Information recommendation method, device and computer readable storage medium

Abstract

The embodiments of the present application disclose an information recommendation method, device and computer-readable storage medium; a project window array containing the window sequence number of each target project most recently displayed in the historical period can be generated according to the time parameter of each target project in the historical period, and a project frequency array containing the project frequency value displayed by each target project in the historical period can be generated, and the project frequency array can be slidingly updated to further determine the project weight under the project frequency array of the current period; then, the behavior weight of the target object is determined based on the project weight, and the interest score of the target object for each project is determined according to the behavior weight and the time decay coefficient, and information recommendation is made according to the interest score; in this way, the frequency of the target project is estimated by sliding update of the array, which saves a large amount of memory space, improves calculation efficiency, and improves information push efficiency.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an information recommendation method, device and computer-readable storage medium. Background Art

[0002] With the development of Internet information technology, people's information acquisition efficiency has been improved, bringing people a rich experience. However, due to the wide variety of Internet information, such as shopping, news and current affairs, travel guides and other information services, multiple types of information will lead to information overload, and invalid information can easily cause trouble to users. In order to push information of interest to users, related technologies first determine the user's interest points when pushing information, specifically, determine the user's interest points for information based on the user's real-time behavior data and historical behavior data, and push Internet information based on the determined interest points.

[0003] During the research and practice of the prior art, the inventors of the present application discovered that when the prior art pushes information, in order to ensure the accuracy of the points of interest in the information push, it is necessary to use a large amount of historical behavior data to calculate and determine the user's points of interest. The calculation process is cumbersome and requires a large amount of storage and computing resources of the system, which affects the computing performance of the system and reduces the efficiency of information push. Summary of the invention

[0004] The embodiments of the present application provide an information recommendation method, device and computer-readable storage medium, which can improve the efficiency of information push.

[0005] The present application provides an information recommendation method, including:

[0006] Collect the time parameters displayed by each target project in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target project;

[0007] Based on the preset iteration value and the item window array, the item frequency value corresponding to each target item in the item frequency array is updated, and the item weight corresponding to each target item is determined according to the updated item frequency array;

[0008] Obtaining the selection frequency of each target item by the target object in the historical period, and calculating the behavior weight of each target item by the target object according to the item weight and the selection frequency corresponding to each target item;

[0009] Determine a time decay coefficient according to a preset decay coefficient and the historical period, and determine the interest score of the target object for each target item according to the time decay coefficient and the behavior weight;

[0010] The information of the items to be recommended is determined according to the interest score of each target item, and the information of the items to be recommended is sent to the target object.

[0011] Accordingly, an embodiment of the present application provides an information recommendation device, including:

[0012] A generating unit, used for collecting time parameters displayed by each target project in a historical period, and generating a project window array and a project frequency array according to the time parameters corresponding to each target project;

[0013] An updating unit, configured to update the item frequency value corresponding to each target item in the item frequency array based on a preset iteration value and the item window array, and determine the item weight corresponding to each target item according to the updated item frequency array;

[0014] an acquisition unit, configured to acquire the selection frequency of each target item by the target object in the historical period, and calculate the behavior weight of each target item by the target object according to the item weight and the selection frequency corresponding to each target item;

[0015] A determination unit, configured to determine a time decay coefficient according to a preset decay coefficient and the historical period, and determine an interest score of the target object for each target item according to the time decay coefficient and the behavior weight;

[0016] The sending unit is used to determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

[0017] In some embodiments, the updating unit is further configured to:

[0018] Extracting the item window sequence number of each target item in the item window array, and extracting the item frequency value of each target item in the item frequency array;

[0019] Performing a weighted summation of the project window sequence number and the project frequency value of each target project according to a preset iteration value to obtain an updated project frequency value corresponding to each target project;

[0020] The item frequency value corresponding to each target item in the item frequency array is updated according to the updated item frequency value corresponding to each target item.

[0021] In some embodiments, the updating unit is further configured to:

[0022] Performing weighted processing on the item frequency value of each target item according to a preset iteration value to obtain a weighted item frequency value;

[0023] Obtain the target time window sequence number currently displayed, and perform weighted processing on the sequence number difference between the target time window sequence number and the project window sequence number according to the preset iteration value to obtain a weighted time window value;

[0024] The weighted item frequency value and the weighted time window value are summed to obtain an updated item frequency value corresponding to each target item.

[0025] In some implementations, the updating unit is further configured to:

[0026] Selecting the updated item frequency value corresponding to each target item from the updated item frequency array;

[0027] The project weight corresponding to each target project is determined according to the updated project frequency value corresponding to each target project.

[0028] In some embodiments, the generating unit is further configured to:

[0029] Generate a corresponding initial project window array and an initial project frequency array according to the time parameters corresponding to each target project;

[0030] Acquire a project identifier corresponding to each target project, and perform mapping processing on the project identifier corresponding to each target project to obtain a project mapping value corresponding to each target project;

[0031] The initial project window array and the initial project frequency array are respectively labeled according to the project mapping value corresponding to each target project to obtain a project window array and a project frequency array.

[0032] In some embodiments, the generating unit is further configured to:

[0033] Acquire a preset hash function corresponding to each target item from a preset function database;

[0034] The project identifier corresponding to each target project is hash mapped according to the preset hash function corresponding to each target project to obtain the project mapping value corresponding to each target project.

[0035] In some embodiments, the determining unit is further configured to:

[0036] Obtaining a target time for the current period, and obtaining a time difference between the target time and the historical period;

[0037] The time-effect decay coefficient of the current time period is determined according to the preset decay coefficient and the time difference.

[0038] In some implementation modes, the sending unit is further configured to:

[0039] Based on the interest score of each target item, determining an interest score sequence including a plurality of target items;

[0040] Determine the project capacity for the current time window;

[0041] Based on the interest score sequence, a target project corresponding to the project capacity is selected, and the target project corresponding to the project capacity is determined as the project information to be recommended.

[0042] In addition, an embodiment of the present application further provides a computer device, including a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the steps in the information recommendation method provided in the embodiment of the present application.

[0043] In addition, an embodiment of the present application further provides a computer-readable storage medium, which stores a plurality of instructions, and the instructions are suitable for a processor to load to execute the steps in any one of the information recommendation methods provided in the embodiments of the present application.

[0044] In addition, an embodiment of the present application also provides a computer program product, including computer instructions, which, when executed, implement the steps of any one of the information recommendation methods provided in the embodiments of the present application.

[0045] The embodiment of the present application can collect the time parameters displayed by each target item in the historical time period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, the project frequency value corresponding to each target item in the project frequency array is updated, and the project weight corresponding to each target item is determined according to the updated project frequency array; the selection frequency of each target item by the target object in the historical time period is obtained, and the behavior weight of the target object for each target item is calculated according to the project weight and selection frequency corresponding to each target item; the time attenuation coefficient is determined according to the preset attenuation coefficient and the historical time period, and the interest score of the target object for each target item is determined according to the time attenuation coefficient and the behavior weight; the information of the project to be recommended is determined according to the interest score of each target item, and the information of the project to be recommended is sent to the target object. It can be concluded that the embodiment of the present application can generate a project window array containing the window serial number of the most recent display of each target project in the historical period according to the time parameters of each target project in the historical period, and generate a project frequency array containing the project frequency value displayed by each target project in the historical period, and perform sliding updates on the project frequency array to further determine the project weights under the project frequency array of the current period; then, determine the behavior weight of the target object based on the project weight, and determine the interest score of the target object for each project according to the behavior weight and the time attenuation coefficient, and make information recommendations based on the interest score; in this way, the frequency of the target project is estimated by sliding updates of the array, which saves a huge amount of memory space, improves computing efficiency, and improves information push efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 is a scenario diagram of an information recommendation system provided by an embodiment of the present application;

[0048] Figure 2 It is a schematic diagram of the steps of the information recommendation method provided in the embodiment of the present application;

[0049] Figure 3 This is another step flow chart of the information recommendation method provided in the embodiment of the present application;

[0050] Figure 4 It is a flowchart of the information recommendation method provided in the embodiment of the present application;

[0051] Figure 5is a schematic diagram of the structure of an information recommendation device provided in an embodiment of the present application;

[0052] Figure 6 It is a schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0054] The embodiments of the present application provide an information recommendation method, device and computer-readable storage medium. Specifically, the embodiments of the present application will be described from the perspective of an information recommendation device, which can be integrated in a computer device, which can be a server or a terminal. Among them, the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, a car terminal, an intelligent voice interaction device, an aircraft, etc., but is not limited to this. The terminal and the server can be directly or indirectly connected via wired or wireless communication, and this application does not limit this.

[0055] The solution provided in the embodiment of the present application involves technologies such as information recommendation in scenarios such as smart transportation and the background of the Internet of Vehicles. The Intelligent Traffic System (ITS), also known as the Intelligent Transportation System (ITS), is an effective and comprehensive application of advanced science and technology (information technology, computer technology, data communication technology, sensor technology, electronic control technology, automatic control theory, operations research, artificial intelligence, etc.) to transportation, service control and vehicle manufacturing, strengthening the connection between vehicles, roads and users, thereby forming a comprehensive transportation system that ensures safety, improves efficiency, improves the environment and saves energy. It is specifically described by the following embodiments:

[0056] For example, see Figure 1 , is a schematic diagram of a scenario of an information recommendation system provided in an embodiment of the present application. The scenario includes a terminal or a server.

[0057] The terminal or server can collect the time parameters of each target item in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, the project frequency value corresponding to each target item in the project frequency array is updated, and the project weight corresponding to each target item is determined according to the updated project frequency array; the selection frequency of each target item by the target object in the historical period is obtained, and the behavior weight of the target object for each target item is calculated according to the project weight and selection frequency corresponding to each target item; the time attenuation coefficient is determined according to the preset attenuation coefficient and the historical period, and the interest score of the target object for each target item is determined according to the time attenuation coefficient and the behavior weight; the information of the items to be recommended is determined according to the interest score of each target item, and the information of the items to be recommended is sent to the target object.

[0058] Among them, information recommendation can include collecting time parameters of target items, generating arrays, updating arrays, determining item weights, calculating behavior weights of target objects, determining interest scores for each target item, pushing information, and other processing methods.

[0059] It should be noted that the order of the following embodiments is not intended to limit the preferred order of the embodiments.

[0060] In the embodiment of the present application, the description will be made from the perspective of a user interface test information recommendation device, and the user interface test information recommendation device can be integrated into a computer device such as a terminal or a server. Figure 2 , Figure 2 This is a schematic diagram of the steps of a method for recommending user interface test information provided in an embodiment of the present application. In the embodiment of the present application, the information recommendation device is specifically integrated on a terminal as an example. When the processor on the terminal executes the program corresponding to the information processing method, the specific process is as follows:

[0061] 101. Collect the time parameters of each target project in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target project.

[0062] The target item may refer to a specific item. For example, the target item may be an advertisement, an article, news, a short video or other specific item. One advertisement or one article corresponds to one target item.

[0063] The historical period may be any period before the current time, such as the past hour, day, week, or a specific period (eg, 12:00-15:00 on xx day).

[0064] The time parameter may be the specific time information or time period information of the target item when it is displayed. For example, taking an article as a target item, the time parameter of the article when it is displayed in the historical time period is 12:30 (specific time information), or the time parameter of the article when it is displayed in the historical time period is 12:00 to 13:00 one day ago (time period information). It should be noted that the time parameter of each target item can be used to determine the window (time window) of the target item when it is displayed.

[0065] The project window array may be an array of time window numbers or time window parameters containing one or more projects, which is used to record the time window number or parameter of each target project when displayed in historical time for subsequent calculations.

[0066] Among them, the time window can be a window set with a specific duration as the time interval, such as setting a 5-minute duration interval as a time window, or other duration intervals, such as 1 minute, 10 minutes, 1 hour, 1 day, etc. It should be noted that each time window can be connected, for example, 0-5 minutes is the first time window, 5-10 minutes is the second time window, and 10-15 minutes is the third time window; in addition, each time window can be discontinuous, such as taking a fixed time period of each date as a time window, for example, 0-5 minutes of the first day is the first time window, 0-5 minutes of the second day is the second time window, and 0-5 minutes of the third day is the third time window; for example, different time periods within the same date are used as time windows, such as 0-1 hours of the same day as the first time window, 6-8 hours of the same day as the second time window, and 12-14 hours of the same day as the third time window. The above is only an example of enumeration about time windows, and the embodiments of the present application are not specifically limited.

[0067] It should be noted that the time window can be used to record the time of the target item displayed within the time interval, such as recording the time window in which the relevant target item is displayed, or recording the frequency of the relevant target item when it is displayed in the time window (on average how many time windows it appears once).

[0068] In order to distinguish each time window, the embodiment of the present application sets a serial number corresponding to each time window, and the time window serial number can be an identifier to distinguish each time window, so as to clarify the specific window where each target project is displayed. For example, based on the time sequence, the first time window is time window 1 (serial number 1), the second time window is time window 2 (serial number 2), and the third time window is time window 3 (serial number 3). For example, the time window where project A was last displayed is time window 3, and the time window where project B was last displayed is time window 1.

[0069] Among them, the item frequency array can be an array containing window frequency values ​​(item frequency values) corresponding to one or more target items, that is, it contains the window frequency value of each target item when it is displayed, such as how many time windows each target item appears (displayed once) on average.

[0070] In order to be able to subsequently count and calculate the display frequency (or display time) of each target item in the streaming data, the embodiment of the present application can collect the time parameters of each target item in the streaming data when it is displayed in real time to generate an array of window parameters for each target item in the streaming data. Specifically, when each target item is detected to be displayed, the time parameters of each target item when it is displayed in the historical period can be collected in real time to generate a project window array containing the window sequence number of the most recent display of each target item in the historical period according to the time parameters of each target item, and generate a project frequency array containing the project frequency value displayed by each target item in the historical period. In this way, the window data of each target item in the streaming data (such as the window sequence number and window frequency value of each target item) is recorded, so as to facilitate the subsequent statistics and calculation of the frequency of the target item based on the above array.

[0071] In some implementations, the step of “generating a project window array and a project frequency array according to the time parameters corresponding to each target project” may include:

[0072] (1) Generate the corresponding initial project window array and initial project frequency array according to the time parameters corresponding to each target project.

[0073] In order to record the display situation (such as display time) of each target item in the streaming data, the embodiment of the present application can generate an initial item window array and an initial item frequency array according to the time parameters corresponding to each target item after collecting the time parameters of each target item when displayed in the historical period. Specifically, multiple time windows contained in the historical period can be obtained, each time window has a corresponding window serial number, and the project time window serial number corresponding to each target item is determined according to the time parameter of each target item, and the initial item window array is generated according to the project time window serial number corresponding to each target item; and, the number of times each target item is displayed in multiple time windows is determined according to the time parameter of each target item, and the project frequency of each target item is determined according to the number of windows and the number of displays corresponding to the multiple time windows, so as to generate the corresponding initial item frequency array according to the project frequency of each target item.

[0074] (2) Obtain the project identifier corresponding to each target project, and perform mapping processing on the project identifier corresponding to each target project to obtain the project mapping value corresponding to each target project.

[0075] The project identifier may be identification information of a related target project, specifically a number or ID of the target project, which is used to identify or mark related target projects so that each target project can be distinguished by its own project identifier.

[0076] Among them, the project mapping value can be a value after mapping processing. It should be noted that in the embodiment of the present application, the project identification of each target project may be large. If the project identification is composed of multi-bit values, it needs to occupy more bytes and other system resources. If the values ​​in the array are directly labeled with the project identification, it will occupy more memory. Therefore, before labeling the parameters related to each target project in the array, the project identification of each target project can be mapped to simplify the project identification of each target project, so as to label the parameters of the relevant target projects in the array with the project mapping value.

[0077] Specifically, in order to reduce the memory occupied by the project identifier of each target project after labeling, the embodiment of the present application can obtain the project identifier corresponding to each target project, and map the project identifier corresponding to each target project to obtain the identifier mapping value after mapping, that is, the project mapping value of each target project, so as to facilitate the subsequent labeling of related parameters in the array according to the project mapping value of each target project.

[0078] In some implementations, the step of “mapping the project identifier corresponding to each target project to obtain a project mapping value corresponding to each target project” may include:

[0079] Obtain a preset hash function corresponding to each target project from a preset function database; perform hash mapping on the project identifier corresponding to each target project according to the preset hash function corresponding to each target project, and obtain a project mapping value corresponding to each target project.

[0080] Among them, the preset function database may include one or more hash functions, each hash function can be used to map the project identification of a specific target project. In addition, each hash function can also be used to map the project identification of the target project within the same historical period (all time windows) or the same time window, without specific limitation.

[0081] In order to map the project identification of each target project, the embodiment of the present application specifically takes each hash function as an example for mapping the project identification of a specific target project, and obtains the preset hash function corresponding to each target project, and performs hash value mapping calculation on the project identification corresponding to each target project according to the preset hash function corresponding to each target project, and obtains the identification hash value corresponding to each target project, that is, the project mapping value. In this way, before marking the relevant parameters of each target project in the array, the project identification of each target project can be mapped to obtain the project mapping value of each target project, so as to facilitate the subsequent use of marking the relevant parameters of each target project in the array, reduce the amount of array data after marking, and reduce the system memory resources occupied when calculating / counting the frequency data of each target project in the streaming data.

[0082] (3) According to the project mapping value corresponding to each target project, the initial project window array and the initial project frequency array are marked respectively to obtain the project window array and the project frequency array.

[0083] In order to obtain the complete record data of each target item in the streaming data when it is displayed in the historical time period, after calculating the project mapping value corresponding to each target item, the embodiment of the present application can use the project mapping value to mark the parameters corresponding to the relevant target items in the array to obtain the complete record data of each target item when it is displayed in the historical time period, that is, the project window array and the project frequency array, so as to facilitate the subsequent use of the array corresponding to the streaming data to perform frequency statistics and calculations for each target item.

[0084] Specifically, in order to obtain a project window array for subsequent calculations, the embodiment of the present application can mark the project window sequence number corresponding to each target project in the initial project window array according to the project mapping value corresponding to each target project to obtain a marked project window array.

[0085] Specifically, in order to obtain an item frequency array for subsequent calculations, the embodiment of the present application can mark the item frequency value corresponding to each target item in the initial item frequency array according to the item mapping value corresponding to each target item to obtain a marked item frequency array.

[0086] Through the above method, the window data of each target item in the streaming data (such as the window sequence number and window frequency value of each target item) can be collected and recorded in array form, so as to facilitate the subsequent statistics and calculation of the frequency of the target item based on the above array.

[0087] 102. Based on the preset iteration value and the project window array, the project frequency value corresponding to each target project in the project frequency array is updated, and the project weight corresponding to each target project is determined according to the updated project frequency array.

[0088] The preset iteration value may be a preset update step value of the array parameter, which may be understood as a learning rate of the parameters in the array, and is used for iteratively updating the array parameters when the time window changes. For example, the preset iteration value is 0.1, and in order to iteratively update each parameter in the array, the parameters in the array may be iteratively updated according to the preset iteration value 0.1.

[0089] The project weight may be the proportion of the target project when it is displayed in the target time window, and is used to indicate the proportion between the target project and other target projects in the target time window.

[0090] In order to estimate the display frequency of each target item in the current time window, after obtaining the project window array and the project frequency array containing the display status of each target item in the historical period, the embodiment of the present application can update the project frequency value in the project frequency array to estimate the project frequency value of each target item in the current time window, specifically, according to the preset iteration value, the project window array and the project frequency array and other parameters, the parameters in the project frequency array of the historical period are updated to obtain the updated project frequency array. In this way, the project frequency value of each target item in the current time window is estimated based on the data of the project window array and the project frequency array of the preset iteration value and the historical period, so as to update the parameters in the project frequency array.

[0091] In some implementations, the step of “updating the item frequency value corresponding to each target item in the item frequency array based on the preset iteration value and the item window array” may include:

[0092] (1) Extract the project window sequence number of each target project in the project window array, and extract the project frequency value of each target project in the project frequency array.

[0093] (2) Perform a weighted summation of the project window sequence number and project frequency value of each target project according to the preset iteration value to obtain the updated project frequency value corresponding to each target project.

[0094] (3) The item frequency value corresponding to each target item in the item frequency array is updated according to the updated item frequency value corresponding to each target item.

[0095] The project window serial number may be the time window serial number in which the corresponding target project was displayed in the historical period, or the time window serial number in which the target project was last displayed in the historical period.

[0096] The item frequency value may be a window frequency value of the corresponding target item when displayed in a historical period, that is, the display frequency of each target item when displayed, such as how many time windows each target item is displayed once on average.

[0097] Specifically, in order to update the project frequency array, the embodiment of the present application can extract the project window sequence number of each target project in the project window array, and extract the project frequency value of each target project from the project frequency array, and perform weighted summation of the project window sequence number and the project frequency value corresponding to each target project through a preset iteration value to obtain an updated project frequency value corresponding to each target project. Furthermore, the updated project frequency value of each target project updates the corresponding original project frequency value in the project frequency array to obtain an updated project frequency value.

[0098] In some implementations, the step of “performing a weighted summation of the project window sequence number and the project frequency value of each target project according to a preset iteration value to obtain an updated project frequency value corresponding to each target project” may include:

[0099] (2.1) The item frequency value of each target item is weighted according to the preset iteration value to obtain the weighted item frequency value.

[0100] (2.2) Obtain the current target time window sequence number, and perform weighted processing on the sequence number difference between the target time window sequence number and the project window sequence number according to a preset iteration value to obtain a weighted time window value.

[0101] (2.3) The weighted item frequency value and the weighted time window value are summed to obtain the updated item frequency value corresponding to each target item.

[0102] It should be noted that since there is a certain time interval between the current time window and the time window in the historical period, the project frequencies of the target projects displayed therein will also be different. The frequency estimation of the target project in the current time window can be participated in by using the preset iteration value, so that in the case of iteration of the time window, the project frequency value of the target project is updated through the learning rate (preset iteration value) to complete the estimation of the project frequency value of the target project in the current time window.

[0103] Specifically, in order to estimate the project frequency value of each target project in the current time window, the embodiment of the present application can use the preset iteration value to perform weighted calculation on the project frequency value of each target project in the historical period to obtain the weighted project frequency value; and, obtain the window sequence number difference between the project window sequence number of each target project when it was last displayed and the target time window sequence number of the current period, so as to perform weighted processing on the window sequence number difference according to the preset iteration value to obtain the weighted time window value; and then, sum the weighted project frequency value and the weighted time window value to obtain the updated project frequency value corresponding to each target project. In this way, the project frequency value of each target project in the current time window is estimated to predict the project display frequency of each target project in the current time window, that is, how many time windows each target project is displayed once on average in the current time window.

[0104] Furthermore, in order to determine the project weight of each target project in the current time window, the embodiment of the present application can determine the project weight corresponding to each target project in the current time window based on the updated project frequency value of each target project in the updated project frequency array.

[0105] In some embodiments, the step of "determining the project weight corresponding to each target project based on the updated project frequency array" may include: selecting the updated project frequency value corresponding to each target project from the updated project frequency array; determining the project weight corresponding to each target project based on the updated project frequency value corresponding to each target project.

[0106] Specifically, in order to determine the project weight of each target project in the current time window, each updated project frequency value in the project frequency array after the update iteration is completed can be obtained to determine the project weight of the corresponding target project according to each updated project frequency value, such as obtaining the project weight of the target project according to the reciprocal ratio of the updated project frequency value. In this way, after estimating the updated project frequency value in the current time window based on the project frequency value of each target project in the historical period, the current project weight of the target project can be estimated based on the estimated updated project frequency value, so as to facilitate the subsequent determination of the behavior weight of the target object based on the target project.

[0107] Through the above method, the item frequency value of each target item when displayed in the historical time period can be updated in combination with the preset iteration value, so as to estimate the window data (such as item frequency value) in the current time window according to the time window data of each target object in the historical time period. In this way, the display frequency of the target item in the current time window (or other target time window) is estimated by array sliding update. There is no need to store massive data in the historical time period, which saves massive memory space and improves the computing efficiency of the system.

[0108] 103. Obtain the selection frequency of each target item by the target object in the historical period, and calculate the behavior weight of the target object for each target item according to the item weight and selection frequency corresponding to each target item.

[0109] Among them, the selection frequency can be the number of times the selection behavior is performed on the relevant target item, reflecting the frequency of the target object on the relevant target item. The selection operation corresponding to the selection frequency is not limited to reading, clicking, collecting, watching, etc.; for example, taking an article or advertisement as a target item, when it is detected that the target object performs any operation such as clicking, reading or collecting on the article, it can be regarded as a selection operation of the target object on the article, and the operation is counted into the selection frequency of the article. It can be understood that the target item that the target object is interested in can be reflected through the selection frequency of each target item by the target object.

[0110] Among them, the behavior weight can be the weight of the relevant target item after being selected by the target object. The behavior weight can represent the proportion of the target object selecting the relevant target item, reflecting the weight of the target item selected by the target object in the time window within the historical period.

[0111] It should be noted that in order to determine the target object's interest score for each target item, the embodiment of the present application, after estimating the project weight of each target item at the current moment, also needs to calculate the target object's behavior weight for each target item based on the project weight, so as to subsequently determine the target object's interest score for the relevant target items based on the behavior weight of each target item.

[0112] Among them, in order to determine the behavior weight of the target object for each target item, the embodiment of the present application needs to collect the number of selection operations of the target object for each target item in the time window of the historical period, that is, the selection frequency, after estimating the item weight of each target item in the current time window; thereby, the behavior weight of the target object is calculated according to the item weight and selection frequency of each target item. It should be noted that the behavior weight can be the behavior weight of the target object for the same target item in all time windows in the historical period. The behavior weights between different target items may be different, which is mainly determined by two factors: the item weight of each target item and the selection frequency of the target object for the target item.

[0113] Through the above method, the target object's behavioral weight for the corresponding target item can be determined based on the project weight of each target item estimated in the current time window and the target object's selection frequency of the relevant target item in the historical period, so as to subsequently determine the interest score of the corresponding target item based on the behavioral weight.

[0114] 104. Determine the time decay coefficient according to the preset decay coefficient and the historical period, and determine the interest score of the target object for each target item according to the time decay coefficient and the behavior weight.

[0115] Among them, the preset attenuation coefficient can be a preset interest attenuation coefficient, which can specifically be a preset fixed value, which is mainly used to indicate the degree of attenuation of the target object's interest in the relevant target project. It is understandable that over time, the user's interest in the relevant target project may decay slightly. For example, if the user read an article an hour ago, then the user's interest in the same article or article-like project at present or for a period of time may decay slightly, and the interest point may be in other projects, such as videos, news, etc. Therefore, the embodiment of the present application sets a preset attenuation coefficient for participating in the calculation to determine the time attenuation coefficient.

[0116] The time decay coefficient may be a time decay value of the target object's interest in each target item, which belongs to a time decay factor and is used to participate in the calculation of the target object's interest score in each target item.

[0117] The interest score may be an interest index of the target object in the relevant target item, and the interest score may reflect the interest level of the target object in the relevant target item. It is understandable that the interest ratio of the target object in different target items may be determined by comparing the interest scores between different target items.

[0118] In order to determine the time-effectiveness attenuation factor of the target object's interest in each target item in the current period, the embodiment of the present application can determine the time-effectiveness attenuation factor according to the preset attenuation coefficient and the historical period. Among them, the step of "determining the time-effectiveness attenuation coefficient according to the preset attenuation coefficient and the historical period" can include: obtaining the target time of the current period, and obtaining the time difference between the target time and the historical period; determining the time-effectiveness attenuation coefficient of the current period according to the preset attenuation coefficient and the time difference.

[0119] It should be noted that the target object's interest in the relevant target project will change over time, and the target object's interest level in the target project will be different at different times from the historical period, that is, the instant effect attenuation coefficient will be different.

[0120] Specifically, in order to determine the time decay coefficient of the current time, the embodiment of the present application calculates the time difference between the historical period and the current time, and calculates the time decay coefficient of the current time based on the time difference and the preset decay coefficient, such as using the time difference as the exponent of the preset decay coefficient to calculate the time decay coefficient of the current time.

[0121] Furthermore, in order to determine the interest score of the target object for each target item, the embodiment of the present application may specifically perform a product based on the current time decay coefficient and the behavior weight to obtain the interest score of the target object for each target item.

[0122] Through the above method, the time decay coefficient at the current time can be obtained to determine the interest score according to the time decay coefficient and the behavior weight of the target object for each target item, so as to facilitate the subsequent determination of the recommendable target item information based on the target object's interest score for each target item and make recommendations.

[0123] 105. Determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

[0124] The information of the item to be recommended may be relevant information of the target item to be recommended, such as text content, video content or voice content of the target item, etc., which is not specifically limited here.

[0125] After obtaining the target object's interest score for each target item, the embodiment of the present application needs to first determine the target item to be recommended based on the interest score, thereby determining the item information to be recommended corresponding to the target item to be recommended, and then sending the item information to be recommended to the target object, so that the item information to be recommended is displayed in the time window of the current time period for the target object to subscribe, click, read, collect, and other operations.

[0126] In some implementations, the step of “determining the information of the item to be recommended according to the interest score of each target item” may include:

[0127] (1) Based on the interest score of each target item, determine an interest score sequence containing multiple target items.

[0128] (2) Determine the project capacity of the current time window.

[0129] (3) Based on the interest score sequence, select the target project corresponding to the project capacity, and determine the target project corresponding to the project capacity as the project information to be recommended.

[0130] Among them, the interest score sequence can be a sequence including one or more interest scores, which includes the interest score corresponding to each target item. The sequence can specifically be a sequence of interest scores of multiple target items, which can be arranged from large to small, or from small to large, and is not limited here.

[0131] Specifically, after obtaining the target object's interest score for each target item, the embodiment of the present application can arrange the target items from large to small according to the interest score of each target item to obtain an interest score sequence containing multiple target items; then, determine the project capacity of the current time window, and select the target item corresponding to the project capacity from the interest score sequence as the item to be recommended, thereby determining the information of the item to be recommended.

[0132] It should be noted that when determining the project capacity of the current time window, the project capacity of the current time window can be estimated based on the project capacity of each time window in the historical period. In addition, when selecting target projects corresponding to the project capacity as the recommended projects, the total project length of one or more target projects with larger scores in the interest score sequence can be estimated first, so as to select one or more target projects corresponding to the total project length that meets the project capacity as the recommended projects.

[0133] In the embodiment of the present application, in the face of a large amount of streaming data, the window data of each target item in the historical period is collected and recorded and updated in real time in the form of an array to realize recording the display data of each target item in the streaming data for subsequent calculation. Then, after obtaining the array containing the window data of each target item when displayed, the item frequency value of each target item in the item frequency array when displayed in the historical period is updated in combination with the preset iteration value, and the updated item frequency array is obtained, which realizes the estimation of the display frequency of the target item in the current time window by the array sliding update method, without the need to store the massive data of the historical period, and improves the calculation efficiency of the system. Then, the project weight of the corresponding target item is determined based on each updated item frequency value estimated in the updated item frequency array, and the target object's behavior weight for each target item is determined in combination with the selected operation and item weight of the target object in the historical period, so as to determine the interest score of the target object for each target item according to the behavior weight. Finally, the information of the project to be recommended is determined based on the interest score of each target item, and the information of the project to be recommended is sent to the target object, so that the information of the project to be recommended is displayed in the time window of the current period. In this way, the computing efficiency is improved and the efficiency of information push is increased.

[0134] As can be seen from the above, the embodiments of the present application can collect the time parameters displayed by each target item in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, the project frequency value corresponding to each target item in the project frequency array is updated, and the project weight corresponding to each target item is determined according to the updated project frequency array; the selection frequency of each target item by the target object in the historical period is obtained, and the behavior weight of the target object for each target item is calculated according to the project weight and selection frequency corresponding to each target item; the time attenuation coefficient is determined according to the preset attenuation coefficient and the historical period, and the interest score of the target object for each target item is determined according to the time attenuation coefficient and the behavior weight; the information of the project to be recommended is determined according to the interest score of each target item, and the information of the project to be recommended is sent to the target object. It can be concluded that the embodiment of the present application can generate a project window array containing the window serial number of the most recent display of each target project in the historical period according to the time parameters of each target project in the historical period, and generate a project frequency array containing the project frequency value displayed by each target project in the historical period, and perform sliding updates on the project frequency array to further determine the project weights under the project frequency array of the current period; then, determine the behavior weight of the target object based on the project weight, and determine the interest score of the target object for each project according to the behavior weight and the time attenuation coefficient, and make information recommendations based on the interest score; in this way, the frequency of the target project is estimated by sliding updates of the array, which saves a huge amount of memory space, improves computing efficiency, and improves information push efficiency.

[0135] According to the method described in the above embodiment, the following is further described in detail with examples.

[0136] The embodiment of the present application takes information recommendation as an example to further describe the information recommendation method provided in the embodiment of the present application.

[0137] Figure 3 is another step flow chart of the information recommendation method provided in the embodiment of the present application, Figure 4 This is a flowchart of the information recommendation method provided in the embodiment of the present application. For ease of understanding, please refer to Figure 3 and Figure 4 , the embodiments of the present application are described.

[0138] In the embodiments of the present application, the information recommendation device will be described from the perspective of the information recommendation device, which can be integrated into a computer device such as a terminal and / or a server. For example, taking the integration into a terminal or a server as an example, when the processor on the terminal or the server executes the program corresponding to the information recommendation method, the specific process of the information recommendation method is as follows:

[0139] 201. Collect time parameters of each target item in a historical period.

[0140] The target item may refer to a specific item. For example, the target item may be an advertisement, an article, news, a short video or other specific item. One advertisement or one article corresponds to one target item.

[0141] The historical period may be any period before the current time, such as the past 5 minutes, 1 hour, 1 day, one week, or a specific period (eg, 12-15 o'clock on xx day).

[0142] The time parameter may be the specific time information or time period information of the target item when it is displayed. For example, taking an article as a target item, the time parameter of the article when it is displayed in the historical time period is 12:30 (specific time information), or the time parameter of the article when it is displayed in the historical time period is 12:00 to 13:00 one day ago (time period information). It should be noted that the time parameter of each target item can be used to determine the window (time window) of the target item when it is displayed.

[0143] 202. Generate a corresponding initial project window array and an initial project frequency array according to the time parameter corresponding to each target project.

[0144] The (initial) project window array may be an array of time window numbers or time window parameters containing one or more projects, which is used to record the time window number or parameter of each target project when displayed in historical time for subsequent calculations.

[0145] Among them, the time window can be a window set with a specific duration as the time interval, such as setting a 5-minute duration interval as a time window, or other duration intervals, such as 1 minute, 10 minutes, 1 hour, 1 day, etc. It should be noted that each time window can be connected, for example, 0-5 minutes is the first time window, 5-10 minutes is the second time window, and 10-15 minutes is the third time window; in addition, each time window can be discontinuous, such as taking a fixed time period of each date as a time window, for example, 0-5 minutes of the first day is the first time window, 0-5 minutes of the second day is the second time window, and 0-5 minutes of the third day is the third time window; for example, different time periods within the same date are used as time windows, such as 0-1 hours of the same day as the first time window, 6-8 hours of the same day as the second time window, and 12-14 hours of the same day as the third time window. The above is only an example of enumeration about time windows, and the embodiments of the present application are not specifically limited.

[0146] It should be noted that the time window can be used to indicate the time of the target item displayed within the time interval, such as recording the time window in which the relevant target item is displayed, or recording the frequency of the relevant target item when it is displayed in the time window (on average how many time windows it appears once).

[0147] Among them, the (initial) item frequency array can be an array containing window frequency values ​​(item frequency values) corresponding to one or more target items, that is, it contains the window frequency value of each target item when displayed, such as how many time windows each target item appears (displayed) on average.

[0148] In order to record the display status (such as display time) of each target item in the streaming data, after collecting the time parameters of each target item when displayed in the historical period, the embodiment of the present application can first initialize the array so that the parameters of the initialized array are 0, and initialize the preset iteration value and the dimension of the array.

[0149] For example, take array A as the project window array and array B as the project frequency array. After initializing array A and array B, Indicates item The sequence number of the window that appeared last time. Indicates item How many windows appear on average? Indicates the dimensions of matrices A and B, which are fixed parameters. Represents the learning rate and sets a fixed initial value.

[0150] Furthermore, an initial project window array and an initial project frequency array can be generated according to the time parameters corresponding to each target project. Specifically, multiple time windows included in the historical period can be obtained, each time window has a corresponding window serial number, and the project time window serial number corresponding to each target project is determined according to the time parameters of each target project, and the initial project window array is generated according to the project time window serial number corresponding to each target project; and the number of times each target project is displayed in multiple time windows is determined according to the time parameters of each target project, and the project frequency of each target project is determined according to the number of windows and the number of displays corresponding to the multiple time windows, so as to generate the corresponding initial project frequency array according to the project frequency of each target project.

[0151] 203. Obtain a project identifier corresponding to each target project, and perform mapping processing on the project identifier corresponding to each target project to obtain a project mapping value corresponding to each target project.

[0152] The project identifier may be identification information of a related target project, specifically a number or ID of the target project, which is used to identify or mark related target projects so that each target project can be distinguished by its own project identifier.

[0153] Among them, the project mapping value can be a value after mapping processing. It should be noted that in the embodiment of the present application, the project identification of each target project may be large. If the project identification is composed of multi-bit values, it needs to occupy more bytes and other system resources. If the values ​​in the array are directly labeled with the project identification, it will occupy more memory. Therefore, before labeling the parameters related to each target project in the array, the project identification of each target project can be mapped to simplify the project identification of each target project, so as to label the parameters of the relevant target projects in the array with the project mapping value.

[0154] Specifically, taking the example that each hash function can be used to map the project identifier of a specific target project, the preset hash function corresponding to each target project can be obtained, and the project identifier corresponding to each target project can be mapped and calculated based on the preset hash function corresponding to each target project to obtain the identifier hash value corresponding to each target project, that is, the project mapping value.

[0155] For example, the embodiment of the present application sets multiple Hash functions. The hash function is , take the Hash functions Map all target project item IDs to [0, H], and the mapped values ​​are , obtained is an array and arrays The subscript of .

[0156] In this way, before labeling the relevant parameters of each target item in the array, the project identifier of each target item can be mapped to obtain the project mapping value of each target item, so as to facilitate the subsequent labeling of the relevant parameters of each target item in the array, reduce the amount of array data after labeling, and reduce the system memory resources occupied when calculating / counting the frequency data of each target item in the streaming data.

[0157] 204. Annotate the initial project window array and the initial project frequency array according to the project mapping value corresponding to each target project to obtain a project window array and a project frequency array.

[0158] In order to obtain the complete record data of each target item in the streaming data when it is displayed in the historical time period, after calculating the project mapping value corresponding to each target item, the embodiment of the present application can use the project mapping value to mark the parameters corresponding to the relevant target items in the array to obtain the complete record data of each target item when it is displayed in the historical time period, that is, the project window array and the project frequency array, so as to facilitate the subsequent use of the array corresponding to the streaming data to perform frequency statistics and calculations for each target item.

[0159] Through the above method, the window data of each target item in the streaming data (such as the window sequence number and window frequency value of each target item) can be collected and recorded in the form of an array.

[0160] 205. Extract the project window sequence number of each target project in the project window array, and extract the project frequency value of each target project in the project frequency array.

[0161] In order to distinguish each time window, the embodiment of the present application sets a serial number corresponding to each time window, and the time window serial number can be an identifier to distinguish each time window, so as to clarify the specific window where each target project is displayed. For example, based on the time sequence, the first time window is time window 1 (serial number 1), the second time window is time window 2 (serial number 2), and the third time window is time window 3 (serial number 3). For example, the time window where project A was last displayed is time window 3, and the time window where project B was last displayed is time window 1.

[0162] The item frequency value may be a window frequency value of the corresponding target item when displayed in a historical period, that is, the display frequency of each target item when displayed, such as how many time windows each target item is displayed once on average.

[0163] 206. Perform weighted summation on the project window sequence number and the project frequency value of each target project according to a preset iteration value to obtain an updated project frequency value corresponding to each target project.

[0164] The preset iteration value may be a preset update step value of the array parameter, which may be understood as a learning rate of the parameters in the array, and is used for iteratively updating the array parameters when the time window changes. For example, the preset iteration value is 0.1, and in order to iteratively update each parameter in the array, the parameters in the array may be iteratively updated according to the preset iteration value 0.1.

[0165] Specifically, the project frequency value of each target project is weighted according to the preset iteration value to obtain the weighted project frequency value; and, the current target time window sequence number is obtained, and the sequence number difference between the target time window sequence number and the project window sequence number is weighted according to the preset iteration value to obtain the weighted time window value; then, the weighted project frequency value and the weighted time window value are summed to obtain the updated project frequency value corresponding to each target project.

[0166] For example, the project frequency value of each target project in the current period is calculated as follows:

[0167]

[0168] in, represents the item frequency value of each target item estimated under the current hash function, Represents the original item frequency value in the item frequency array, Represents the item window sequence number in the item window array.

[0169] 207. Update the item frequency value corresponding to each target item in the item frequency array according to the updated item frequency value corresponding to each target item.

[0170] After obtaining the updated item frequency value corresponding to each target item, the embodiment of the present application replaces the item frequency value corresponding to each target item in the original item frequency array according to the updated item frequency value corresponding to each target item, thereby obtaining an updated item frequency array. In this way, the item frequency value in the item frequency array is updated by an array sliding update method.

[0171] 208. Determine the project weight corresponding to each target project according to the updated project frequency array.

[0172] The project weight may be the proportion of the target project when it is displayed in the target time window, and is used to indicate the proportion between the target project and other target projects in the target time window.

[0173] Specifically, an updated item frequency value corresponding to each target item is selected from the updated item frequency array; and an item weight corresponding to each target item is determined according to the updated item frequency value corresponding to each target item.

[0174] In the embodiment of the present application, multiple sets of hash functions and arrays A and B can be used to calculate multiple different candidate item frequency values, and the maximum item frequency value is determined from the multiple different candidate item frequency values ​​as the updated item frequency value. The updated item frequency value is expressed as follows:

[0175]

[0176] Among them, the Indicates the updated item frequency value corresponding to each target item, Represents the candidate item frequency value of each target item estimated under different hash functions.

[0177] For example, the project weight of the target project is obtained according to the inverse ratio of the updated project frequency value. The calculation process is as follows:

[0178]

[0179] in, represents the project weight of each target project, and i represents the identifier of each target project.

[0180] 209. Obtain the selection frequency of each target item by the target object in the historical period, and calculate the behavior weight of the target object for each target item according to the item weight and selection frequency corresponding to each target item.

[0181] Among them, the selection frequency can be the number of times the selection behavior is performed on the relevant target item, reflecting the frequency of the target object on the relevant target item. The selection operation corresponding to the selection frequency is not limited to reading, clicking, collecting, watching, etc.; for example, taking an article or advertisement as a target item, when it is detected that the target object performs any operation such as clicking, reading or collecting on the article, it can be regarded as a selection operation of the target object on the article, and the operation is counted into the selection frequency of the article. It can be understood that the target item that the target object is interested in can be reflected through the selection frequency of each target item by the target object.

[0182] Among them, the behavior weight can be the weight of the relevant target item after being selected by the target object. The behavior weight can represent the proportion of the target object selecting the relevant target item, reflecting the weight of the target item selected by the target object in the time window within the historical period.

[0183] For example, the user's behavior weight For a fixed time window User behavior The cumulative sum is calculated as follows:

[0184]

[0185] Among them, u represents the user, i represents the item i, and t represents the tag.

[0186] 210. Determine the time-effect decay coefficient based on the preset decay coefficient and historical time period.

[0187] Among them, the preset attenuation coefficient can be a preset interest attenuation coefficient, which can specifically be a preset fixed value, which is mainly used to indicate the degree of attenuation of the target object's interest in the relevant target project. It is understandable that over time, the user's interest in the relevant target project may decay slightly. For example, if the user read an article an hour ago, then the user's interest in the same article or article-like project at present or for a period of time may decay slightly, and the interest point may be in other projects, such as videos, news, etc. Therefore, the embodiment of the present application sets a preset attenuation coefficient for participating in the calculation to determine the time attenuation coefficient.

[0188] The time decay coefficient may be a time decay value of the target object's interest in each target item, which belongs to a time decay factor and is used to participate in the calculation of the target object's interest score in each target item.

[0189] In order to determine the time-effectiveness attenuation factor of the target object's interest in each target item in the current period, the embodiment of the present application can determine the time-effectiveness attenuation factor according to the preset attenuation coefficient and the historical period. Specifically, the target time of the current period is obtained, and the time difference between the target time and the historical period is obtained; the time-effectiveness attenuation coefficient of the current period is determined according to the preset attenuation coefficient and the time difference.

[0190] For example, to calculate the time-efficiency decay factor of a user, the time-efficiency decay coefficient is calculated as follows:

[0191]

[0192] in, Represents the time-dependent attenuation coefficient, which decays with time As a fixed decay coefficient, k is the number of days from now, and T represents all time windows in the historical period.

[0193] 211. Determine the target object's interest score for each target item based on the time decay coefficient and behavior weight.

[0194] The interest score may be an interest index of the target object in the relevant target item, and the interest score may reflect the interest level of the target object in the relevant target item. It is understandable that the interest ratio of the target object in different target items may be determined by comparing the interest scores between different target items.

[0195] In order to determine the interest score of the target object for each target item, the embodiment of the present application can specifically perform a product based on the current time decay coefficient and the behavior weight to obtain the interest score of the target object for each target item. For each target project The process of behavioral interest scoring is as follows:

[0196]

[0197] Through the above method, the time decay coefficient at the current time can be obtained to determine the interest score according to the time decay coefficient and the behavior weight of the target object for each target item, so as to determine the recommendable target item information and make recommendations based on the target object's interest score for each target item.

[0198] 212. Determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

[0199] The information of the item to be recommended may be relevant information of the target item to be recommended, such as text content, video content or voice content of the target item, etc., which is not specifically limited here.

[0200] After obtaining the target object's interest score for each target item, the embodiment of the present application needs to first determine the target item to be recommended based on the interest score, thereby determining the item information to be recommended corresponding to the target item to be recommended, and then sending the item information to be recommended to the target object, so that the item information to be recommended is displayed in the time window of the current time period for the target object to subscribe, click, read, collect, and other operations.

[0201] Through the above, good results can be achieved in large-scale user behavior interest scores, which can be applied to scenarios such as user profiling and advertising to empower advertising.

[0202] From the above, it can be seen that the embodiment of the present application can collect the time parameters displayed by each target item in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, the project frequency value corresponding to each target item in the project frequency array is updated, and the project weight corresponding to each target item is determined according to the updated project frequency array; the selection frequency of each target item by the target object in the historical period is obtained, and the behavior weight of the target object for each target item is calculated according to the project weight and selection frequency corresponding to each target item; the time attenuation coefficient is determined according to the preset attenuation coefficient and the historical period, and the interest score of the target object for each target item is determined according to the time attenuation coefficient and the behavior weight; the information of the project to be recommended is determined according to the interest score of each target item, and the information of the project to be recommended is sent to the target object. It can be concluded that the embodiment of the present application can generate a project window array containing the window serial number of the most recent display of each target project in the historical period according to the time parameters of each target project in the historical period, and generate a project frequency array containing the project frequency value displayed by each target project in the historical period, and perform sliding updates on the project frequency array to further determine the project weights under the project frequency array of the current period; then, determine the behavior weight of the target object based on the project weight, and determine the interest score of the target object for each project according to the behavior weight and the time attenuation coefficient, and make information recommendations based on the interest score; in this way, the frequency of the target project is estimated by sliding updates of the array, which saves a huge amount of memory space, improves computing efficiency, and improves information push efficiency.

[0203] In order to better implement the above method, an embodiment of the present application further provides an information recommendation device, which can be integrated into a network device, such as a server or a terminal. The terminal may include a tablet computer, a laptop computer and / or a personal computer.

[0204] For example, Figure 5 As shown, the information recommendation device may include a generating unit 501 , an updating unit 502 , an acquiring unit 503 , a determining unit 504 and a sending unit 505 .

[0205] The generating unit 501 is used to collect the time parameters displayed by each target project in the historical period, and generate the project window array and the project frequency array according to the time parameters corresponding to each target project;

[0206] An updating unit 502, configured to update the project frequency value corresponding to each target project in the project frequency array based on a preset iteration value and the project window array, and determine the project weight corresponding to each target project according to the updated project frequency array;

[0207] The acquisition unit 503 is used to acquire the selection frequency of each target item by the target object in the historical period, and calculate the behavior weight of the target object for each target item according to the item weight corresponding to each target item and the selection frequency;

[0208] A determination unit 504 is used to determine a time decay coefficient according to a preset decay coefficient and a historical period, and to determine the interest score of the target object for each target item according to the time decay coefficient and the behavior weight;

[0209] The sending unit 505 is used to determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

[0210] In some embodiments, the updating unit 502 is further configured to:

[0211] Extract the project window sequence number of each target project in the project window array, and extract the project frequency value of each target project in the project frequency array; perform weighted summation of the project window sequence number and the project frequency value of each target project according to a preset iteration value to obtain an updated project frequency value corresponding to each target project; update the project frequency value corresponding to each target project in the project frequency array according to the updated project frequency value corresponding to each target project.

[0212] In some embodiments, the updating unit 502 is further configured to:

[0213] The project frequency value of each target project is weighted according to the preset iteration value to obtain the weighted project frequency value; the currently displayed target time window sequence number is obtained, and the sequence number difference between the target time window sequence number and the project window sequence number is weighted according to the preset iteration value to obtain the weighted time window value; the weighted project frequency value and the weighted time window value are summed to obtain the updated project frequency value corresponding to each target project.

[0214] In some embodiments, the updating unit 502 is further configured to:

[0215] Selecting the updated item frequency value corresponding to each target item from the updated item frequency array; and determining the item weight corresponding to each target item according to the updated item frequency value corresponding to each target item.

[0216] In some implementations, the generating unit 501 is further configured to:

[0217] Generate a corresponding initial project window array and an initial project frequency array according to the time parameters corresponding to each target project; obtain a project identifier corresponding to each target project, and map the project identifier corresponding to each target project to obtain a project mapping value corresponding to each target project; respectively mark the initial project window array and the initial project frequency array according to the project mapping value corresponding to each target project to obtain a project window array and a project frequency array.

[0218] In some implementations, the generating unit 501 is further configured to:

[0219] Obtain a preset hash function corresponding to each target project from a preset function database; perform hash mapping on the project identifier corresponding to each target project according to the preset hash function corresponding to each target project, and obtain a project mapping value corresponding to each target project.

[0220] In some implementations, the determining unit 504 is further configured to:

[0221] Obtain the target time of the current period and the time difference between the target time and the historical period; determine the time decay coefficient of the current period according to the preset decay coefficient and the time difference.

[0222] In some implementations, the sending unit 505 is further configured to:

[0223] Based on the interest score of each target item, an interest score sequence including multiple target items is determined; the project capacity of the current time window is determined; based on the interest score sequence, the target items corresponding to the project capacity are selected, and the target items corresponding to the project capacity are determined as the project information to be recommended.

[0224] As can be seen from the above, the embodiment of the present application can collect the time parameters displayed by each target item in the historical period through the generation unit 501, and generate the project window array and the project frequency array according to the time parameters corresponding to each target item; through the updating unit 502, based on the preset iteration value and the project window array, the project frequency value corresponding to each target item in the project frequency array is updated, and the project weight corresponding to each target item is determined according to the updated project frequency array; through the acquisition unit 503, the selection frequency of each target item by the target object in the historical period is obtained, and the behavior weight of the target object for each target item is calculated according to the project weight and selection frequency corresponding to each target item; through the determination unit 504, the time attenuation coefficient is determined according to the preset attenuation coefficient and the historical period, and the interest score of the target object for each target item is determined according to the time attenuation coefficient and the behavior weight; through the sending unit 505, the information of the project to be recommended is determined according to the interest score of each target item, and the information of the project to be recommended is sent to the target object. It can be concluded that the embodiment of the present application can generate a project window array containing the window serial number of the most recent display of each target project in the historical period according to the time parameters of each target project in the historical period, and generate a project frequency array containing the project frequency value displayed by each target project in the historical period, and perform sliding updates on the project frequency array to further determine the project weights under the project frequency array of the current period; then, determine the behavior weight of the target object based on the project weight, and determine the interest score of the target object for each project according to the behavior weight and the time attenuation coefficient, and make information recommendations based on the interest score; in this way, the frequency of the target project is estimated by sliding updates of the array, which saves a huge amount of memory space, improves computing efficiency, and improves information push efficiency.

[0225] The present application also provides a computer device, such as Figure 6 As shown, it shows a schematic diagram of the structure of the computer device involved in the embodiment of the present application, specifically:

[0226] The computer device may include components such as a processor 601 with one or more processing cores, a memory 602 with one or more computer-readable storage media, a power supply 603, and an input unit 604. Those skilled in the art will appreciate that Figure 6 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. Among them:

[0227] The processor 601 is the control center of the computer device. It uses various interfaces and lines to connect various parts of the entire computer device. By running or executing software programs and / or modules stored in the memory 602, and calling data stored in the memory 602, it executes various functions of the computer device and processes data, thereby performing overall detection of the computer device. Optionally, the processor 601 may include one or more processing cores; preferably, the processor 601 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 601.

[0228] The memory 602 can be used to store software programs and modules. The processor 601 executes various functional applications and information recommendations by running the software programs and modules stored in the memory 602. The memory 602 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 602 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 602 may also include a memory controller to provide the processor 601 with access to the memory 602.

[0229] The computer device also includes a power supply 603 for supplying power to various components. Preferably, the power supply 603 can be logically connected to the processor 601 through a power management system, so as to manage charging, discharging, and power consumption through the power management system. The power supply 603 can also include any components such as one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, and power status indicators.

[0230] The computer device may further include an input unit 604, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0231] Although not shown, the computer device may also include a display unit, etc., which will not be described in detail here. Specifically, in the embodiment of the present application, the processor 601 in the computer device will load the executable file corresponding to the process of one or more application programs into the memory 602 according to the following instructions, and the processor 601 will run the application program stored in the memory 602, thereby realizing various functions, as follows:

[0232] Collect the time parameters displayed for each target item in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, update the project frequency value corresponding to each target item in the project frequency array, and determine the project weight corresponding to each target item according to the updated project frequency array; obtain the selection frequency of each target item by the target object in the historical period, and calculate the target object's behavior weight for each target item according to the project weight and selection frequency corresponding to each target item; determine the time attenuation coefficient according to the preset attenuation coefficient and the historical period, and determine the target object's interest score for each target item according to the time attenuation coefficient and the behavior weight; determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

[0233] The specific implementation of the above operations can be found in the previous embodiments and will not be described in detail here.

[0234] From the above, it can be seen that the embodiment of the present application can collect the time parameters displayed by each target item in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, the project frequency value corresponding to each target item in the project frequency array is updated, and the project weight corresponding to each target item is determined according to the updated project frequency array; the selection frequency of each target item by the target object in the historical period is obtained, and the behavior weight of the target object for each target item is calculated according to the project weight and selection frequency corresponding to each target item; the time attenuation coefficient is determined according to the preset attenuation coefficient and the historical period, and the interest score of the target object for each target item is determined according to the time attenuation coefficient and the behavior weight; the information of the project to be recommended is determined according to the interest score of each target item, and the information of the project to be recommended is sent to the target object. It can be concluded that the embodiment of the present application can generate a project window array containing the window serial number of the most recent display of each target project in the historical period according to the time parameters of each target project in the historical period, and generate a project frequency array containing the project frequency value displayed by each target project in the historical period, and perform sliding updates on the project frequency array to further determine the project weights under the project frequency array of the current period; then, determine the behavior weight of the target object based on the project weight, and determine the interest score of the target object for each project according to the behavior weight and the time attenuation coefficient, and make information recommendations based on the interest score; in this way, the frequency of the target project is estimated by sliding updates of the array, which saves a huge amount of memory space, improves computing efficiency, and improves information push efficiency.

[0235] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0236] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the steps in any one of the information recommendation methods provided in the embodiments of the present application. For example, the instructions can execute the following steps:

[0237] Collect the time parameters displayed for each target item in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; based on the preset iteration value and the project window array, update the project frequency value corresponding to each target item in the project frequency array, and determine the project weight corresponding to each target item according to the updated project frequency array; obtain the selection frequency of each target item by the target object in the historical period, and calculate the target object's behavior weight for each target item according to the project weight and selection frequency corresponding to each target item; determine the time attenuation coefficient according to the preset attenuation coefficient and the historical period, and determine the target object's interest score for each target item according to the time attenuation coefficient and the behavior weight; determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

[0238] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0239] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0240] Since the instructions stored in the computer-readable storage medium can execute the steps in any information recommendation method provided in the embodiments of the present application, the beneficial effects that can be achieved by any information recommendation method provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0241] The above is a detailed introduction to an information recommendation method, device and computer-readable storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. An information recommendation method, characterized in that: include: Collect the time parameters of each target item displayed in the historical period, and generate a project window array and a project frequency array according to the time parameters corresponding to each target item; wherein the time parameter of each target item is used to determine the time window of the target item when it is displayed; the project window array is an array containing the time window sequence number or time window parameter of one or more items, which is used to record the time window sequence number or parameter of each target item when it is displayed in the historical time; Based on the preset iteration value and the item window array, the item frequency value corresponding to each target item in the item frequency array is updated, and the item weight corresponding to each target item is determined according to the updated item frequency array; Obtaining the selection frequency of each target item by the target object in the historical period, and calculating the behavior weight of each target item by the target object according to the item weight and the selection frequency corresponding to each target item; Determine a time decay coefficient according to a preset decay coefficient and the historical period, and determine the interest score of the target object for each target item according to the time decay coefficient and the behavior weight; The information of the items to be recommended is determined according to the interest score of each target item, and the information of the items to be recommended is sent to the target object.

2. The method according to claim 1, characterized in that The updating of the item frequency value corresponding to each target item in the item frequency array based on the preset iteration value and the item window array includes: Extracting the item window sequence number of each target item in the item window array, and extracting the item frequency value of each target item in the item frequency array; Performing a weighted summation of the project window sequence number and the project frequency value of each target project according to a preset iteration value to obtain an updated project frequency value corresponding to each target project; The item frequency value corresponding to each target item in the item frequency array is updated according to the updated item frequency value corresponding to each target item.

3. The method according to claim 2, characterized in that The step of performing weighted summation on the project window identifier and the project frequency value of each target project according to the preset iteration value to obtain the updated project frequency value corresponding to each target project includes: Performing weighted processing on the item frequency value of each target item according to a preset iteration value to obtain a weighted item frequency value; Obtaining the current target time window sequence number, and weighting the sequence number difference between the target time window sequence number and the project window sequence number according to the preset iteration value to obtain a weighted time window value; The weighted item frequency value and the weighted time window value are summed to obtain an updated item frequency value corresponding to each target item.

4. The method according to any one of claims 1 to 3, characterized in that: The determining the project weight corresponding to each target project according to the updated project frequency array includes: Selecting the updated item frequency value corresponding to each target item from the updated item frequency array; The project weight corresponding to each target project is determined according to the updated project frequency value corresponding to each target project.

5. The method according to claim 1, characterized in that The generating of the project window array and the project frequency array according to the time parameter corresponding to each target project includes: Generate a corresponding initial project window array and an initial project frequency array according to the time parameters corresponding to each target project; Acquire a project identifier corresponding to each target project, and perform mapping processing on the project identifier corresponding to each target project to obtain a project mapping value corresponding to each target project; The initial project window array and the initial project frequency array are respectively labeled according to the project mapping value corresponding to each target project to obtain a project window array and a project frequency array.

6. The method according to claim 5, characterized in that The mapping process is performed on the project identifier corresponding to each target project to obtain the project mapping value corresponding to each target project, including: Acquire a preset hash function corresponding to each target item from a preset function database; The project identifier corresponding to each target project is hash mapped according to the preset hash function corresponding to each target project to obtain the project mapping value corresponding to each target project.

7. The method according to claim 1, characterized in that The determining of the time-effect decay coefficient according to the preset decay coefficient and the historical period includes: Obtaining a target time for the current period, and obtaining a time difference between the target time and the historical period; The time-effect decay coefficient of the current time period is determined according to the preset decay coefficient and the time difference.

8. The method according to claim 1, characterized in that: The determining of the item information to be recommended according to the interest score of each target item includes: Based on the interest score of each target item, determining an interest score sequence including a plurality of target items; Determine the project capacity for the current time window; Based on the interest score sequence, a target item corresponding to the item capacity is selected, and the target item corresponding to the item capacity is determined as the item information to be recommended.

9. An information recommendation device, characterized in that: include: A generating unit, used for collecting the time parameters of each target item displayed in the historical period, and generating a project window array and a project frequency array according to the time parameters corresponding to each target item; wherein the time parameter of each target item is used to determine the time window of the target item when it is displayed; the project window array is an array containing the time window sequence number or time window parameter of one or more items, which is used to record the time window sequence number or parameter of each target item when it is displayed in the historical time; An updating unit, configured to update the item frequency value corresponding to each target item in the item frequency array based on a preset iteration value and the item window array, and determine the item weight corresponding to each target item according to the updated item frequency array; an acquisition unit, configured to acquire the selection frequency of each target item by the target object in the historical period, and calculate the behavior weight of each target item by the target object according to the item weight and the selection frequency corresponding to each target item; A determination unit, configured to determine a time decay coefficient according to a preset decay coefficient and the historical period, and determine an interest score of the target object for each target item according to the time decay coefficient and the behavior weight; The sending unit is used to determine the information of the items to be recommended according to the interest score of each target item, and send the information of the items to be recommended to the target object.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is computer-readable and stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor to execute the steps of the information recommendation method according to any one of claims 1 to 8.

11. A computer device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the steps in the information recommendation method according to any one of claims 1 to 8.

12. A computer program product, characterized in that The method comprises computer instructions, which, when executed, implement the steps of the information recommendation method according to any one of claims 1 to 8.

Images