Software pushing method and device, equipment, storage medium and program product

Abstract

The invention relates to a software pushing method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: identifying evaluation information of an updated version according to user comment information of a corresponding updated version of cloud platform software; identifying a use probability value of the updated version according to historical version use data of the cloud platform software by the user to be pushed, and performing software pushing classification according to the evaluation information and the use probability value to obtain a software pushing classification result; and pushing the software to the terminal of the to-be-pushed user based on the software pushing classification result. By adopting the method, the accuracy and safety of software updating can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of cloud computing, and in particular to a software push method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the development of cloud computing technology, users can connect to remote cloud servers and use cloud platform services with just a monitor and a thin terminal. Usually, cloud platforms not only contain the required virtualization hardware, but also have built-in software necessary for users, such as browsers, antivirus software, encryption software, cloud disks, communication software, etc. When the software version in the cloud platform is too low to match the hardware or operating system standards, or when users need a higher version of the software, users usually download the required software manually.

[0003] However, it is troublesome for users to manually update the software, and it is easy for them to download unsafe software, resulting in reduced accuracy and security of software updates. Summary of the invention

[0004] Based on this, it is necessary to provide a software push method, apparatus, computer device, computer-readable storage medium and computer program product that can improve the accuracy and security of software updates in response to the above technical problems.

[0005] In a first aspect, the present application provides a software push method, comprising:

[0006] Identify evaluation information of the updated version based on user review information of the corresponding updated version of the cloud platform software;

[0007] Identify the usage probability value of the updated version based on the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push based on the evaluation information and the usage probability value to obtain the software push classification result;

[0008] Based on the software push classification results, the software is pushed to the terminal of the user to be pushed.

[0009] In one embodiment, identifying evaluation information of the updated version according to user review information corresponding to the updated version of the cloud platform software includes:

[0010] Inputting the updated version of user review information into the review recognition model for recognition, and obtaining the output updated version of evaluation information;

[0011] According to the historical version usage data of the cloud platform software by the users to be pushed, the usage probability value of the updated version is identified, and the software push classification is performed according to the evaluation information and the usage probability value to obtain the software push classification results, including:

[0012] The evaluation information and historical version usage data are input into the software push classification model to identify the usage probability value of the updated version using the historical version usage data through the software push classification model, and the software push is classified according to the evaluation information and the usage probability value to obtain the output software push classification result.

[0013] In one of the embodiments, the software push classification model includes a recognition network, a fusion network, and a discriminant network;

[0014] The software push classification model uses the historical version usage data to identify the usage probability value of the updated version, and classifies the software push according to the evaluation information and the usage probability value to obtain the output software push classification results, including:

[0015] Input the historical version usage data into the recognition network for recognition, and obtain the usage probability value of the updated version as output;

[0016] The evaluation information and the usage probability value are fused to obtain the fusion feature;

[0017] The fused features are input into the fusion network for recognition, and the updated version of the comprehensive usage probability value is output;

[0018] The comprehensive usage probability value is input into the discriminant network to perform software push classification discrimination, and the output software push classification result is obtained, and the software push classification result includes a software push degree value;

[0019] Push software to users to be pushed based on the software push classification results, including:

[0020] When the software push degree value is greater than or equal to the push threshold, the updated version of the cloud platform software is pushed to the user to be pushed;

[0021] When the software push degree value is less than the push threshold, stop pushing the updated version of the cloud platform software to the users to be pushed.

[0022] In one embodiment, the training of the software push classification model includes the following steps:

[0023] Obtaining training evaluation information corresponding to the training version, where the training evaluation information is obtained by identifying user comment information corresponding to the training version of the training software;

[0024] Obtaining training data corresponding to the training user and installation tags corresponding to the training data, wherein the training data includes historical version usage data of the training software and historical version usage data of related software, where the related software is cloud platform software related to the training software;

[0025] Input the training evaluation information and the training data into the initial software push classification model to obtain the output initial software push classification result;

[0026] The initial software push classification model is trained based on the initial software push classification results and the installation labels to obtain a software push classification model.

[0027] In one embodiment, the updated versions include at least two, and the users to be pushed include at least two; identifying the evaluation information of the updated version according to the user review information corresponding to the updated version of the cloud platform software includes:

[0028] Obtain a user review information matrix corresponding to the cloud platform software, where the review information matrix includes user review information for each updated version;

[0029] Identify evaluation information of each updated version according to the user review information matrix to obtain an evaluation matrix;

[0030] According to the historical version usage data of the cloud platform software by the users to be pushed, the usage probability value of the updated version is identified, and the software push classification is performed according to the evaluation information and the usage probability value to obtain the software push classification results, including:

[0031] Obtaining the historical version usage data corresponding to each user, and identifying the usage probability value of the updated version according to the historical version usage data corresponding to each user, to obtain a usage probability value matrix;

[0032] The software push classification is performed according to the evaluation matrix and the use probability value matrix to obtain a software push classification result matrix, wherein the software push classification result matrix includes the software push classification results of each updated version being pushed to each user respectively.

[0033] In one embodiment, before identifying the evaluation information of the updated version according to the user review information corresponding to the updated version of the cloud platform software, the following steps are included:

[0034] Monitor software update information of cloud platform software;

[0035] When the cloud platform software has corresponding software update information, download the software installation package of the corresponding updated version from the software supply platform based on the software update information;

[0036] The updated version of the software installation package is saved in the software database, and the number of installation packages of the cloud platform software is detected. When the number of installation packages exceeds a preset number, the software installation package of the cloud platform software corresponding to the target historical version is deleted;

[0037] Obtain corresponding user review information according to the updated version;

[0038] The software push classification result includes a software push degree value. The software is pushed to the user to be pushed based on the software push classification result, including:

[0039] When the software push degree value is greater than or equal to the push threshold, the updated version of the software installation package is pushed to the terminal of the user to be pushed, so that the terminal installs the updated version of the software installation package.

[0040] In a second aspect, the present application also provides a software push device, including:

[0041] An identification module, used to identify evaluation information of the updated version according to user review information of the corresponding updated version of the cloud platform software;

[0042] The classification module is used to identify the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push according to the evaluation information and the usage probability value to obtain the software push classification result;

[0043] The push module is used to push software to the terminal of the user to be pushed based on the software push classification result.

[0044] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0045] Identify evaluation information of the updated version based on user review information of the corresponding updated version of the cloud platform software;

[0046] Identify the usage probability value of the updated version based on the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push based on the evaluation information and the usage probability value to obtain the software push classification result;

[0047] Based on the software push classification results, the software is pushed to the terminal of the user to be pushed.

[0048] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0049] Identify evaluation information of the updated version based on user review information of the corresponding updated version of the cloud platform software;

[0050] Identify the usage probability value of the updated version based on the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push based on the evaluation information and the usage probability value to obtain the software push classification result;

[0051] Based on the software push classification results, the software is pushed to the terminal of the user to be pushed.

[0052] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0053] Identify evaluation information of the updated version based on user review information of the corresponding updated version of the cloud platform software;

[0054] Identify the usage probability value of the updated version based on the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push based on the evaluation information and the usage probability value to obtain the software push classification result;

[0055] Based on the software push classification results, the software is pushed to the terminal of the user to be pushed.

[0056] The above-mentioned software push method, device, computer equipment, computer-readable storage medium and computer program product identify the evaluation information of the updated version according to the user comment information of the corresponding updated version of the cloud platform software; identify the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push according to the evaluation information and the usage probability value to obtain the software push classification result; push the software to the terminal of the user to be pushed based on the software push classification result. That is, by combining the comment information and the historical version usage data to classify the software push, the accuracy of the software push classification result is improved, and then the software is pushed to the terminal of the user to be pushed according to the software push classification result, and the updated version that meets the usage habits of the user to be pushed and the positive user evaluation can be pushed, thereby realizing personalized push, and then improving the accuracy of software push. And updating the cloud platform software by software push can avoid users manually downloading unsafe software, thereby improving the security of software updates. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0058] FIG1 is an application environment diagram of a software push method in one embodiment;

[0059] FIG2 is a schematic diagram of a flow chart of a software push method in one embodiment;

[0060] FIG3 is a schematic diagram of a process flow of model prediction in one embodiment;

[0061] FIG4 is a schematic diagram of a model architecture of software push in one embodiment;

[0062] FIG5 is a schematic diagram of a process for obtaining user comment information in one embodiment;

[0063] FIG6 is a schematic diagram of the overall framework of software push in one embodiment;

[0064] FIG7 is a structural block diagram of a software push device in one embodiment;

[0065] FIG. 8 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0067] The software push method provided in the embodiment of the present application can be applied to the application environment shown in FIG. 1. The cloud server 104 communicates with the cloud platform terminals of each user through the network, including the terminal 102. The cloud server can obtain the historical version usage data of the cloud platform software by the user to be pushed from each cloud platform terminal. The cloud server 104 can also obtain the user comment information of the updated version from the software supply platform. Then the cloud server 104 identifies the evaluation information of the updated version according to the user comment information corresponding to the updated version of the cloud platform software. The cloud server 104 identifies the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classifies the software push according to the evaluation information and the usage probability value to obtain the software push classification result. The cloud server 104 pushes the software to the terminal 102 of the user to be pushed based on the software push classification result. The terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head mounted device may be a virtual reality (VR) device, an augmented reality (AR) device, smart glasses, etc. The cloud server 104 is a cloud server that provides cloud computing services.

[0068] In an exemplary embodiment, as shown in FIG2 , a software push method is provided, which is described by taking the method applied to the cloud server in FIG1 as an example, and includes the following steps 202 to 206. Among them:

[0069] S202, identifying evaluation information of the updated version according to user review information of the corresponding updated version of the cloud platform software.

[0070] Among them, the user review information is the review information after the user uses the updated version of the cloud platform software, which may include user reviews of multiple users. Usually, after the latest version of the software is released, there will be users who experience it and then publish different user reviews. The user review information can directly and indirectly reflect the user experience and stability of the latest version of the software. When the user evaluation is positive, it indicates that the user experience and stability of the latest version of the software are good, and when the user evaluation is negative, it indicates that the latest version of the software has poor user experience and stability. That is, the evaluation information of the updated version can be identified through the user review information. The evaluation information is used to reflect the user preference of the updated version of the cloud platform software. The evaluation information may be the evaluation value of the updated version obtained by identification. The higher the evaluation value, the better the user experience of the updated version, and the lower the evaluation value, the worse the user experience of the updated version. The evaluation information may also be the user evaluation level of the updated version obtained by identification. The higher the user evaluation level, the better the user experience of the updated version, and the lower the user evaluation level, the worse the user experience of the updated version.

[0071] It is understandable that when the cloud server detects that an updated version of the cloud platform software has been released, user review information of the updated version can be collected from the software platform corresponding to the cloud platform software or the software supply platform to which the cloud platform software belongs. At this time, the cloud server can identify the collected user review information, and when the user review information is a user evaluation score, the user review information can be directly used as the evaluation information of the updated version. When the user review information is text information, the text information can be converted into a vector, and then the semantics in the vector can be identified through an intelligent algorithm to obtain the evaluation information of the updated version.

[0072] It is understandable that when the cloud server detects that the cloud platform software has released an updated version, the cloud server can obtain user comment information corresponding to the updated version of the cloud platform software in different ways. Among them, the cloud server can obtain user comment information that the software platform or software supply platform of the cloud platform software actively pushes the updated version. The cloud server can also collect user comment information of the updated version from the software platform or software supply platform of the cloud platform software. Then, the cloud server can determine the corresponding identification method to identify the evaluation information according to the information type of the user comment information, wherein, when the user comment information is a numerical type of information, the user comment information can be directly used as comment information. When the user comment information is text type information, the text comment information can be identified by an intelligent algorithm for text recognition to obtain evaluation information. When the user comment information is image type information, the image comment information can be identified as text information, and then the text recognition intelligent algorithm is used to identify it to obtain evaluation information, or the intelligent algorithm for image recognition can be directly used to identify it to obtain evaluation information.

[0073] In some embodiments, when the cloud server simultaneously monitors the release of updated versions of multiple cloud platform software, it can simultaneously obtain user review information of the corresponding updated versions of the cloud platform software, and then simultaneously identify the user review information of all updated versions to obtain evaluation information of each updated version.

[0074] S204, identifying the usage probability value of the updated version based on the historical version usage data of the cloud platform software by the user to be pushed, and classifying the software push based on the evaluation information and the usage probability value to obtain the software push classification result.

[0075] The historical version usage data refers to the usage data of the historical version of the cloud platform software by the user to be pushed, which may include but is not limited to the number of times and duration of use. Preferably, the usage data of multiple historical versions of the cloud platform software can be obtained, for example, the usage data of 5 historical versions can be obtained.

[0076] The usage probability value is used to characterize the possibility of a user using an updated version. The larger the usage probability value, the greater the possibility of a user using an updated version. The smaller the usage probability value, the smaller the possibility of a user using an updated version.

[0077] The software push classification result may include a software push category result and a software push prohibition category result. The software push category result is used to indicate that the software can be pushed to the terminal of the user to be pushed, and the software push prohibition category result is used to indicate that the software does not need to be pushed to the terminal of the user to be pushed.

[0078] In this embodiment, the cloud server can record the usage data of each historical version of the cloud platform software by the user to be pushed. Preferably, the cloud server can record the usage data of all historical versions of all cloud platform users. Then the cloud server can obtain the historical version usage data of the cloud platform software by the user to be pushed from the database. Then the cloud server uses the historical version usage data to predict the usage probability value of the updated version by the user to be pushed. Usually, a machine learning algorithm is used to predict the usage probability value. The machine learning algorithm can be a linear regression algorithm or a neural network algorithm. Finally, the cloud server uses the evaluation information and the usage probability value to classify the software push and obtain the software push classification result.

[0079] It is understandable that usually the software push classification can be two categories, including software push category and software push prohibited category. In some scenarios, the software push classification can also be multiple categories, for example, it can be three categories, including software push category, software push prohibited category and undetermined category. When it is identified that the software category cannot be determined, a prompt can be sent to the user to be pushed to ask whether to push, and the push can be made when the user chooses to push.

[0080] S206: Push the software to the terminal of the user to be pushed based on the software push classification result.

[0081] The terminal to be pushed may be a cloud terminal, and the cloud terminal may be composed of a display and a thin terminal. For example, a cloud server provided by a remote cloud vendor may be connected only through a display and a thin terminal.

[0082] In some embodiments, when the software push classification result is a software push category result, the cloud server can push the installation package of the corresponding updated version of the cloud platform software to the terminal of the user to be pushed, and the user can directly use the updated version of the installation package to update the cloud platform software. The cloud server can also push the update link of the corresponding updated version of the cloud platform software to the terminal of the user to be pushed, and the user can update the cloud platform software through the update link. When the software push classification result is a software push prohibited category result, the cloud server does not perform push processing.

[0083] Preferably, the cloud server can simultaneously classify software pushes for multiple cloud platform software, and obtain software push classification results corresponding to updated versions of multiple cloud platform software. Then, push screening is performed based on the software push classification results, wherein, usually, updated versions of some cloud platform software can be pushed to the terminals of the users to be pushed, and updated versions of some cloud platform software are not pushed. The cloud server can also push updated versions of all cloud platform software, or it can not push updated versions of all cloud platform software.

[0084] The above-mentioned software push method, device, computer equipment, computer-readable storage medium and computer program product identify the evaluation information of the updated version according to the user comment information of the corresponding updated version of the cloud platform software; identify the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push according to the evaluation information and the usage probability value to obtain the software push classification result; push the software to the terminal of the user to be pushed based on the software push classification result. That is, by combining the comment information and the historical version usage data to classify the software push, the accuracy of the software push classification result is improved, and then the software is pushed to the terminal of the user to be pushed according to the software push classification result, and the updated version that meets the usage habits of the user to be pushed and the positive user evaluation can be pushed, thereby realizing personalized push, and then improving the accuracy of software push. And updating the cloud platform software by software push can avoid users manually downloading unsafe software, thereby improving the security of software updates.

[0085] In an exemplary embodiment, as shown in FIG3 , identifying evaluation information of an updated version according to user review information corresponding to an updated version of the cloud platform software includes:

[0086] S302: input the updated version of the user review information into the review recognition model for recognition, and obtain the output updated version of the evaluation information.

[0087] Before using the comment recognition model, it is necessary to train the comment recognition model using historical data. That is, the cloud server uses the historical version of user comment information as the input of the comment recognition model to be trained, and then uses the loss between the model output and the training evaluation label to iterate the model parameters in the comment recognition model to be trained until the training completion condition is met, and a trained comment recognition model is obtained, wherein the comment recognition model to be trained can be a model established using an artificial intelligence algorithm, and the artificial intelligence algorithm can be a neural network algorithm or a linear regression algorithm. The training evaluation label can be a pre-set evaluation grade value label. At this time, the comment recognition model can be a multi-classification model, for example, the comment recognition model can be an LSTM (Long Short-Term Memory) model.

[0088] At this time, the cloud server deploys and uses the trained review recognition model. When in use, when the cloud server obtains the updated version of user review information, it calls the deployed review recognition model to recognize the updated version of the user review information and obtains the output updated version of the evaluation grade value.

[0089] Furthermore, when the cloud server detects that an updated version of the cloud platform software has been released, the cloud server can collect user comment information corresponding to each updated version of the cloud platform software. For cloud platform users, whether the software needs to be updated depends not only on whether the new version of the software is powerful, but also on whether the stability and user experience of the new version of the software are good. When the updated version of the software is released, it is usually a test version, and it still needs to be iterated from the stable version. Then, after the updated version of the software is released, user comments can be obtained, and the user comments can directly and indirectly reflect the user experience and stability of the updated version of the software. When the user evaluation is more positive, it indicates that the updated version of the software has a better user experience. When the user evaluation is negative, it indicates that the updated version of the software has poor stability and poor user experience. Therefore, when the cloud platform software iterates from the test updated version to the stable updated version, n user comments of each updated version can be collected, and then identified at the same time to obtain the evaluation level value corresponding to each updated version of the corresponding cloud platform software.

[0090] In some embodiments, before inputting the updated version of the user comment information into the comment recognition model, the updated version of the user comment information may be vectorized. For example, the user comment information may be converted into a feature vector through a word embedding model.

[0091] S204, that is, identifying the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classifying the software push according to the evaluation information and the usage probability value, to obtain the software push classification result, including:

[0092] S304, input the evaluation information and historical version usage data into the software push classification model to identify the usage probability value of the updated version using the historical version usage data through the software push classification model, and perform software push classification based on the evaluation information and the usage probability value to obtain the output software push classification result.

[0093] The software push classification model is pre-trained and deployed in the cloud server. When the evaluation information and historical version usage data are obtained, the software push classification model can be directly called to perform software push classification and obtain the output software push classification result.

[0094] In an exemplary embodiment, the software push classification model includes a recognition network, a fusion network, and a discriminant network;

[0095] S304, using the historical version usage data through the software push classification model to identify the usage probability value of the updated version, and performing software push classification according to the evaluation information and the usage probability value, to obtain the output software push classification result, including:

[0096] S304a, inputting the historical version usage data into the recognition network for recognition, and obtaining the output usage probability value of the updated version;

[0097] S304b, fusing the evaluation information and the usage probability value to obtain a fusion feature;

[0098] S304c, inputting the fusion feature into the fusion network for identification, and obtaining an outputted comprehensive usage probability value of the updated version;

[0099] S304d, inputting the comprehensive use probability value into the discrimination network to perform software push classification discrimination, and obtaining an output software push classification result, the software push classification result including the software push degree value;

[0100] S206, pushing the software to the user to be pushed based on the software push classification result, including:

[0101] When the software push degree value is greater than or equal to the push threshold, the updated version of the cloud platform software will be pushed to the users to be pushed; when the software push degree value is less than the push threshold, the updated version of the cloud platform software will stop being pushed to the users to be pushed.

[0102] The recognition network can be established using Transformer (a neural network architecture based on the self-attention mechanism). The fusion network can be established using random forests, and the discriminant network can be established using a binary classification judger to determine whether to push the updated version to the cloud platform users.

[0103] For example, as shown in Figure 4, it is a schematic diagram of the model architecture used when pushing software, in which the cloud server can obtain user comment information of the updated version of the cloud platform software, and then convert the user comment information into a digital form that the model can recognize through the word embedding model, and then use the LSTM model to predict the software public opinion evaluation and output the evaluation level. The evaluation level includes 1 to 10, ten levels, where 1 represents the worst evaluation and 10 represents the best evaluation. At the same time, the cloud server inputs the process usage time T, the number of times C, and the software version V of the updated version into the Transformer model to obtain the output usage probability value, and then uses the normalization method to set the value range of the usage probability value to between 1 and 10. The formula can be used Normalize, where P is the normalized value, Pi is the output of the Transformer model, Pmax refers to the maximum value of the Transformer model output, which can be the maximum output of each version of the same software, and Pmin refers to the minimum value of the Transformer model output. Then combine the usage probability value and the evaluation level. During the combination process, set the Transformer output value P with a weight of 0.7, and the evaluation level with a weight of 0.3. Specifically, calculate and obtain the fusion feature M=0.7*P+0.3*N, and input the fusion feature into the random forest model to obtain the comprehensive usage probability value Q. The value range of Q is between 1 and 10, which represents the model evaluation of the software update version presented in combination with user usage habits and comments. Finally, the discriminator is used to determine whether the updated version needs to be pushed to the user cloud platform. The discriminator consists of a function with an input value of Q and an output of a Boolean variable Push. For each element q in Q, , the push threshold is preset to 5. Then when push is true, it means that the software version is pushed to the user cloud platform, and when push is false, it means that the software version is not pushed.

[0104] In the above embodiment, the evaluation information of the updated version is obtained by using the comment recognition model. Then, the software push classification model including the recognition network, the fusion network and the discriminant network is used to perform software push classification, which can improve the accuracy of the software push classification, that is, improve the accuracy of the obtained software push degree value, and then when the software push degree value is greater than or equal to the push threshold, the updated version of the cloud platform software is pushed to the user to be pushed. When the software push degree value is less than the push threshold, the updated version of the cloud platform software is stopped from being pushed to the user to be pushed, which can improve the accuracy of the software push.

[0105] It is understandable that before using the software push classification model, in order to improve the prediction accuracy of the model, the software push classification model needs to be trained. In an exemplary embodiment, the training of the software push classification model includes the following steps:

[0106] Obtaining training evaluation information corresponding to the training version, where the training evaluation information is obtained by identifying user comment information corresponding to the training version of the training software;

[0107] Obtaining training data corresponding to the training user and installation tags corresponding to the training data, wherein the training data includes historical version usage data of the training software and historical version usage data of related software, where the related software is cloud platform software related to the training software;

[0108] Input the training evaluation information and the training data into the initial software push classification model to obtain the output initial software push classification result;

[0109] The initial software push classification model is trained based on the initial software push classification results and the installation labels to obtain a software push classification model.

[0110] In this embodiment, the installation tag is used to characterize whether the historical version of the software is installed on the cloud platform, including the installation tag and the non-installation tag. The related software is the cloud platform software related to the training software, that is, the related software and the training software are the same type of software, for example, both are video software, both are instant messaging software, both are security software, etc. At this time, during training, the cloud server can use the historical version usage data of the training software and the historical version usage data of the related software to train the initial software push classification model, and when the training completion condition is met, the software push classification model is obtained.

[0111] In some embodiments, the server can train the comment recognition model and the software push classification model at the same time, that is, use the comment recognition model and the software push classification model together as the software push model, and then use the user comment information of the training version, the historical version usage data of the training software and the historical version usage data of related software as the input of the software push model, and use the information of whether the training version is installed as the training label for training to obtain the software push model, and then the software push model can be directly used to identify the installation probability value of the updated version. When the installation probability value exceeds the installation threshold, the updated version of the cloud platform software will be pushed.

[0112] In the above embodiment, the initial software push classification model is trained by using the historical version usage data of the training software and the historical version usage data of related software. When the training completion conditions are met, the software push classification model is obtained. Since the historical version usage data of the related software is used for training, the accuracy of the model training can be further improved.

[0113] In an exemplary embodiment, the updated versions include at least two, and the users to be pushed include at least two; S202, identifying evaluation information of the updated version according to user review information corresponding to the updated version of the cloud platform software, including:

[0114] A user review information matrix corresponding to the cloud platform software is obtained, wherein the review information matrix includes user review information of each updated version; and evaluation information of each updated version is identified according to the user review information matrix to obtain an evaluation matrix.

[0115] In this embodiment, the updated versions of the cloud platform software may include at least two, and then the cloud server obtains user comments on each updated version to obtain a user comment matrix, for example, the user comment matrix , where n is the number of comments and m is the number of software versions. represents the nth comment of the mth software version. Each software version takes the same number of comment information, which is then converted into a digital form that the model can recognize through the word embedding model, which can be recorded as At this time, the evaluation information of each updated version is identified according to the user review information matrix to obtain an evaluation matrix, which includes the evaluation level corresponding to each updated version.

[0116] S204, identifying the usage probability value of the updated version based on the historical version usage data of the cloud platform software by the user to be pushed, and classifying the software push based on the evaluation information and the usage probability value to obtain the software push classification result, including:

[0117] The historical version usage data corresponding to each user is obtained, and the usage probability value of the updated version is identified according to the historical version usage data corresponding to each user to obtain a usage probability value matrix; the software push classification is performed according to the evaluation matrix and the usage probability value matrix to obtain a software push classification result matrix, which includes the software push classification results of each updated version pushed to each user.

[0118] In this embodiment, the users to be pushed may include at least two, that is, the users of the cloud platform may include multiple users. At this time, when a software update version is detected, it is necessary to determine whether to push to multiple cloud platform users at the same time. That is, the cloud server performs weighted sum calculation on the evaluation matrix and the use probability value matrix to obtain a fusion feature matrix, and then the fusion feature matrix can be input into the fusion network for identification to obtain an output comprehensive use probability value matrix. For example, the obtained comprehensive use probability value matrix can be ,in, It represents the comprehensive usage probability value of the mth software version of the nth user. Each value in Q ranges from 1 to 10. Then the software push classification is judged by the discriminant network. For example, the software push is performed when the comprehensive usage probability value is greater than or equal to the push threshold 5, otherwise it is not pushed.

[0119] In some embodiments, in order to improve training efficiency, multiple training data can also be obtained at the same time, and training can be performed through the training data matrix to obtain the output software push classification result matrix, and the loss can be calculated according to the software push classification result matrix and the corresponding installation label matrix, and training can be performed using the gradient descent algorithm to obtain a trained model for software push prediction.

[0120] In the above embodiment, when there are multiple update versions and multiple users of the cloud platform, software push classification and identification can be performed on multiple update versions and multiple users at the same time in a matrix manner, so as to obtain software push classification results for pushing each update version to each user respectively, thereby improving the efficiency of obtaining software push classification results and further improving the efficiency of software push.

[0121] It is understandable that the user review information needs to be obtained before identifying the user review information. In an exemplary embodiment, as shown in FIG5 , before S202 , before identifying the evaluation information of the updated version according to the user review information corresponding to the updated version of the cloud platform software, the following steps are included:

[0122] S502, monitoring software update information of the cloud platform software.

[0123] The software database can be pre-established in the cloud server. The software installation packages frequently used by each user in the cloud platform, such as browsers, anti-virus software, encryption software, cloud disks, communication software, etc., are stored in the software database. Each software installation package will form a multi-version sequence to obtain the software database, which can be represented by a software matrix. , where S 11 To S 1n It represents n different versions of the same software from the oldest to the newest. Each software can maintain 30 formal stable versions.

[0124] At this time, a software real-time update platform can be installed in the cloud server to update the software in the software database in real time. The software real-time update platform is used to monitor the software update and release progress of each software supply platform in real time, that is, the software update information refers to the software update and release information.

[0125] S504: When the cloud platform software has corresponding software update information, download the software installation package of the corresponding updated version from the software supply platform based on the software update information.

[0126] In this embodiment, the software supply platform is a platform that provides software, which can be a software manufacturer platform, a software publishing platform, etc. The cloud server can periodically obtain the latest software version number from the software supply platform. When the obtained software version number is consistent with the latest software version number stored in the software database, it means that there is no corresponding software update information for the cloud platform software. At this time, the cloud server does not process it. When it is detected that the software version number is inconsistent with the latest software version number stored in the software database, it means that there is a corresponding updated version of the cloud platform software. At this time, the cloud server can download the software installation package of the corresponding updated version from the software supply platform.

[0127] For example, the cloud server obtains the latest version of software A as version a. Then, when the latest software version saved in the software database is a, no processing is performed. When the latest software version saved in the software database is not version a, the cloud server downloads the software installation package of version a from the software manufacturer, and can also download the software update package of version a.

[0128] S506, save the updated version of the software installation package to the software database, and detect the number of installation packages of the cloud platform software. When the number of installation packages exceeds a preset number, delete the software installation package of the cloud platform software corresponding to the target historical version.

[0129] It is understandable that in order to prevent the amount of data in the software database from being too large, the old version of the software installation package can be deleted. Among them, the number of software installation packages can be used to determine whether the software installation needs to be deleted. For example, when the software installation packages of the same software exceed the preset number of 30, the oldest version of the software installation package is deleted, thereby ensuring the safe and stable operation of the database.

[0130] In some embodiments, the cloud server may also calculate the amount of data in the software database, and when the amount of data exceeds a threshold, the software installation package corresponding to the oldest version of the cloud platform software may be deleted.

[0131] S508: Obtain corresponding user comment information according to the updated version.

[0132] In this embodiment, a software public opinion monitoring system is also established in the cloud server, through which user review information of the corresponding updated version of the software can be obtained. Among them, user review information of all real-time updated software can be collected from different software manufacturers or software publishing platforms in real time. Then, the user review information is identified by an artificial intelligence algorithm to obtain evaluation information of the updated version, thereby improving the accuracy of the obtained evaluation information.

[0133] The software push classification result includes a software push degree value. S206, based on the software push classification result, the software is pushed to the user to be pushed, including the steps of:

[0134] When the software push degree value is greater than or equal to the push threshold, the updated version of the software installation package is pushed to the terminal of the user to be pushed, so that the terminal installs the updated version of the software installation package.

[0135] The software push degree value is used to characterize the possibility of pushing software to the user to be pushed. The larger the software push degree value is, the greater the possibility of pushing the updated version of the software installation package to the terminal of the user to be pushed and installing it. The smaller the software push degree value is, the smaller the possibility of pushing the updated version of the software installation package to the terminal of the user to be pushed and installing it. The software push degree value can be a probability value or a value within a pre-set degree range. The push threshold is a pre-set threshold used to determine whether to push the updated version of the software installation package to the terminal of the user to be pushed for installation.

[0136] In some embodiments, when the software push degree value is less than the push threshold, the cloud server does not process it, that is, it does not install the updated version of the software including the terminal of the cloud platform user.

[0137] In some embodiments, when there are multiple software, the software update version with a software push degree value greater than or equal to the push threshold can be pushed to the cloud platform user's terminal, and the software update version with a software push degree value less than the push threshold will not be pushed.

[0138] In some embodiments, when the terminal is installing according to the updated version of the software installation package, a prompt message of whether to install can be displayed through a push window. When the push window can be displayed on the terminal for a preset time, for example, it can be displayed for 1 minute. At this time, the user can choose to accept the installation or refuse the installation. If the user clicks to accept the installation or does not process this pop-up window, the terminal can automatically decompress and install the updated version of the software, realizing the non-sensing installation of the software.

[0139] In the above embodiment, by using the comparison result between the software push degree value and the push threshold to determine whether to push the updated version of the software installation package to the terminal of the user to be pushed, the updated version of the software less than the push threshold is avoided from being pushed, which can improve the accuracy of software push.

[0140] In a specific embodiment, when pushing software, a software database server can be established first to maintain multiple versions of different software. Then, the software update and release progress of each software manufacturer is monitored in real time through the software real-time update platform, and the software database is updated in real time. At the same time, the cloud platform monitors the usage time, number of times used and version number of all software installed by users on the cloud platform. For example, the software can be recorded as S=S1, S2...Sn, the software version can be recorded as V=V1, V2...Vn, and the usage time of the software process can be recorded at the same time, and the usage time can be recorded as T=T1, T2...Tn. At the same time, the number of times the software is used is recorded, and the number of times used is recorded as C=C1, C2...Cn. And the user can view the number of times the software process is used and the usage time. At the same time, the cloud platform monitoring will also monitor whether the user chooses to install the software version, and whether the installation is set to two values ​​of 0 and 1. When the software version is installed, the corresponding installation value 1 is recorded. When the software version is not installed, the corresponding installation value 0 is recorded. Then, a software public opinion monitoring system and a public opinion monitoring model are established. User comments from various software platforms are collected through the software public opinion monitoring system, and the collected user comment information is evaluated or predicted by the public opinion monitoring model established by the LSTM algorithm, and the specific evaluation level value of the software version is output. Then, a design software push classification model can be established. The software push classification model is obtained by combining the Transformer model, random forest and a judge. Then, the process usage time and number of times of the software version can be used as input features, and the software version and the corresponding evaluation level value can be used as feature factors to input the software push classification model, and the information on whether the software version is worth being pushed can be output. Finally, a monitoring management server cluster and a monitoring management platform are established. The monitoring management server cluster stores and analyzes the information recorded by the user using the software, and pushes the new version of the software required by the user in real time to achieve non-sensing installation. Monitoring management can be achieved through the monitoring management platform. For example, as shown in Figure 6, a schematic diagram of the overall framework of software push is provided, and the overall framework of software push is applied to the cloud server cluster, wherein the comment information of each software is pulled in real time through the public opinion monitoring system, and the software stored in the software library server cluster is updated in real time through the real-time update platform. At the same time, through the monitoring and management platform, the monitoring and management platform can be established through the monitoring and management server cluster in the cloud service cluster. The monitoring and management server cluster can be aggregated by virtualizing multiple monitoring and management servers, and has a huge database built in for storing the usage data of user software versions of each cloud platform recorded by the cloud platform monitoring.At the same time, the monitoring and management server cluster also deploys a deep learning model for software push, which can push the new version of software required by users in real time. That is, it can realize personalized push based on the user's usage habits and comment information of the updated version, thereby improving the accuracy of software updates. At the same time, it prevents users from downloading and installing some software versions with poor quality or low security, improves system security during software updates, and realizes seamless software updates for users, without the need for users to update manually, thereby improving user experience.

[0141] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0142] Based on the same inventive concept, the embodiment of the present application also provides a software push device for implementing the software push method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more software push device embodiments provided below can refer to the limitations on the software push method above, and will not be repeated here.

[0143] In an exemplary embodiment, as shown in FIG. 7 , a software push device 700 is provided, comprising: an identification module 702, a classification module 704 and a push module 706, wherein:

[0144] Identification module 702, used to identify evaluation information of the updated version according to user review information corresponding to the updated version of the cloud platform software;

[0145] The classification module 704 is used to identify the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push according to the evaluation information and the usage probability value to obtain the software push classification result;

[0146] The push module 706 is used to push the software to the terminal of the user to be pushed based on the software push classification result.

[0147] In one embodiment, the identification module 702 includes:

[0148] A model recognition unit, used for inputting the updated version of the user review information into the review recognition model for recognition, and obtaining the output updated version of the evaluation information;

[0149] The classification module 704 includes:

[0150] The model classification unit is used to input the evaluation information and historical version usage data into the software push classification model, so as to identify the usage probability value of the updated version by using the historical version usage data through the software push classification model, and perform software push classification according to the evaluation information and the usage probability value to obtain the output software push classification result.

[0151] In one embodiment, the software push classification model includes a recognition network, a fusion network, and a discriminant network;

[0152] Model classification units, including:

[0153] The network classification unit is used to input the historical version usage data into the recognition network for identification, and obtain the output usage probability value of the updated version; fuse the evaluation information and the usage probability value to obtain the fusion feature; input the fusion feature into the fusion network for identification, and obtain the output comprehensive usage probability value of the updated version; input the comprehensive usage probability value into the discrimination network for software push classification discrimination, and obtain the output software push classification result, which includes the software push degree value;

[0154] The push module 706 includes:

[0155] A push unit, used to push the updated version of the cloud platform software to the user to be pushed when the software push degree value is greater than or equal to the push threshold;

[0156] The push stopping unit is used to stop pushing the updated version of the cloud platform software to the users to be pushed when the software push degree value is less than the push threshold.

[0157] In one embodiment, the software push device 700 further includes:

[0158] The model training module is used to obtain the training evaluation information corresponding to the training version, which is obtained by identifying the user comment information of the training software corresponding to the training version; obtain the training data corresponding to the training user and the installation label corresponding to the training data, the training data includes the historical version usage data of the training software and the historical version usage data of the related software, and the related software is the cloud platform software related to the training software; input the training evaluation information and the training data into the initial software push classification model to obtain the output initial software push classification result; train the initial software push classification model based on the initial software push classification result and the installation label to obtain the software push classification model.

[0159] In one embodiment, the update version includes at least two, and the users to be pushed include at least two; the identification module 702 includes:

[0160] A matrix acquisition unit, used to acquire a user review information matrix corresponding to the cloud platform software, wherein the review information matrix includes user review information of each updated version;

[0161] A matrix identification unit, used to identify evaluation information of each updated version according to the user comment information matrix to obtain an evaluation matrix;

[0162] The classification module 704 includes:

[0163] A usage probability identification unit is used to obtain the historical version usage data corresponding to each user, and identify the usage probability value of the updated version according to the historical version usage data corresponding to each user, so as to obtain a usage probability value matrix;

[0164] The push classification unit is used to classify software pushes according to the evaluation matrix and the use probability value matrix to obtain a software push classification result matrix, which includes the software push classification results of pushing each updated version to each user.

[0165] In one embodiment, the software push device 700 further includes:

[0166] The comment information acquisition module is used to monitor the software update information of the cloud platform software; when the cloud platform software has corresponding software update information, download the corresponding updated version of the software installation package from the software supply platform based on the software update information; save the updated version of the software installation package to the software database, and detect the number of installation packages of the cloud platform software. When the number of installation packages exceeds the preset number, delete the software installation package of the cloud platform software corresponding to the target historical version; obtain the corresponding user comment information according to the updated version;

[0167] The software push classification result includes a software push degree value. The push module 706 is further used to push the updated version of the software installation package to the terminal of the user to be pushed when the software push degree value is greater than or equal to the push threshold, so that the terminal can install the updated version of the software installation package.

[0168] Each module in the above software push device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0169] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG8. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store user comment information, historical version usage data of cloud platform software, cloud platform software installation package data, etc. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a software push method is implemented.

[0170] Those skilled in the art will understand that the structure shown in FIG. 8 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0171] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0172] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0173] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0174] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0175] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0176] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0177] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A software push method, characterized in that: Applied in a cloud server, the method includes: Identify evaluation information of the updated version according to user review information of the updated version of the cloud platform software; Identify the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push according to the evaluation information and the usage probability value to obtain the software push classification result; The software is pushed to the terminal of the user to be pushed based on the software push classification result.

2. The method according to claim 1, characterized in that The step of identifying evaluation information of the updated version according to user review information of the updated version of the cloud platform software includes: Inputting the user review information of the updated version into the review recognition model for recognition, and obtaining the evaluation information of the updated version as output; The identifying the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classifying the software push according to the evaluation information and the usage probability value to obtain the software push classification result includes: The evaluation information and the historical version usage data are input into a software push classification model so that the software push classification model uses the historical version usage data to identify the usage probability value of the updated version, and performs software push classification based on the evaluation information and the usage probability value to obtain an output software push classification result.

3. The method according to claim 2, characterized in that The software push classification model includes a recognition network, a fusion network and a discrimination network; The software push classification model uses the historical version usage data to identify the usage probability value of the updated version, and performs software push classification according to the evaluation information and the usage probability value to obtain an output software push classification result, including: Inputting the historical version usage data into a recognition network for identification, and obtaining an output usage probability value of the updated version; Fusing the evaluation information and the usage probability value to obtain a fusion feature; Inputting the fusion feature into the fusion network for identification, and obtaining an output comprehensive use probability value of the updated version; Inputting the comprehensive usage probability value into a discrimination network to perform software push classification discrimination, and obtaining an output software push classification result, wherein the software push classification result includes a software push degree value; The pushing of software to the to-be-pushed user based on the software push classification result includes: When the software push degree value is greater than or equal to the push threshold, the updated version of the cloud platform software is pushed to the user to be pushed; When the software push degree value is less than the push threshold, the updated version of the cloud platform software is stopped from being pushed to the user to be pushed.

4. The method according to any one of claims 2 to 3, characterized in that: The training of the software push classification model includes the following steps: Acquire training evaluation information corresponding to the training version, where the training evaluation information is obtained by identifying user review information corresponding to the training version of the training software; Acquire training data corresponding to the training user and installation tags corresponding to the training data, wherein the training data includes historical version usage data of the training software and historical version usage data of related software, and the related software is cloud platform software related to the training software; Inputting the training evaluation information and the training data into an initial software push classification model to obtain an output initial software push classification result; The initial software push classification model is trained based on the initial software push classification result and the installation label to obtain the software push classification model.

5. The method according to claim 1, characterized in that The updated versions include at least two, and the users to be pushed include at least two; the step of identifying the evaluation information of the updated version according to the user review information corresponding to the updated version of the cloud platform software includes: Obtaining a user review information matrix corresponding to the cloud platform software, wherein the review information matrix includes user review information of each updated version; Identify evaluation information of each updated version according to the user review information matrix to obtain an evaluation matrix; The identifying the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classifying the software push according to the evaluation information and the usage probability value to obtain the software push classification result includes: Acquire the historical version usage data corresponding to each user, and identify the usage probability value of the updated version according to the historical version usage data corresponding to each user, to obtain a usage probability value matrix; Software push classification is performed according to the evaluation matrix and the usage probability value matrix to obtain a software push classification result matrix, wherein the software push classification result matrix includes software push classification results of each updated version being pushed to each user.

6. The method according to claim 1, characterized in that Before identifying the evaluation information of the updated version according to the user review information corresponding to the updated version of the cloud platform software, the method includes: Monitoring software update information of the cloud platform software; When the cloud platform software has corresponding software update information, downloading the software installation package of the corresponding updated version from the software supply platform based on the software update information; The software installation package of the updated version is saved in a software database, and the number of installation packages of the cloud platform software is detected. When the number of installation packages exceeds a preset number, the software installation package of the cloud platform software corresponding to the target historical version is deleted; Obtain corresponding user review information according to the updated version; The software push classification result includes a software push degree value, and the software is pushed to the to-be-pushed user based on the software push classification result, including: When the software push degree value is greater than or equal to the push threshold, the updated version of the software installation package is pushed to the terminal of the user to be pushed, so that the terminal installs the updated version of the software installation package.

7. A software push device, characterized in that: The device comprises: An identification module, used to identify evaluation information of the updated version of the cloud platform software according to user review information corresponding to the updated version; A classification module, used to identify the usage probability value of the updated version according to the historical version usage data of the cloud platform software by the user to be pushed, and classify the software push according to the evaluation information and the usage probability value to obtain the software push classification result; The push module is used to push the software to the terminal of the user to be pushed based on the software push classification result.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images