Combining the three dimensions of accuracy, coverage, and consistency can avoid the limitations of a single indicator: accuracy ensures "accurate recommendations", coverage ensures "complete recommendations", and consistency ensures "smooth processes". The coordinated optimization of the three can achieve a balance between accuracy, diversity, and logic in recommendation results.

[0068] Combined with the above Figure 1 The optimization method for database interactive recommendation provided in the embodiment of the present application is introduced in detail. The apparatus and device provided in the embodiment of the present application will be introduced below with reference to the accompanying drawings.

[0069] like Figure 2 As shown in FIG, this figure is a schematic diagram of an optimization device for database interactive recommendation provided by an embodiment of the present application, the device comprising: Acquisition module 201 is used to acquire a first operation of a user and a set of candidate recommended operations corresponding to the first operation; and acquire a second operation actually selected by the user; Determination module 202, configured to determine the user's precision based on the second operation, determine the user's consistency based on the second operation and the first operation, determine the user's coverage based on the candidate recommended operation set; and determine a feedback optimization coefficient for the user based on the precision, coverage, and consistency. The optimization module 203 is configured to optimize the candidate recommended operation set according to the feedback optimization coefficient and the process information of the user selecting the second operation.

[0070] Optionally, the acquisition module 201 is further configured to acquire a third operation selected by the user; The optimization module 203 is specifically used to calculate the similarity between the third operation and all candidate operations within a preset time interval; if there is a target candidate operation among all candidate operations whose similarity is greater than or equal to the similarity threshold, the recommendation probability of the target candidate operation is increased and optimized according to the target similarity between the target candidate operation and the third operation.

[0071] Optionally, the optimization module 203 is specifically used to update the initial learning rate according to the feedback optimization coefficient to obtain a dynamic learning rate; determine the user's initial thermal index for the candidate operation in the candidate recommended operation set according to the process information of the user selecting the second operation; update the initial thermal index according to the sensitivity coefficient to obtain a dynamic thermal index; obtain the initial recommendation probability of the candidate operation in the candidate recommended operation set; update the initial recommendation probability according to the dynamic learning rate and the dynamic thermal index to obtain the dynamic recommendation probability corresponding to the candidate operation.

[0072] Optionally, the optimization module 203 is specifically used to determine the target correlation between the target candidate operation and the third operation based on the historical recommendation probability of the target candidate operation, the target similarity between the target candidate operation and the third operation, and the target time interval between the third operation and the target candidate operation; and according to the target correlation, increase and optimize the recommendation probability of the target candidate operation.

[0073] Optionally, the acquisition module 201 is also used to obtain the target role label of the user; the determination module 202 is used to determine the target weight coefficient corresponding to the target role label based on the mapping relationship between the pre-configured reference role label and the reference weight coefficient, the target weight coefficient including the target accuracy coefficient, the target coverage coefficient and the target coherence coefficient, and determine the feedback optimization coefficient of the user based on the accuracy, the target accuracy coefficient, the coverage, the target coverage coefficient, the coherence and the target coherence coefficient.

[0074] Optionally, the precision is determined by the hit status of the second operation actually selected by the user in the set of candidate recommended operations corresponding to the first operation; the coverage is obtained by the reasonable candidate operations included in the set of candidate recommended operations corresponding to the first operation and all reasonable candidate operations corresponding to the first operation; the coherence is obtained by the correlation between the second operation and the first operation.

[0075] Optionally, the process information of the user selecting the second operation includes: the time the control cursor stays on each candidate operation and the position of each candidate operation during the process of the user selecting the second operation.

[0076] The database interactive recommendation optimization device according to the embodiment of the present application may correspond to the method described in the embodiment of the present application, and the above-mentioned other operations and / or functions of each module / unit of the database interactive recommendation optimization device are respectively to achieve Figure 1 For the sake of brevity, the corresponding processes of the various methods in the illustrated embodiments are not described here in detail.

[0077] The present application also provides a computing device. Figure 3 As shown, this figure is a schematic diagram of a computing device provided by an embodiment of the present application, and the computing device 400 includes a bus 401, a processor 402, a communication interface 403 and a memory 404. The processor 402, the memory 404 and the communication interface 403 communicate with each other via the bus 401.

[0078] The bus 401 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0079] The processor 402 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0080] The communication interface 403 is used for communicating with the outside.

[0081] Memory 404 may include volatile memory, such as random access memory (RAM). Memory 404 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0082] The memory 404 stores executable codes, and the processor 402 executes the executable codes to perform the aforementioned optimization method for database interactive recommendation.

[0083] Specifically, in the implementation Figure 2 In the case of the embodiment shown, and Figure 2 When each module or unit of the database interactive recommendation optimization device described in the embodiment is implemented by software, Figure 2 The software or program code required for the functions of each module / unit in the system may be partially or completely stored in the memory 404. The processor 402 executes the program code corresponding to each unit stored in the memory 404 to perform the aforementioned optimization method for database interactive recommendation.

[0084] Embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center that contains one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, hard disk, or magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to execute the aforementioned database interactive recommendation optimization method.

[0085] The present application also provides a computer program product comprising one or more computer instructions that, when loaded and executed on a computing device, fully or partially generate the process or function described in the present application.

[0086] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, or data center to another website, computer, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0087] When the computer program product is executed by a computer, the computer performs any of the aforementioned optimization methods for database interactive recommendation. The computer program product may be a software installation package, and when any of the aforementioned optimization methods for database interactive recommendation is needed, the computer program product may be downloaded and executed on the computer.

[0088] The descriptions of the processes or structures corresponding to the above figures have different emphases. For parts that are not described in detail in a certain process or structure, please refer to the relevant descriptions of other processes or structures.

[0089] The above description is only a specific implementation method of the present application, but the protection scope of the present application is not limited thereto. Any changes or replacements within the technical scope disclosed in the present application should be included in the protection scope of the present application.

Claims

1. A database interactive recommendation optimization method, characterized in that: The method comprises: Obtaining a first operation of the user and a set of candidate recommended operations corresponding to the first operation; obtaining a second operation actually selected by the user; determining the user's accuracy based on the second operation, determining the user's consistency based on the second operation and the first operation, and determining the user's coverage based on the candidate recommended operation set; determining a feedback optimization coefficient for the user according to the accuracy, the coverage, and the consistency; The candidate recommended operation set is optimized according to the feedback optimization coefficient and the process information of the user selecting the second operation.

2. The method according to claim 1, characterized in that After obtaining the second operation actually selected by the user, the method further includes: Obtaining a third operation selected by the user; Calculating similarity between the third operation and all candidate operations within a preset time interval; If there is a target candidate operation with a similarity greater than or equal to the similarity threshold among all the candidate operations, the recommendation probability of the target candidate operation is increased and optimized according to the target similarity between the target candidate operation and the third operation.

3. The method according to claim 1, characterized in that Optimizing the candidate recommended operation set according to the feedback optimization coefficient and the process information of the user selecting the second operation includes: The initial learning rate is updated according to the feedback optimization coefficient to obtain a dynamic learning rate; determining, based on process information of the user selecting the second operation, an initial heat index of the user for a candidate operation in the set of candidate recommended operations; and updating the initial heat index based on a sensitivity coefficient to obtain a dynamic heat index; Obtaining initial recommendation probabilities of candidate operations in the candidate recommended operation set; The initial recommendation probability is updated according to the dynamic learning rate and the dynamic heat index to obtain the dynamic recommendation probability corresponding to the candidate operation.

4. The method according to claim 2, characterized in that The increasing and optimizing the recommendation probability of the target candidate operation according to the target similarity between the target candidate operation and the third operation includes: determining a target relevance between the target candidate operation and the third operation based on a historical recommendation probability of the target candidate operation, a target similarity between the target candidate operation and the third operation, and a target time interval between the third operation and the target candidate operation; According to the target relevance, the recommendation probability of the target candidate operation is increased and optimized.

5. The method according to claim 1, wherein The method further comprises: Obtain the target role tag of the user; Determine a target weight coefficient corresponding to the target role label according to a pre-configured mapping relationship between the reference role label and the reference weight coefficient, wherein the target weight coefficient includes a target accuracy coefficient, a target coverage coefficient, and a target coherence coefficient; The determining of the user feedback optimization coefficient according to the accuracy, the coverage, and the coherence includes: A feedback optimization coefficient for the user is determined according to the accuracy, the target accuracy coefficient, the coverage, the target coverage coefficient, the coherence, and the target coherence coefficient.

6. The method according to claim 1, characterized in that The precision is determined by whether the second operation actually selected by the user hits the set of candidate recommended operations corresponding to the first operation; the coverage is obtained by including the reasonable candidate operations in the set of candidate recommended operations corresponding to the first operation and all reasonable candidate operations corresponding to the first operation; The coherence is obtained from the correlation between the second operation and the first operation.

7. The method according to any one of claims 1 to 6, characterized in that The process information of the user selecting the second operation includes: the stay time of the control cursor on each candidate operation and the position of each candidate operation during the process of the user selecting the second operation.

8. A database interactive recommendation optimization device, characterized in that: The device comprises: An acquisition module, configured to acquire a first operation of a user and a set of candidate recommended operations corresponding to the first operation; and acquire a second operation actually selected by the user; a determination module, configured to determine the user's precision rate based on the second operation, determine the user's consistency based on the second operation and the first operation, determine the user's coverage rate based on the candidate recommended operation set; and determine a feedback optimization coefficient for the user based on the precision rate, the coverage rate, and the consistency; An optimization module is configured to optimize the candidate recommended operation set according to the feedback optimization coefficient and process information of the user selecting the second operation.

9. A computing device, characterized in that including memory and processor; One or more computer programs are stored in the memory, and the one or more computer programs include instructions; when the instructions are executed by the processor, the computing device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium is used to store a computer program, and the computer program is used to execute the method according to any one of claims 1 to 7.

Images