Cache management method and device based on interface value evaluation

Abstract

The invention discloses a cache management method based on interface value evaluation, and relates to the technical field of computer science, in particular to a cache management method and device based on interface value evaluation. According to the method, the weight of each evaluation field is automatically calculated by utilizing a machine learning algorithm, and the value of each interface is comprehensively evaluated, so that the cache is dynamically managed, the interface cache can be efficiently managed, a low-value interface is prevented from occupying precious cache resources, and the performance and efficiency of a system can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer science and technology, and particularly to a cache management method and device based on interface value evaluation. Background Art

[0002] In modern Web applications, the efficient management of WebService interfaces is crucial for system performance and resource utilization. With the increase in the number of user requests and the improvement of business complexity, how to dynamically manage the cache of WebService interfaces under limited resources (especially memory) has become an urgent problem to be solved. Traditional cache management methods usually rely on fixed policies, such as LRU (Least Recently Used) or LFU (Least Frequently Used), but these methods often seem inadequate when facing complex and changing business requirements.

[0003] In order to improve the intelligence and refinement level of cache management, many studies have proposed dynamic cache management methods based on machine learning. These methods automatically adjust the cache policy by analyzing the characteristics of user requests to maximize system performance and resource utilization. However, existing methods still have some limitations when dealing with multi-dimensional features and user-defined requirements. For example:

[0004] 1. Chinese Patent No. CN116848516A, "A Distributed Cache System and Data Caching Method", which relates to the storage field and is used to improve the bandwidth of the distributed cache system. The distributed cache system includes a data management node, a data request node, and a memory. The data management node is used to manage cache consistency for the data in the memory; the data request node is used to send a request message to the data management node to request caching of the target data and update the cache expiration time according to time.

[0005] Disadvantages: This method mainly focuses on the bandwidth and cache consistency management of the distributed cache system, but lacks flexibility when dealing with multi-dimensional features and user-defined requirements.

[0006] Reason: Since this method mainly relies on fixed cache policies and cannot dynamically adjust the cache policy according to real-time business requirements and user feedback, it performs poorly in complex and changing business environments.

[0007] 2. Chinese Patent No. CN116383227A, "A Distributed Cache and Data Storage Consistency Processing System and Method", which relates to distributed cache and data storage consistency processing, and realizes the consistent storage of transaction data of distributed applications through a cache consistency processing method, device, electronic device, and storage medium.

[0008] Disadvantages: This method focuses on the transactional data consistency storage of distributed applications, but it has deficiencies in the intelligence and refinement of cache management.

[0009] Reason: This method mainly solves the data consistency problem and does not fully consider the dynamic management and optimization of cache resources, resulting in limitations in resource utilization and system performance.

[0010] 3. Chinese Patent "A Cache Management Method, System, Electronic Device and Storage Medium" with Publication No. CN115237539A. This method reduces the performance overhead during memory management by setting cache management policies. The specific implementation includes that the cache controller receives and caches read and write requests, arbitrates commands, and puts the commands into the command queue, takes out commands from the command queue and initiates read and write access operations to the SRAM unit.

[0011] Disadvantages: This method reduces the performance overhead during memory management by setting cache management policies, but it lacks flexibility and intelligence when dealing with multi-dimensional features and user-defined requirements.

[0012] Reason: This method mainly relies on preset cache management policies and cannot dynamically adjust cache policies according to real-time business requirements and user feedback, resulting in poor performance in complex and changing business environments.

[0013] 4. Chinese Patent "A Cache Processing Method, Cache Controller and Readable Storage Medium" with Publication No. CN118426701A. This method receives and caches read and write requests in the cache controller, arbitrates commands, and puts the commands into the command queue, takes out commands from the command queue and initiates read and write access operations to the SRAM unit, and determines and controls the target SRAM unit to enter the low-power mode based on the access volume.

[0014] Disadvantages: This method receives and caches read and write requests in the cache controller, arbitrates commands, and puts the commands into the command queue, but it lacks flexibility and intelligence when dealing with multi-dimensional features and user-defined requirements.

[0015] Reason: This method mainly focuses on the hardware implementation of the cache controller and the low-power mode, and does not fully consider the dynamic management and optimization of cache resources, resulting in limitations in resource utilization and system performance.

[0016] The above existing technical solutions solve the cache management problem to a certain extent, but there are still some limitations when dealing with multi-dimensional features and user-defined requirements. The main reason is that most of these methods rely on fixed cache policies, lack flexibility and intelligence, and cannot dynamically adjust cache policies according to real-time business requirements and user feedback. Summary of the Invention

[0017] The object of the present invention is to avoid the deficiencies in the prior art and provide a cache management technology that can automatically calculate the weights of each evaluation field using machine learning algorithms, thereby dynamically determining which interfaces should be retained in the cache.

[0018] The object of the present invention is achieved by the following technical solutions:

[0019] Therefore, according to one aspect disclosed by the present invention, there is provided a cache management method based on interface value evaluation, including the following steps:

[0020] S1: Obtain user request data sent by the interface;

[0021] S2: Perform feature transformation on the user request data to generate corresponding feature vectors; wherein, the feature vectors include a number of feature indicators;

[0022] S3: Input the feature vectors into a pre-trained logistic regression model to calculate the predicted probability that the interface needs to be retained in the cache; wherein, the logistic regression model includes a prediction function model and a probability distribution model;

[0023] S4: Determine whether the predicted probability calculated in step S3 is greater than a preset decision threshold. If so, retain the interface data in the cache; if not, remove the interface data from the cache.

[0024] Specifically, step S3 includes the following steps:

[0025] S31: Input the feature vectors into the pre-trained prediction function model and calculate the corresponding output value; wherein, the prediction function model includes various feature variables; when inputting the feature vectors, each feature indicator in the feature vectors is respectively input into the corresponding feature variable;

[0026] S32: Perform Sigmoid function transformation on the output value calculated in step S31 through the probability distribution model to calculate the corresponding predicted probability.

[0027] More specifically, the prediction function model includes a parameter vector composed of a number of model parameters, and each model parameter corresponds to each feature variable.

[0028] More specifically, the expression of the prediction function model is: h θ (X) = θ 1 x 1 + θ 2 x 1 + … + θ i x i + b;

[0029] wherein, X represents the feature vector; x 1 , x2 ,...,x i are the respective feature variables; θ 1 , θ 2 ,..., θ u are the model parameters corresponding to the respective feature variables; b is the bias parameter.

[0030] More specifically, the expression of the probability distribution model is: P(y|X; θ) = (h θ (X)) y (1 - h θ (X)) 1-y ; where y is a binary classification label used to indicate the cache status of the interface during this time period.

[0031] As above, the respective model parameters in the parameter vector are calculated by the maximum likelihood estimation method.

[0032] Furthermore, the respective model parameters in the parameter vector are iteratively updated by the gradient ascent method.

[0033] According to another aspect disclosed by the present invention, there is provided a cache management device based on interface value evaluation, which adopts the steps of the above-mentioned cache management method based on interface value evaluation, and includes:

[0034] An acquisition module, configured to obtain user request data sent by the interface;

[0035] A feature conversion module, configured to perform feature conversion on the user request data to generate a corresponding feature vector;

[0036] An operation module, which has a logistic regression model built in, and is configured to calculate the predicted probability that the interface needs to be retained in the cache; wherein, the logistic regression model includes a prediction function model and a probability distribution model;

[0037] A decision module, configured to determine whether the predicted probability calculated by the operation module is greater than a preset decision threshold. If so, retain the interface data in the cache. If not, remove the interface data from the cache.

[0038] According to yet another aspect disclosed by the present invention, there is provided a computing device, including a memory, a processor, and computer instructions stored on the memory and executable on the processor. When the processor executes the instructions, it implements the steps of the above-mentioned cache management method based on interface value evaluation.

[0039] According to another aspect disclosed by the present invention, there is provided a computer-readable storage medium, which stores computer instructions. When the instructions are executed by the processor, it implements the steps of the above-mentioned cache management method based on interface value evaluation.

[0040] Advantages of the present invention: The present invention provides a cache management method based on interface value evaluation. By constructing a corresponding user request feature data set and using a logistic regression model, the weights of each evaluation field are automatically calculated using machine learning algorithms, and the value of each interface is comprehensively evaluated, thereby dynamically managing the cache. It can not only efficiently manage the interface cache, prevent low-value interfaces from occupying precious cache resources, but also improve the performance and efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The present invention can be better understood by describing the exemplary embodiments disclosed in the present invention in conjunction with the accompanying drawings. In the drawings:

[0042] Figure 1 FIG. shows a schematic flowchart of a cache management method based on interface value evaluation according to Embodiment 1 of the present invention disclosed;

[0043] Figure 2 FIG. shows a schematic diagram of program modules of a cache management device based on interface value evaluation according to Embodiment 1 of the present invention disclosed;

[0044] Figure 3 FIG. shows a schematic hardware structure diagram of a computing device according to Embodiment 1 of the present invention disclosed. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The following will describe the specific embodiments of the present invention. It should be noted that in the process of the specific description of these embodiments, for the sake of concise description, this specification cannot describe all the features of the actual embodiments in detail. It should be understood that in the actual implementation process of any one of the embodiments, just as in the process of any engineering project or design project, in order to achieve the specific goals of the developer and to meet the system-related or business-related restrictions, various specific decisions are often made, and these decisions will also change from one embodiment to another. In addition, it can also be understood that although the efforts made in this development process may be complex and lengthy, for those of ordinary skill in the art related to the content disclosed in the present invention, some design, manufacturing or production changes based on the technical content disclosed in the present invention are only conventional technical means and should not be understood as the content of the present invention being insufficient.

[0046] Unless otherwise defined, technical terms or scientific terms used in the claims and the specification shall have the ordinary meanings as understood by those of ordinary skill in the technical field to which the present invention pertains. The terms "first", "second" and similar terms used in the specification and claims of this patent application for invention do not denote any order, quantity or importance, but are merely used to distinguish different components. The terms such as "a" or "an" do not denote a quantity limitation, but mean that there is at least one. The terms such as "comprising" or "including" mean that the elements or items appearing before "comprising" or "including" cover the elements or items listed after "comprising" or "including" and their equivalent elements, and do not exclude other elements or items. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, nor are they limited to direct or indirect connections.

[0047] Embodiment 1

[0048] Please refer to Figure 1 , this embodiment proposes a cache management method based on interface value evaluation, including the following steps S1 to S4:

[0049] S1: Obtain the user request data sent by the interface.

[0050] In this embodiment, collecting user request data includes the request time, request path, request parameters, etc. of the user. This data will be collected through website server logs or a dedicated request recording module.

[0051] As shown in Table 1, the collected data items include: request timestamp, request frequency, response time, data size, error rate, cache hit rate, request priority, business importance, user satisfaction, request method, response status code, and request path.

[0052]

[0053] Table 1 Data items of user requests and their processing methods for feature conversion

[0054] S2: Perform feature conversion on the user request data to generate corresponding feature vectors; wherein, the feature vectors include several feature indicators, corresponding to the above-mentioned respective data items, and the request characteristics and behavior patterns of the interface within this time period are jointly described by the feature indicators corresponding to the respective data items.

[0055] In this embodiment, each feature indicator is the result output after the respective data items in Table 1 are converted according to the corresponding processing methods. The set of indicators for each WebService interface within a specific historical time period is represented as a feature vector X. Specifically, the feature vector X = (x 1 ,x2 ,..., x i ), where x 1 , x 2 ,..., x i represent different characteristic indicators respectively.

[0056] S3: Input the feature vector into the pre-trained logistic regression model to calculate the predicted probability that the interface needs to be retained in the cache; wherein, the logistic regression model includes a prediction function model and a probability distribution model; the prediction function model includes a parameter vector composed of several model parameters, and each model parameter corresponds to each feature variable respectively.

[0057] Specifically, step S3 includes the following steps:

[0058] S31: Input the feature vector into the pre-trained prediction function model and calculate the corresponding output value; wherein, the prediction function model includes each feature variable; when inputting the feature vector, each characteristic indicator in the feature vector is input into the corresponding feature variable respectively;

[0059] S32: Perform Sigmoid function conversion on the output value calculated in step S31 through the probability distribution model to calculate the corresponding predicted probability.

[0060] S4: Determine whether the predicted probability calculated in step S3 is greater than a preset decision threshold (usually 0.5). If so, retain the interface data in the cache; if not, remove the interface data from the cache.

[0061] In this embodiment, the prediction method adopted is a binary classification problem, aiming to determine whether the data of the WebService interface should be retained in the cache or eliminated from the cache. To achieve this goal, a prediction function based on a machine learning model (i.e., the above-mentioned logistic regression model) is constructed, and this function uses a set of metrics collected from historical data to predict the cache status of the interface.

[0062] More specifically, the expression of the prediction function model is: h θ (X) = θ 1 x 1 + θ 2 x 1 + … + θ i x i + b;

[0063] wherein, X represents the feature vector; x 1 , x 2 ,..., x i are each feature variable; θ 1 , θ 2 ,..., θ iθ is the model parameter corresponding to each feature variable; b is the bias parameter.

[0064] More specifically, corresponding to the feature vector X is the label y, which is a binary classification marker used to indicate the cache status of the interface during this time period.

[0065] More specifically, when the value of y is 1, it means that the data of this interface should be retained in the cache, and when the value of y is 0, it means that the data of this interface should be evicted from the cache.

[0066] Logistic regression (Logistic Function) uses a function to normalize the y value so that the value of y is within the interval (0, 1), and the value of y represents the probability that the interface is predicted to need to be retained in the cache.

[0067] In this embodiment, the function used is also called the Sigmoid function, and its expression is:

[0068]

[0069] Given the feature vector X and the model parameter θ, the prediction function h θ (X) can be written as:

[0070]

[0071] Here, is the dot product of the model parameter and the feature vector. The Sigmoid function is used to convert this dot product into a probability value, representing the probability that the interface is predicted to need to be retained in the cache.

[0072] For a given input, the probabilities that the interface belongs to y = 1 (needs caching) and y = 0 (evicts the cache) are respectively:

[0073] P(y = 1|X; θ) = h θ (X)

[0074] P(y = 0|X; θ) = 1 - h θ (X)

[0075] Combining the above conditional probability expressions, the expression of the probability distribution model is obtained as:

[0076] P(y|X; θ) = (h θ (X)) y (1 - h θ (X)) 1-y ;

[0077] where y is a binary classification marker used to indicate the cache status of the interface during this time period.

[0078] Further, each model parameter θ in the parameter vector is calculated by the maximum likelihood estimation method and iteratively updated by the gradient ascent method. The specific steps are as follows:

[0079] According to the expression of the above probability distribution model, we use the maximum likelihood estimation method to estimate the appropriate value. The maximum likelihood function L θ is, where m is the number of samples and n is a certain input sample, and the following expression is obtained:

[0080]

[0081] For convenience in taking the partial derivative of θ, take the logarithm of the likelihood function L(θ) to obtain the following expression:

[0082]

[0083] Then, we take the partial derivative of J(θ) with respect to each parameter θ j to find the gradient, where θ j represents the j-th parameter in the parameter vector θ:

[0084]

[0085]

[0086] Use the gradient ascent method to update each parameter θ j , where α is the learning rate:

[0087]

[0088] When the change in parameter update is less than the preset precision threshold, the iteration stops. The final parameter vector θ is used to predict the cache state of new samples.

[0089] In the present invention, by adopting the logistic regression model pre-trained by the above method, when in use, the user request is converted into a feature vector X and input into the prediction function model in the logistic regression model, and then the output of h θ (X) is calculated through the prediction function model, that is, θ 1 x 1 +θ 2 x 1 +…+θ i x i +b. Finally, h θThe output of (X) is transformed by the Sigmoid function through a probability distribution model to obtain the predicted probability P(y|X; θ). If P(y = 1|X) is greater than the decision threshold, the prediction result is 1, and the system will perform a cache retention operation, that is, retain the interface data in the cache. Otherwise, it is 0, and the system will perform a cache eviction operation, that is, remove the interface data from the cache.

[0090] Please continue to refer to Figure 2 , which shows a cache management device based on interface value evaluation. In this embodiment, a cache management device based on interface value evaluation may include or be divided into one or more program modules. One or more program modules are stored in a storage medium and executed by one or more processors to complete the present invention and implement the above-mentioned cache management method based on interface value evaluation. The program modules referred to in the present invention refer to a series of computer program instruction segments that can complete specific functions, and are more suitable for describing the execution process of a cache management device based on interface value evaluation in a storage medium than the program itself. The following description will specifically introduce the functions of each program module in this embodiment:

[0091] The acquisition module is used to obtain the user request data sent by the interface;

[0092] The feature transformation module is used to perform feature transformation on the user request data to generate corresponding feature vectors;

[0093] The operation module is built-in with a logistic regression model and is used to calculate the predicted probability that the interface needs to be retained in the cache; among them, the logistic regression model includes a prediction function model and a probability distribution model;

[0094] The decision module is used to determine whether the predicted probability calculated by the operation module is greater than a preset decision threshold. If so, retain the interface data in the cache; if not, remove the interface data from the cache.

[0095] According to another aspect disclosed in the present invention, a computing device is provided, including a memory, a processor, and computer instructions stored on the memory and executable on the processor. When the processor executes the instructions, the steps of the above-mentioned cache management method based on interface value evaluation are implemented.

[0096] This embodiment also provides a computing device, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack-mounted server, a blade server, a tower server, or a cabinet server (including an independent server or a server cluster composed of multiple servers) that can execute programs. The computing device 20 in this embodiment includes at least, but is not limited to: a memory 21 and a processor 22 that can communicate with each other through a system bus, as Figure 3 shown. It should be noted that Figure 3Only the computing device 20 with components 21-22 is shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively.

[0097] In this embodiment, the memory 21 (i.e., the readable storage medium) includes flash memory, hard disk, multimedia card, card-type memory (e.g., 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, magnetic disk, optical disk, etc. In some embodiments, the memory 21 may be an internal storage unit of the computing device 20, such as the hard disk or memory of the computing device 20. In other embodiments, the memory 21 may also be an external storage device of the computing device 20, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc. equipped on the computing device 20. Of course, the memory 21 may also include both the internal storage unit and the external storage device of the computing device 20. In this embodiment, the memory 21 is generally used to store the operating system and various application software installed on the computing device 20, such as the program code of a cache management device based on interface value evaluation in Embodiment 1. In addition, the memory 21 may also be used to temporarily store various data that have been output or will be output.

[0098] In some embodiments, the processor 22 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 22 is generally used to control the overall operation of the computing device 20. In this embodiment, the processor 22 is used to run the program code stored in the memory 21 or process data, such as running a cache management device based on interface value evaluation to implement a cache management method based on interface value evaluation in Embodiment 1.

[0099] This embodiment also provides a computer-readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disc, a server, an App application store, etc. A computer program is stored thereon, and when the program is executed by a processor, corresponding functions are implemented. The computer-readable storage medium of this embodiment is used to store a cache management device based on interface value evaluation, and when executed by a processor, it implements a cache management method based on interface value evaluation in Embodiment 1.

[0100] In summary, according to the exemplary embodiment, a cache management method and device based on interface value evaluation of the present invention. The method constructs a corresponding user request feature data set, and through a logistic regression model, uses a machine learning algorithm to automatically calculate the weights of each evaluation field, comprehensively evaluate the value of each interface, and thus dynamically manage the cache; it can not only efficiently manage the interface cache, prevent low-value interfaces from occupying precious cache resources, and the defined h θ (X) and its regression strategy can efficiently obtain a fitted function curve, and based on the parameter θ obtained by regression, the cache management strategy can obtain a prediction result based on simple operations, improving the performance and efficiency of the system and having strong real-time performance.

[0101] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0102] Any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of executable instructions including one or more steps for implementing a specific logical function or process. The scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, and this should be understood by those skilled in the technical field to which the embodiments of the present invention belong.

[0103] Those of ordinary skill in the art in this technical field can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0104] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0105] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0106] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A cache management method based on interface value evaluation, characterized in that: The following steps are involved: S1: Get the user request data sent by the interface; S2: Perform feature conversion on the user request data to generate a corresponding feature vector; wherein the feature vector includes a number of feature indicators; S3: Input the feature vector into a pre-trained logistic regression model to calculate the predicted probability that the interface needs to be retained in the cache; wherein the logistic regression model includes a prediction function model and a probability distribution model; S4: Determine whether the predicted probability calculated in step S3 is greater than a preset decision threshold. If so, retain the interface data in the cache; if not, remove the interface data from the cache.

2. A cache management method based on interface value evaluation according to claim 1, characterized in that: The step S3 comprises the following steps: S31: Input the feature vector into the pre-trained prediction function model and calculate the corresponding output value; Wherein, the prediction function model includes various characteristic variables; when the characteristic vector is input, each characteristic index in the characteristic vector is respectively input into the corresponding characteristic variable; S32: Performing Sigmoid function conversion on the output value calculated in step S31 through the probability distribution model to calculate the corresponding prediction probability.

3. The cache management method based on interface value evaluation according to claim 2, characterized in that: The prediction function model includes a parameter vector composed of a number of model parameters, and each model parameter corresponds to each characteristic variable.

4. The cache management method based on interface value evaluation according to claim 3 is characterized in that: The expression of the prediction function model is: θ (X) = θ1x1 + θ2x1 + ... + θ i x i +b; Where X represents the feature vector; x1, x2, ..., x i For each characteristic variable; θ1, θ2, ..., θ i are model parameters corresponding to each of the characteristic variables; b is the bias parameter.

5. The cache management method based on interface value evaluation according to claim 4 is characterized in that: The expression of the probability distribution model is: P(y|X;θ)=(h θ (X) y (1-h θ (X) 1-y ; Wherein, y is a binary label, which is used to indicate the cache status of the interface in the time period.

6. A cache management method based on interface value evaluation according to any one of claims 3 to 5, characterized in that: Each model parameter in the parameter vector is calculated by maximum likelihood estimation method.

7. The cache management method based on interface value evaluation according to claim 6, characterized in that: Each model parameter in the parameter vector is iteratively updated by a gradient ascent method.

8. A cache management device based on interface value evaluation, adopting a cache management method based on interface value evaluation according to any one of claims 1 to 7, characterized in that: include: The acquisition module is used to obtain the user request data sent by the interface; A feature conversion module is used to perform feature conversion on user request data and generate corresponding feature vectors; A computing module, having a built-in logistic regression model, for calculating the predicted probability that the interface needs to be retained in the cache; wherein the logistic regression model includes a prediction function model and a probability distribution model; The decision module is used to determine whether the prediction probability calculated by the operation module is greater than a preset decision threshold. If so, the interface data is retained in the cache; if not, the interface data is removed from the cache.

9. A computing device comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, characterized in that: When the processor executes the instructions, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer instructions, characterized in that: When the instruction is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images