A channel page recommendation method and device, a computer device and a storage medium

Abstract

Embodiments of the present application disclose a channel page recommendation method and device, computer equipment and a storage medium. The present application obtains browsing behavior data of a target client, and filters target browsing behavior data related to a candidate channel page from the browsing behavior data. The target browsing behavior data is sorted based on access time to obtain a page access sequence. A preset recall model is used to extract features from the target browsing behavior data to obtain short-term interest features of a target user. The short-term interest features are given corresponding weights based on the position of the target browsing behavior data in the page access sequence. The recall model is used to predict the interest degree of the target user in the candidate channel page based on the short-term interest features with the given weights. The channel page to be recommended is determined based on the interest degree of the target user in the candidate channel page, and the page identifier of the channel page to be recommended is recommended to the target user, thereby meeting the personalized needs of the user.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of channel page recommendation, and in particular to a channel page recommendation method and device, a computer device, and a storage medium. BACKGROUND

[0002] With the development of society, more and more web pages or clients recommend a certain channel page required by a user to the user, so as to quickly enter the channel page through the recommended channel page. However, the existing recommendation method is to recommend through some preset fixed commonly used channel pages, which cannot meet the personalized needs of users. SUMMARY

[0003] The present application provides a channel page recommendation method and device, a computer device, and a storage medium, which can be used in the field of financial technology or other related fields, and can meet the personalized needs of users.

[0004] The present application provides a channel page recommendation method, which comprises the following steps:

[0005] obtaining browsing behavior data of a target user to be recommended when accessing at least one access page of a target client within a preset time, and filtering target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data;

[0006] sorting the target browsing behavior data based on the access time of the target user to the at least one access page, to obtain a page access sequence;

[0007] extracting features of the target browsing behavior data in the page access sequence through a preset recall model, to obtain short-term interest features of the target user, and assigning a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence;

[0008] predicting the interest degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values through the recall model;

[0009] determining a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommending a page identifier of the channel pages to be recommended to the target user.

[0010] Correspondingly, the present application also provides a channel page recommendation device, which comprises:

[0011] The data screening module is configured to obtain browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, and screen target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data.

[0012] The sorting module is configured to sort the target browsing behavior data based on access time of the target user to the at least one access page, to obtain a page access sequence.

[0013] The feature extraction module is configured to extract features of the target browsing behavior data in the page access sequence based on a preset recall model, to obtain short-term interest features of the target user, and assign a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on a position of the target browsing behavior data in the page access sequence.

[0014] The prediction module is configured to predict an interest degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values by using the recall model.

[0015] The recommendation module is configured to determine a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommend a page identifier of the channel pages to be recommended to the target user.

[0016] Correspondingly, the embodiment of the present application further provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the channel page recommendation method provided by any of the embodiments of the present application.

[0017] Correspondingly, the embodiment of the present application further provides a storage medium, which stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the channel page recommendation method as above.

[0018] The embodiment of the application obtains the browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, filters target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data, sorts the target browsing behavior data based on the access time of the target user to the at least one access page, obtains a page access sequence, extracts features of the target browsing behavior data in the page access sequence through a preset recall model, obtains short-term interest features of the target user, and assigns a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence. The embodiment of the application predicts the interest degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values through the recall model, determines a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommends the page identifier of the channel pages to be recommended to the target user, so as to intelligently recommend the channel pages required by the user to the user by predicting the browsing data of the user within a preset time, thereby meeting the personalized needs of the user. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 A flowchart of a channel page recommendation method provided by an embodiment of the present application.

[0021] Figure 2 A structural diagram of a channel page recommendation device provided by an embodiment of the present application.

[0022] Figure 3 A structural diagram of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0024] The embodiment of the present application provides a channel page recommendation method and device, a storage medium and a computer device. Specifically, the channel page recommendation method of the embodiment of the present application can be executed by a computer device, which can be a server or a terminal or the like. The server can be a physical server, a server cluster or a distributed system formed by a plurality of physical servers, a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms. The terminal can be a smart phone, a desktop computer, a notebook computer, a tablet computer, and the like, but is not limited thereto. The terminal and the server can be directly or indirectly connected through wired or wireless communication, which is not limited in the present application.

[0025] For example, the computer device can be a terminal, which can acquire browsing behavior data of a target user to be recommended when the target user accesses at least one access page within a preset time, filter target browsing behavior data related to at least one candidate channel page of the target user from the browsing behavior data, sort the target browsing behavior data based on the access time of the target user to the at least one access page, obtain a page access sequence, extract features of short-term interest of the target user from the target browsing behavior data in the page access sequence through a preset recall model, assign a corresponding weight value to the short-term interest feature corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence, predict the interest degree of the target user to the at least one candidate channel page based on the short-term interest feature with the assigned weight value through the recall model, determine a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommend a page identifier of the channel page to be recommended to the target user.

[0026] Based on the above problems, the embodiment of the present application provides a channel page recommendation method, device, computer device and storage medium, which can meet the personalized needs of users.

[0027] The following will be described in detail. It should be noted that the order of the following embodiments is not limited as the preferred order of the embodiments.

[0028] The embodiment of the present application provides a channel page recommendation method, which can be executed by a terminal or a server. The embodiment of the present application takes the channel page recommendation method executed by a terminal as an example for description.

[0029] Please refer to Figure 1 , Figure 1A flowchart of a channel page recommendation method provided in an embodiment of the present application. The specific flow of the channel page recommendation method can be as follows:

[0030] 101. Obtain browsing behavior data of a target user to be recommended when accessing at least one access page of a target client within a preset time, and filter target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data.

[0031] The preset time is a preset time before the current time (including the current time), and the specific preset time value can be set according to requirements.

[0032] The access page is a page currently accessed by the target user in the target client, for example, a page related to “life discount”. The browsing behavior data of the user to the access page includes but is not limited to access time, specific access information, etc. Since not all data in the browsing behavior data of a page is related to the subsequently recommended channel page, it is necessary to determine which data in the browsing behavior data of the access page is related to the candidate channel page based on the recommendable candidate channel page, so as to filter.

[0033] The channel page is a fixed channel page corresponding to each channel in the client, for example, if the access page is a page related to “life discount”, the corresponding channel page can be the channel page of “life discount”.

[0034] The target client can be displayed in the terminal held by the user in the form of an APP, or in the form of a web page, which is not limited here.

[0035] In this embodiment, the terminal obtains the browsing data of the target user to be recommended within a short time, so as to recommend a channel page for the target user based on the browsing behavior data, thereby recommending the target user based on the instant interest of the target user, and improving the personalized needs of the target user.

[0036] 102. Sort the target browsing behavior data based on the access time of the target user to the at least one access page, to obtain a page access sequence.

[0037] In this embodiment, since the recommendation is made for the target user based on the browsing data of the target user in a short time, in order to better achieve the recommendation effect, the access time of the target user browsing each access page needs to be determined, so that the page order of the target user browsing the access page can be obtained based on the access time, the target flow column behavior data of each page is sorted based on the page order, and a page access sequence is composed. The page access sequence includes the target browsing behavior data arranged in the order of browsing time, so that the user is recommended based on the target browsing behavior data in a specific order, so as to improve the accuracy of recommending channel pages to the user.

[0038] In some embodiments, the above sorting of the target browsing behavior data based on the access time of the target user to the at least one access page to obtain the page access sequence can include: determining the target access time of the access page where the target browsing behavior data is located based on the access time of the target user to the at least one access page; and sorting the target browsing behavior data from early to late based on the target access time corresponding to the target browsing behavior data, to obtain the page access sequence.

[0039] 103. By the preset recall model, the target browsing behavior data in the page access sequence is subjected to feature extraction to obtain the short-term interest feature of the target user, and each target browsing behavior data is assigned a corresponding weight value based on the position of each target browsing behavior data in the page access sequence.

[0040] In this embodiment, the target browsing behavior data is subjected to feature extraction by the recall model to obtain the short-term interest feature corresponding to the target browsing behavior data. Since the target browsing behavior data in the page sequence is sorted according to the access time, the short-term interest feature corresponding to the target browsing behavior data needs to be assigned a weight value based on the access time, so as to realize accurate recommendation of channel pages to the user based on the short-term interest feature after the weight value is assigned.

[0041] It can be understood that since the data accessed at different times has different degrees of influence on the user, and the short-term behavior of the user has time sequence and correlation, generally the closer the behavior to the recommendation time, the higher the importance, so as to perceive the immediate interest of the user in real time, the closer the access data to the recommendation time, the higher the importance, and the higher the weight value assigned to the access data.

[0042] In some embodiments, if the page access sequence is obtained by sorting the target browsing behavior data in the order of first to last, after assigning a corresponding weight to the short-term interest feature corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence, the result presented is that the weight of the short-term interest feature corresponding to the target browsing behavior data in the front position is less than the weight of the short-term interest feature corresponding to the target browsing behavior data in the rear position.

[0043] In some embodiments, due to the time sequence and correlation between the behavior data of the user, the feature extraction of the target browsing behavior data in the page access sequence by the preset recall model to obtain the short-term interest feature of the target user can specifically include: through the recall model, the terminal extracts the relative browsing order between the target browsing behavior data in the page access sequence and the association between the access pages where the target browsing behavior data in the page access sequence is located to obtain the short-term interest feature of the target user.

[0044] In some embodiments, before the feature extraction of the target browsing behavior data in the page access sequence by the preset recall model to obtain the short-term interest feature of the target user, a corresponding recall model needs to be generated, for example, a recall model specific to a certain user, that is, the recall model is trained based on the historical data of the user, thereby generating a recall model specific to the user. For another example, the user attributes of the user are obtained, which include but are not limited to the user gender, the user age, etc., thereby based on the user attributes, the historical data of the users with consistent attributes are selected to train the recall model, thereby generating a recall model specific to the users of this type.

[0045] Specifically, it can include: obtaining a sample set, each sample in the sample set including historical browsing behavior data of a historical access user when accessing at least one access page of the target client within the preset time, and the label of the sample being a channel page identifier, the channel page identifier being used to indicate the channel page accessed by the historical access user. Then, according to the access time of the historical access user of each sample in the sample set, the sample set is divided into a training set, a test set and a validation set, and the recall model is constructed through the training set, the test set and the validation set. Wherein, the historical browsing behavior data can also be stored in a sequence, which can be specifically referred to the above step 102, and the data length of the sequence can be set to 100, that is, there are 100 historical browsing behavior data.

[0046] Exemplarily, the last preset number of samples of the access time can be taken as a test set, the last preset number of samples of the access time in samples other than those in the test set can be taken as a validation set, and the remaining samples can be taken as a training set.

[0047] Specifically, since the next page to be accessed by the user is recommended to the user through the short-term behavior of the user in the embodiment, and the short-term behavior of the user has time sequence and correlation, a sequence recommendation model based on self-attention mechanism (denoted as SASRec) can be used to construct the recall model. The recall model includes an embedding layer, a Self-Attention layer and a prediction layer. The embedding layer includes item embedding and position embedding. The recall model can learn the correlation between items and the time sequence.

[0048] In some embodiments, the construction of the recall model through the training set, the test set and the validation set includes: training the recall model through the samples in the training set, and obtaining a loss function corresponding to the recall model; if the loss function corresponding to the recall model does not meet a preset condition, continuing to train the recall model through the samples in the sample set until the loss function corresponding to the recall model meets the preset condition, and stopping the training of the recall model.

[0049] Specifically, in the process of training the recall model, a negative sample can be randomly generated for each sample in the training process, so as to calculate a corresponding cross-entropy loss function, so as to prevent gradient explosion and overfitting in the training process based on the cross-entropy loss function. In addition, the recall model can be further gradient clipped and dropout rate adjusted.

[0050] In some embodiments, since there can be some noise data in the browsing behavior data of the user on the page, i.e., data irrelevant to the construction of the recall model, such as irrelevant data reported by user login and other buried points, i.e., non-real user behavior data, the browsing behavior data constituting the sample set needs to be filtered to retain relevant data before the sample set is obtained.

[0051] Specifically, it can include: obtaining historical browsing behavior data of at least one historical access user when accessing a page of the target client; filtering the historical browsing behavior data corresponding to the at least one historical access user based on at least one of the number of access pages accessed by the historical access user, the page type of the access page and the data attribute of the historical browsing behavior data, to obtain filtered historical browsing behavior data; and generating the sample set according to the filtered historical browsing behavior data.

[0052] For example, if the number of pages accessed by the user is less than a preset threshold, such as less than 3, it indicates that the user is a user with low access frequency, and therefore the browsing behavior data of the user is not used as sample data.

[0053] For example, if the page type of the accessed page is a high-frequency type, such as the "first", "my", and "search page" of the target client, the accessed page not only has no effect on the access channel page, but also can cause the influence of false prediction, and therefore the browsing behavior data corresponding to the accessed page is not used as sample data.

[0054] For example, if the browsing time in the data attribute of the set historical browsing behavior data is very short, it indicates that the page or the browsing behavior data in the page can be generated due to user error, and therefore the historical browsing behavior data is not used as sample data.

[0055] 104. Through the recall model, the interest degree of the target user to the at least one candidate channel page is predicted based on the assigned value of the short-term interest feature.

[0056] In this embodiment, the terminal can predict the interest degree of the target user to the at least one candidate channel page based on the assigned value of the short-term interest feature corresponding to the at least one target browsing behavior data through the recall model, and determine the to-be-recommended channel page in the at least one candidate channel page recommended to the target user based on the interest degree of the target user to the at least one candidate channel page.

[0057] In some embodiments, since the recall model includes an embedding layer, a Self-Attention layer, and a prediction layer, and the position embedding in the embedding layer can obtain the maximum value of the short-term interest feature corresponding to the last target browsing behavior data in the weight values of the short-term interest features corresponding to each target browsing behavior data in the page access sequence arranged from early to late, that is, the full score of the short-term interest feature corresponding to the last target browsing behavior data accessed by the user is the maximum, and therefore when prediction is performed through the prediction layer, the embedding vector corresponding to the last target browsing behavior data in the Self-Attention layer can be used as the user short-term interest representation to match the embedding vectors corresponding to the at least one candidate channel, and thus the matching values corresponding to each candidate channel, that is, the interest degrees of the target user to each candidate channel, are obtained.

[0058] 105、based on the degree of interest of the target user in the at least one candidate channel page, determine a preset number of channel pages to be recommended, and recommend the page identifiers of the channel pages to be recommended to the target user.

[0059] In this embodiment, the terminal can sort each candidate channel page in descending order of the degree of interest of the target user in the candidate channel page, and select a preset number of channel pages in the front of the sorting as the channel pages to be recommended. The preset number can be set according to user demand, for example, 3, so as to perceive the immediate interest of the user in real time through the behavior of the user in real-time browsing the page and the relative order of the browsing behavior, and make recommendations based on the immediate interest, thereby improving the recommendation effect and meeting the personalized needs of the user.

[0060] Specifically, the terminal can display the page identifiers of the channel pages to be recommended in the page currently browsed by the target user, so that the target user can enter the corresponding channel page by clicking the page identifier in the currently browsed page; or the terminal can display the page identifiers of the channel pages to be recommended in the "home page" of the target client currently browsed by the target user, so that the target user can enter the corresponding channel page by clicking the page identifier in the "home page".

[0061] The embodiment of the application discloses a channel page recommendation method, which comprises: obtaining browsing behavior data of a target user to be recommended when accessing at least one access page of a target client within a preset time, and screening target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data; sorting the target browsing behavior data based on the access time of the target user to the at least one access page to obtain a page access sequence; extracting features of the target browsing behavior data in the page access sequence through a preset recall model to obtain short-term interest features of the target user, and assigning a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence; predicting the degree of interest of the target user in the at least one candidate channel page based on the short-term interest features with the assigned weight values through the recall model; and determining a preset number of channel pages to be recommended based on the degree of interest of the target user in the at least one candidate channel page, and recommending the page identifiers of the channel pages to be recommended to the target user. In this way, the channel pages required by the user can be intelligently recommended to the user by predicting the browsing data of the user within a preset time, so as to meet the personalized needs of the user.

[0062] To facilitate better implementation of the channel page recommendation method provided in the embodiments of the present application, the embodiments of the present application further provide a channel page recommendation device based on the above channel page recommendation method. The meanings of the terms are the same as those in the above channel page recommendation method, and the specific implementation details can be referred to the description in the method embodiments.

[0063] Please refer to Figure 2 , Figure 2 A structural block diagram of a channel page recommendation device provided in the embodiments of the present application is shown in FIG. 2. The device includes:

[0064] A data screening module 201 is configured to obtain browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, and screen target browsing behavior data related to at least one candidate channel page of the target client from the above browsing behavior data.

[0065] An ordering module 202 is configured to order the target browsing behavior data based on the access time of the target user to the at least one access page, to obtain a page access sequence.

[0066] A feature extraction module 203 is configured to extract features of the target browsing behavior data in the page access sequence by a preset recall model, to obtain short-term interest features of the target user, and to assign a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence.

[0067] A prediction module 204 is configured to predict the interest degree of the target user to the at least one candidate channel page based on the short-term interest features with assigned weight values by the recall model.

[0068] A recommendation module 205 is configured to determine a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and to recommend the page identifiers of the channel pages to be recommended to the target user.

[0069] In some embodiments, the channel page recommendation device further includes:

[0070] An acquisition module is configured to obtain a sample set, wherein each sample in the sample set includes historical browsing behavior data of a historical access user to the at least one access page of the target client within the preset time, and the label of the sample is a channel page identifier, which is used to indicate the channel page accessed by the historical access user.

[0071] The dividing module is configured to divide the sample set into a training set, a test set, and a verification set according to the access time of the historical access user of each sample in the sample set.

[0072] The model constructing module is configured to construct the recall model by using the training set, the test set, and the verification set.

[0073] In some embodiments, the model constructing module includes:

[0074] The first training unit is configured to train the recall model by using the samples in the training set and obtain a loss function corresponding to the recall model.

[0075] The second training unit is configured to continue training the recall model by using the samples in the sample set if the loss function corresponding to the recall model does not meet the preset condition, and stop training the recall model when the loss function corresponding to the recall model meets the preset condition.

[0076] In some embodiments, the channel page recommendation device further includes:

[0077] The data obtaining module is configured to obtain historical browsing behavior data of at least one historical access user when the historical access user accesses a target page of the target client;

[0078] The data processing module is configured to filter the historical browsing behavior data corresponding to the at least one historical access user based on at least one of the number of accessed pages of the historical access user, the page type of the accessed page, and a data attribute of the historical browsing behavior data, and obtain filtered historical browsing behavior data.

[0079] The generating module is configured to generate the sample set according to the filtered historical browsing behavior data.

[0080] In some embodiments, the sorting module 202 includes:

[0081] The time determining unit is configured to determine a target access time of an accessed page where the target browsing behavior data is located based on the access time of the target user to the at least one accessed page.

[0082] The sorting unit is configured to sort the target browsing behavior data from early to late based on the target access time corresponding to the target browsing behavior data, and obtain the page access sequence.

[0083] In some embodiments, the weight of the short-term interest feature corresponding to the position-in-advance target browsing behavior data is smaller than the weight of the short-term interest feature corresponding to the position-later target browsing behavior data.

[0084] In some embodiments, the feature extraction module 203 comprises:

[0085] The feature extraction unit is configured to extract features of the target browsing behavior data in the page access sequence and the association between the target browsing behavior data in the page access sequence by using the recall model, to obtain the short-term interest features of the target user.

[0086] The embodiment of the present application discloses a channel page recommendation device. The data filtering module 201 is configured to obtain browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, and filter target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data. The sorting module 202 is configured to sort the target browsing behavior data based on the access time of the target user to the at least one access page, to obtain a page access sequence. The feature extraction module 203 is configured to extract features of the target browsing behavior data in the page access sequence by using a preset recall model, to obtain short-term interest features of the target user, and assign a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence. The prediction module 204 is configured to predict the interest degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values by using the recall model. The recommendation module 205 is configured to determine a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommend page identifiers of the channel pages to be recommended to the target user. In this way, the channel pages required by the user are intelligently recommended to the user by predicting the browsing data of the user within a preset time, to meet the personalized needs of the user.

[0087] Correspondingly, the embodiment of the present application also provides a computer device, which can be a terminal. As shown in Figure 3 Figure 3 The computer device provided by the embodiment of the present application is a structure diagram. The computer device 300 includes a processor 301 with one or more processing cores, a memory 302 with one or more computer readable storage media, and a computer program stored on the memory 302 and executable on the processor. The processor 301 is electrically connected to the memory 302. Those skilled in the art can understand that the computer device structure shown in the figure does not constitute a limitation on the computer device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0088] ​The processor 301 is a control center of the computer device 300, connects all parts of the computer device 300 through various interfaces and lines, executes various functions of the computer device 300 and processes data by running or loading software programs and / or modules stored in the memory 302 and calling data stored in the memory 302, thereby monitoring the computer device 300 as a whole.

[0089] In the embodiment of the present application, the processor 301 in the computer device 300 loads the instructions corresponding to the processes of one or more application programs into the memory 302, and runs the application programs stored in the memory 302 by the processor 301, so as to realize various functions according to the following steps:

[0090] Obtain the browsing behavior data of the target user to be recommended when accessing at least one access page of the target client within a preset time, and filter target browsing behavior data related to at least one candidate channel page of the target client from the above browsing behavior data;

[0091] Sort the target browsing behavior data based on the access time of the target user to the at least one access page, to obtain a page access sequence;

[0092] Extract features from the target browsing behavior data in the page access sequence through a preset recall model, to obtain short-term interest features of the target user, and assign a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on the position of each target browsing behavior data in the page access sequence;

[0093] Predict the interest degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values through the recall model;

[0094] Determine a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommend the page identifier of the channel pages to be recommended to the target user.

[0095] The specific implementation of each operation can be referred to the foregoing embodiments, which will not be described here.

[0096] Optionally, as shown in Figure 3 The computer device 300 further includes a touch display screen 303, a radio frequency circuit 304, an audio circuit 305, an input unit 306, and a power supply 307. The processor 301 is electrically connected with the touch display screen 303, the radio frequency circuit 304, the audio circuit 305, the input unit 306, and the power supply 307, respectively. Those skilled in the art can understand that Figure 3The computer device structure shown in the figure is not a limitation of the computer device, and can include more or fewer components than shown, or combine certain components, or arrange different components.

[0097] The touch display screen 303 can be used to display a graphical user interface and receive operation instructions generated by user acting on the graphical user interface. The touch display screen 303 can include a display panel and a touch panel. The display panel can be used to display messages input by the user or provided to the user and various graphical user interfaces of the computer device, which can be composed of graphics, text, icons, videos and any combination thereof. Optionally, the display panel can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. The touch panel can be used to collect touch operations (such as operations of the user using a finger, a stylus or any suitable object or accessory on or near the touch panel) of the user thereon or therearound, and generate corresponding operation instructions, and the operation instructions execute corresponding programs. Optionally, the touch panel can include two parts of a touch detection device and a touch controller. The touch detection device detects the touch position of the user and detects signals caused by the touch operation, and transmits the signals to the touch controller; the touch controller receives the touch message from the touch detection device, and converts it into touch coordinates, and sends it to the processor 301, and can receive commands from the processor 301 and execute them. The touch panel can cover the display panel, and when the touch panel detects a touch operation thereon or therearound, it transmits to the processor 301 to determine the type of the touch event, and then the processor 301 provides corresponding visual output on the display panel according to the type of the touch event. In the embodiments of the present application, the touch panel and the display panel can be integrated into the touch display screen 303 to realize input and output functions. However, in some embodiments, the touch panel and the touch panel can realize input and output functions as two independent components. That is, the touch display screen 303 can also realize input functions as part of the input unit 306.

[0098] The radio frequency circuit 304 can be used to transceive radio frequency signals to establish wireless communication with network devices or other computer devices, and transceive signals between network devices or other computer devices.

[0099] The audio circuit 305 can be used to provide an audio interface between the user and the computer device through a speaker and a microphone. The audio circuit 305 can convert received audio data into an electrical signal and transmit the electrical signal to the speaker for conversion into an audible signal output by the speaker. On the other hand, the microphone can collect a sound signal and convert the sound signal into an electrical signal, which is received by the audio circuit 305 and converted into audio data. The audio data is then output to the processor 301 for processing, and then transmitted to another computer device through the radio frequency circuit 304, or output to the memory 302 for further processing. The audio circuit 305 can also include a jack for a headset to provide communication between the headset and the computer device.

[0100] The input unit 306 can be used to receive inputted numbers, character messages or user feature messages (e.g. fingerprint, iris, face message, etc.), and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0101] The power supply 307 is used to supply power to various components of the computer device 300. Optionally, the power supply 307 can be logically connected to the processor 301 through a power management system, so that the power management system can be used to manage charging, discharging and power consumption management, etc. The power supply 307 can also include one or more DC or AC power sources, recharging systems, power failure detection circuits, power converters or inverters, power status indicators, etc.

[0102] Although Figure 3 The computer device 300 can also include a camera, a sensor, a wireless fidelity module, a Bluetooth module, etc. which are not shown in the embodiments and will not be described herein.

[0103] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0104] From the above, the computer device provided in the embodiment can obtain browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, and filter target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data; sort the target browsing behavior data based on access time of the target user to the at least one access page, to obtain a page access sequence; perform feature extraction on the target browsing behavior data in the page access sequence through a preset recall model, to obtain short-term interest features of the target user, and assign a corresponding weight value to a short-term interest feature corresponding to each target browsing behavior data based on a position of each target browsing behavior data in the page access sequence; predict an interested degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values through the recall model; and determine a preset number of channel pages to be recommended based on the interested degree of the target user to the at least one candidate channel page, and recommend a page identifier of the channel pages to be recommended to the target user.

[0105] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by related hardware controlled by the instructions, which can be stored in a computer readable storage medium and loaded and executed by a processor.

[0106] To this end, the embodiment of the present application provides a computer readable storage medium, which stores a plurality of computer programs capable of being loaded by a processor to execute the steps in any channel page recommendation method provided by the embodiment of the present application. For example, the computer program can execute the following steps:

[0107] Obtain browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, and filter target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data;

[0108] Sort the target browsing behavior data based on access time of the target user to the at least one access page, to obtain a page access sequence;

[0109] Perform feature extraction on the target browsing behavior data in the page access sequence through a preset recall model, to obtain short-term interest features of the target user, and assign a corresponding weight value to a short-term interest feature corresponding to each target browsing behavior data based on a position of each target browsing behavior data in the page access sequence;

[0110] The interest degree of the target user in the at least one candidate channel page is predicted based on the short-term interest features with assigned weights through the recall model;

[0111] Based on the interest degree of the target user in the at least one candidate channel page, a preset number of channel pages to be recommended are determined, and the page identifiers of the channel pages to be recommended are recommended to the target user.

[0112] The specific implementation of each operation can refer to the foregoing embodiments, and will not be described here.

[0113] The storage medium can include a read-only memory (ROM, Read Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, etc.

[0114] Due to the computer program stored in the storage medium, the steps of any channel page recommendation method provided by the embodiments of the present application can be executed, and thus the beneficial effects of any channel page recommendation method provided by the embodiments of the present application can be achieved. Details can be found in the foregoing embodiments, and will not be described here.

[0115] The above describes in detail a channel page recommendation method, device, computer device and storage medium provided by the embodiments of the present application. The principle and implementation manner of the present application are described by applying specific examples. The above embodiment description is only used to help understand the method and its core idea of the present application. Meanwhile, for those skilled in the art, the specific implementation manner and application range can be changed according to the idea of the present application. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method of recommending a channel page, characterized by, The method comprises: obtaining browsing behavior data of a target user to be recommended when the target user accesses at least one access page of a target client within a preset time, and screening target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data; sorting the target browsing behavior data based on access time of the target user to the at least one access page, to obtain a page access sequence; extracting features of the target browsing behavior data in the page access sequence through a preset recall model, to obtain short-term interest features of the target user, and assigning a corresponding weight value to the short-term interest features corresponding to each target browsing behavior data based on a position of each target browsing behavior data in the page access sequence; predicting an interest degree of the target user to the at least one candidate channel page based on the short-term interest features with the assigned weight values through the recall model; determining a preset number of channel pages to be recommended based on the interest degree of the target user to the at least one candidate channel page, and recommending a page identifier of the channel pages to be recommended to the target user.

2. The method of claim 1, wherein, Before the feature extraction of the target browsing behavior data in the page access sequence through the preset recall model, the method further comprises: obtaining a sample set, wherein each sample in the sample set comprises historical browsing behavior data of a historical access user to the target client when the historical access user accesses at least one access page of the target client within the preset time, and a label of the sample is a channel page identifier, which is used to indicate a channel page accessed by the historical access user; dividing the sample set into a training set, a test set and a validation set according to access time of the historical access user of each sample in the sample set; constructing the recall model through the training set, the test set and the validation set.

3. The method of claim 2, wherein, The construction of the recall model through the training set, the test set and the validation set comprises: training the recall model through samples in the training set, and obtaining a loss function corresponding to the recall model; if the loss function corresponding to the recall model does not meet a preset condition, continuing to train the recall model through samples in the sample set until the loss function corresponding to the recall model meets the preset condition, and stopping the training of the recall model.

4. The method of claim 2, wherein, Before the obtaining of the sample set, the method further comprises: obtaining historical browsing behavior data of at least one historical access user when the historical access user accesses an access page of the target client; screening historical browsing behavior data corresponding to at least one historical access user based on at least one of the following: a number of access pages accessed by the historical access user, a page type of the access page, and a data attribute of the historical browsing behavior data, to obtain screened historical browsing behavior data; generating the sample set according to the screened historical browsing behavior data.

5. The method of claim 1, wherein, The sorting of the target browsing behavior data based on the access time of the target user to the at least one access page, to obtain a page access sequence, comprises: determine a target access time of an access page where the target browsing behavior data is located based on an access time of the target user to the at least one access page; sort the target browsing behavior data from early to late based on the target access time corresponding to the target browsing behavior data, to obtain the page access sequence.

6. The method of claim 5, wherein, The weight value of the short-term interest feature corresponding to the target browsing behavior data in the front position is smaller than the weight value of the short-term interest feature corresponding to the target browsing behavior data in the rear position.

7. The method according to any one of claims 1 to 6, characterized in that, The feature extraction of the target browsing behavior data in the page access sequence is performed by the preset recall model to obtain the short-term interest feature of the target user, including: The feature extraction of the relative browsing order between the target browsing behavior data in the page access sequence and the association between the access pages where the target browsing behavior data in the page access sequence is located is performed by the recall model to obtain the short-term interest feature of the target user.

8. A channel page recommendation apparatus characterized by comprising: The device comprises: A data screening module is configured to obtain browsing behavior data of a target user to be recommended when accessing at least one access page of a target client within a preset time, and screen target browsing behavior data related to at least one candidate channel page of the target client from the browsing behavior data; An ordering module is configured to order the target browsing behavior data based on an access time of the target user to the at least one access page, to obtain a page access sequence; A feature extraction module is configured to perform feature extraction of the target browsing behavior data in the page access sequence by a preset recall model, to obtain a short-term interest feature of the target user, and assign a corresponding weight value to the short-term interest feature corresponding to each target browsing behavior data based on a position of each target browsing behavior data in the page access sequence; A prediction module is configured to predict an interested degree of the target user to the at least one candidate channel page based on the short-term interest feature with the assigned weight value by the recall model; A recommendation module is configured to determine a preset number of channel pages to be recommended based on the interested degree of the target user to the at least one candidate channel page, and recommend a page identifier of the channel page to be recommended to the target user.

9. A computer device, comprising: The memory, the processor, and the computer program stored in the memory and running on the processor are included, wherein the processor implements the channel page recommendation method of any one of claims 1 to 7 when executing the program.

10. A storage medium, characterized by The storage medium stores a plurality of instructions, and the instructions are adapted to be loaded by the processor to execute the channel page recommendation method of any one of claims 1 to 7.

Images