Online self-media interaction management system and method based on artificial intelligence

Abstract

The invention discloses an online self-media interaction management system and method based on artificial intelligence, and belongs to the technical field of interaction management. After user authorization, extracting starting time, ending time and browsed content types of the user during historical online activities; constructing a browsing content category set and an activity information association library; calculating the duration ratio of the content type in the historical online activity, and constructing a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtaining high-frequency content types corresponding to wave crests of the content proportion fluctuation curve, and constructing a high-frequency content type set; and calculating the occurrence probability of the high-frequency content type when the user performs the next online activity, presetting a threshold value, and analyzing and recommending the content type. According to the method, through combination of historical behavior analysis, time sequence modeling and intelligent recommendation technologies, in-depth insight of user behaviors and optimization of interactive experience are achieved, and interactivity and attraction of online self-media are effectively enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of interaction management, and in particular to an online self-media interaction management system and method based on artificial intelligence. Background Art

[0002] With the rapid development of artificial intelligence and big data technologies, online self-media has become an important platform for people to obtain information, exchange views and entertain; however, due to the massive and complex amount of online content, users face the problem of information overload when browsing; in order to solve this problem, many self-media platforms try to combine users' browsing history, content preferences and behavior patterns, and use artificial intelligence technology to optimize content recommendation systems; however, existing recommendation algorithms mostly focus on simple collaborative filtering or classification methods based on user portraits, and these technologies are difficult to deeply explore the regular characteristics behind user behavior; in addition, with the diversification of user interaction behaviors, traditional recommendation models have shown certain limitations in processing time series analysis and high-frequency content recognition, resulting in limited accuracy of recommendation results and user satisfaction.

[0003] Existing recommendation technologies usually rely on static user portraits and ignore the dynamic changes in user preferences in different time periods. For example, users may show high attention to specific types of content in a short period of time, but traditional technologies cannot capture and respond to these fluctuating characteristics in a timely manner. In addition, existing technologies also process user behavior data in a relatively extensive manner, lacking in-depth exploration of the relationship between browsed content types and time tags. These deficiencies lead to obvious shortcomings in the accuracy and practicality of content recommendation systems, and users' personalized needs are not fully met. Summary of the invention

[0004] The purpose of the present invention is to provide an online self-media interactive management system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] An online self-media interaction management method based on artificial intelligence, the method comprising the following steps: after authorization by the user, obtaining the network log of the user in the online self-media, extracting the start time, end time and browsed content type of the user in the historical online activities; constructing a browsed content type set and an activity information association library; initializing the user's online activity history time series, adding time tags to the start time, end time and browsed content type; calculating the duration ratio of the content type in the historical online activities, and constructing a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtaining the high-frequency content type corresponding to the peak of the content ratio fluctuation curve, and constructing a high-frequency content type set; calculating the probability of the high-frequency content type appearing in the user's next online activity, presetting a threshold, and analyzing and recommending content types.

[0007] As a preferred solution of the online self-media interactive management method based on artificial intelligence described in the present invention, after the user's authorization, the network log of the user's online activities in the online self-media is obtained, the start time and end time of the user's single online activity are extracted from the network log, and the natural language processing technology is used to extract all the browsing content types of the user in the single online activity to construct a browsing content type set.

[0008] The start time, end time and browsed content category set of the user during the single online activity are associated to construct an activity profile generated by the user during the single online activity, wherein the activity profile records the associated start time, end time and browsed content category set of the user during the single online activity; and an activity information association library is constructed, wherein the activity information association library records the activity profile generated by the user during each online activity in chronological order.

[0009] As a preferred solution of the online self-media interaction management method based on artificial intelligence described in the present invention, the user's online activity history time series is initialized, denoted as TI = {Y t |t∈[1,T]}, where Y t represents the user’s t-th historical online activity, and T represents the total number of historical online activities of the user.

[0010] Based on the activity information association library, extract the user's historical online activity Y t The activity file generated when the browsing content type set recorded in the activity file is recorded as Q[ST(Y t )→ET(Y t )], where ST(Y t ) represents the user's historical online activity Y t The starting time of ET(Y t ) represents the user's historical online activity Yt The end time of Q[ST(Y t )→ET(Y t )]={BC a (Y t )|a∈[1,A]}, where BC a (Y t ) represents the user's historical online activity Y t The ath content category browsed at the time, A represents the user's historical online activity Y t The total number of content types viewed at that time.

[0011] As a preferred solution of the online self-media interactive management method based on artificial intelligence described in the present invention, based on the user's historical online activity Y t Content types viewed at BC a (Y t ), calculate the content type BC a (Y t ) in History Online Activity Y t The calculation formula is as follows:

[0012]

[0013] Among them, P a (Y t ) indicates the content type BC a (Y t ) in History Online Activity Y t The proportion of time in Represents the user's historical online activities Y t Browse content types BC a (Y t ) duration.

[0014] Through the user's historical online activities Y t The duration ratio of each content type browsed during the browsing time is calculated, and a content type-duration ratio coordinate system is constructed.

[0015] The horizontal axis of the content type-duration ratio coordinate system is the user's historical online activity Y t The vertical axis of the content type-duration ratio coordinate system is the user's historical online activity Y t The duration ratio of each content type.

[0016] It should be noted that this formula calculates the content type BC a (Y t ) in History Online Activity Y t The proportion of time in Represents the user's historical online activities Yt Browse content types BC a (Y t ) duration, the denominator ET(Y t )-ST(Y t ) indicates historical online activity Y t By calculating the proportion of time spent on different content types, we can understand the user's preference for different content types during a specific online activity time period, thereby providing basic data for subsequent user behavior analysis and content recommendation. In practical applications, such as online media platforms, this formula can be used to analyze the proportion of users' browsing time on different content types such as news, videos, and articles in different time periods, so that the platform can optimize content recommendation strategies based on these data and improve user experience. For example, if it is found that users spend a high proportion of time watching video content between 7 and 9 p.m., the platform can give priority to recommending popular videos during this time period.

[0017] Obtain the coordinate points of the content types and the duration ratios corresponding to the content types in the content type-duration ratio coordinate system; obtain all the coordinate points, and connect them in sequence to form a fluctuation curve, and record the fluctuation curve as the online activity Y t Content ratio fluctuation curve CRC(Y t ).

[0018] Get the content ratio fluctuation curve CRC(Y t ) is recorded as the high-frequency content category BC a,top [CRC(Y t )], among which BC a,top [CRC(Y t )] represents the content ratio fluctuation curve CRC(Y t ) is the ath content category; let t = t + 1, obtain all high-frequency content categories of the user during the total T online activities in history, and construct a high-frequency content category set, denoted as HCC = {BC a,top [CRC(Y t )]|a∈[1,A],t∈[1,T]}.

[0019] It should be noted that, firstly, by analyzing the proportion of content browsing time in users' historical online activities, the proportion of different content types is determined, and then the content type corresponding to the peak is found from the content proportion fluctuation curve, which helps to screen out the content types that users are most interested in or most frequently browsed. These high-frequency content types are the key to subsequent accurate content recommendation. For example, on an e-commerce platform, this method can find out the product categories that users browse most frequently, such as electronic products, clothing, etc.; and it provides an important basis for user behavior analysis. The platform can understand the core needs and consumption tendencies of users based on high-frequency content types; in actual e-commerce operations, by determining high-frequency content types, the platform can optimize the display of these popular product categories, such as focusing on displaying product categories that users are interested in in homepage recommendations, personalized recommendations, etc., to increase product exposure and sales conversion rate; at the same time, it also has guiding significance for the planning of marketing activities, such as launching promotional activities for high-frequency browsed product categories to attract more users to participate.

[0020] As a preferred solution of the online self-media interactive management method based on artificial intelligence described in the present invention, based on the high-frequency content category set HCC={BC a,top [CRC(Y t )]|a∈[1,A],t∈[1,T]}, calculate the high-frequency content category BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence is calculated as follows:

[0021]

[0022] Among them, P(BC a,top [CRC(Y T+1 )]|HCC) indicates high frequency content category BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence when P(BC a,top [CRC(Y T+1 )]∩HCC) represents the high-frequency content category BC a,top [CRC(Y t )] represents the probability of appearing in the high-frequency content category set HCC, and P(HCC) represents the probability of appearing in the high-frequency content category set HCC.

[0023] It should be noted that this formula can predict the user's next online activity behavior, especially the content that the user may be interested in. For example, on social media platforms, this method can predict the hot topics or content types that the user may be interested in the next time he logs in. It provides a decision-making basis for the content recommendation system, so that the recommendation system can prepare and push content that the user may be interested in in advance. In social media operations, it can prepare content that the user may be interested in in advance, and improve the timeliness and accuracy of content push. For example, if it is predicted that the user has a high probability of being interested in a certain type of news (such as technology news) the next time he logs in, the platform can push relevant news to the user in advance, thereby increasing the user's stickiness and activity on the platform, and at the same time improving the user's satisfaction with the recommended content.

[0024] Preset probability threshold, if high-frequency content type BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence P(BC a,top [CRC(Y T+1 )]|HCC) is greater than the probability threshold, then in the user's next online activity Y T+1 When the user clicks the ath content type, the ath content type is recommended to the user first.

[0025] It should be noted that by setting a probability threshold, the predicted probability of high-frequency content types appearing in the next online activity is screened. When the probability is greater than the set threshold, it is considered that the content type has a sufficiently high probability of being of interest to the user, and is thus given priority recommendation. This step avoids excessive recommendations or recommendations of irrelevant content and improves the accuracy of recommendations. For example, on an online music platform, by setting a probability threshold, it can ensure that the music recommended to users is what the users are really likely to like, rather than randomly recommending a large amount of music. It can also optimize the performance of the recommendation system, reduce unnecessary calculations and recommendation operations, and improve system operating efficiency.

[0026] An online self-media interactive management system based on artificial intelligence, the system includes: a user activity data collection module, a historical activity information processing module, a content analysis and feature extraction module, and a content recommendation and optimization module.

[0027] The user activity data collection module obtains the user's network log in the online self-media after authorization by the user, extracts the start time, end time and browsing content type of the user's historical online activities, and constructs a browsing content type set and activity information association library.

[0028] The historical activity information processing module is used to initialize the user's online activity history time sequence and add time tags to the start time, end time and browsing content type.

[0029] The content analysis and feature extraction module calculates the duration ratio of the content type in historical online activities, and constructs a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtains high-frequency content types corresponding to the peaks of the content ratio fluctuation curve, and constructs a high-frequency content type set.

[0030] The content recommendation and optimization module calculates the probability of occurrence of the high-frequency content type during the user's next online activity, presets a threshold, and analyzes and recommends content types.

[0031] Furthermore, the user activity data collection module includes a log extraction unit and an activity information association library generation unit.

[0032] The log extraction unit: after authorization by the user, obtains the network log of the user's online activities in the online self-media, extracts the start time and end time of the user's single online activity from the network log, and uses natural language processing technology to extract all the browsing content types of the user during the single online activity, and constructs a browsing content type set.

[0033] The activity information association library generating unit: associates the start time, end time and browsed content type set of the user during the single online activity, and constructs an activity file generated by the user during the single online activity, wherein the activity file records the associated start time, end time and browsed content type set of the user during the single online activity; and constructs an activity information association library, wherein the activity information association library records the activity files generated by the user during each online activity in chronological order.

[0034] Furthermore, the historical activity information processing module includes an encoding unit.

[0035] The encoding unit initializes the user's online activity history time series, extracts the activity files generated by the user's historical online activities based on the activity information association library, and encodes the browsing content category set recorded in the activity files.

[0036] Furthermore, the content analysis and feature extraction module includes a coordinate system construction unit and a high-frequency content category set construction unit.

[0037] The coordinate system construction unit: calculates the time proportion of the content type in the historical online activity based on the content type browsed by the user during the historical online activity; constructs a content type-time proportion coordinate system through the time proportion of each content type browsed by the user during the historical online activity; the horizontal axis of the content type-time proportion coordinate system is the content type browsed by the user during the historical online activity, and the vertical axis of the content type-time proportion coordinate system is the time proportion corresponding to the content type during the user's historical online activity.

[0038] The high-frequency content category set construction unit: obtains the coordinate points of the content category and the duration ratio corresponding to the content category in the content category-duration ratio coordinate system; obtains all the coordinate points and connects them in sequence to form a fluctuation curve, and records the fluctuation curve as the content ratio fluctuation curve during online activities; obtains the content category corresponding to the peak of the content ratio fluctuation curve, and records it as the high-frequency content category; obtains all the high-frequency content categories of the user in all historical online activities, and constructs a high-frequency content category set.

[0039] Furthermore, the probability calculation and content recommendation module includes a probability calculation unit and a content recommendation unit.

[0040] The probability calculation unit calculates the probability of occurrence of the high-frequency content category during the user's next online activity based on the high-frequency content category set.

[0041] The content recommendation unit presets a probability threshold, and if the probability of a high-frequency content type appearing in the user's next online activity is greater than the probability threshold, the high-frequency content type is preferentially recommended to the user in the user's next online activity.

[0042] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: in an online self-media interactive management system and method based on artificial intelligence provided by the present invention, by obtaining user authorization and extracting network logs, analyzing the start time, end time and browsing content types of users in historical online activities, constructing a browsing content type set and an activity information association library, the systematic organization of user activity data is realized, and a reliable data basis is provided for subsequent analysis. Secondly, by initializing the user's online activity history time series, adding time tags to key time nodes, ensuring the time correlation of activity data, and providing accurate support for dynamic behavior analysis; by calculating the duration ratio of content types, generating a content ratio fluctuation curve and extracting a high-frequency content type set, the user's content preference can be intuitively presented and its changing rules can be quantified, which is helpful to identify user interest hotspots; by analyzing the probability of occurrence of high-frequency content types in the user's next online activity, and combining with preset thresholds, the content types are intelligently recommended to users, realizing personalized content recommendation, and improving the accuracy of user experience and content push; this method combines historical behavior analysis, time series modeling and intelligent recommendation technology, achieving in-depth insight into user behavior and optimization of interactive experience, and effectively enhancing the interactivity and attractiveness of online self-media. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0044] Figure 1 It is a schematic diagram of the steps of an online self-media interactive management method based on artificial intelligence of the present invention;

[0045] Figure 2 It is a structural schematic diagram of an online self-media interactive management system based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1 In the first embodiment of the present invention, an online self-media interaction management method based on artificial intelligence is provided, and the method comprises the following steps:

[0048] Step S1: After authorization by the user, obtain the user's network log in the online self-media, extract the start time, end time and browsing content type of the user's historical online activities; and build a browsing content type set and activity information association library.

[0049] Specifically, after the user's authorization, the network log of the user's online activities in the online self-media is obtained, the start time and end time of the user's single online activity are extracted from the network log, and the natural language processing technology is used to extract all the types of content browsed by the user during the single online activity, and a set of browsed content types is constructed.

[0050] Furthermore, the start time, end time and browsed content category set of the user during the single online activity are associated to construct an activity profile generated by the user during the single online activity, wherein the activity profile records the associated start time, end time and browsed content category set of the user during the single online activity; and an activity information association library is constructed, wherein the activity information association library records the activity profile generated by the user during each online activity in chronological order.

[0051] Step S2: Initialize the user's online activity history time sequence, and add time tags to the start time, end time, and browsing content type.

[0052] Specifically, initialize the user's online activity history time series, denoted as TI = {Y t |t∈[1,T]}, where Y t represents the user’s t-th historical online activity, and T represents the total number of historical online activities of the user.

[0053] Further, based on the activity information association library, extract the user's historical online activity Y t The activity file generated when the browsing content type set recorded in the activity file is recorded as Q[ST(Y t )→ET(Y t )], where ST(Y t ) represents the user's historical online activity Y t The starting time of ET(Y t ) represents the user's historical online activity Y t The end time of Q[ST(Y t )→ET(Y t )]={BC a (Y t )|a∈[1,A]}, where BC a (Y t ) represents the user's historical online activity Y tThe ath content category browsed at the time, A represents the user's historical online activity Y t The total number of content types viewed at that time.

[0054] Step S3: Calculate the duration ratio of the content type in historical online activities, and construct a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtain high-frequency content types corresponding to the peaks of the content ratio fluctuation curve, and construct a high-frequency content type set.

[0055] Specifically, based on the user's historical online activity Y t Content types viewed at BC a (Y t ), calculate the content type BC a (Y t ) in History Online Activity Y t The calculation formula is as follows:

[0056]

[0057] Among them, P a (Y t ) indicates the content type BC a (Y t ) in History Online Activity Y t The proportion of time in Represents the user's historical online activities Y t Browse content types BC a (Y t ) duration.

[0058] For example, suppose a user has historical online activity Y t The starting time ST(Y t ) is 9:00, the user's historical online activity is Y t The end time ET(Y t ) is 10:00, the user's historical online activity Y t Browse content types BC a (Y t ) is 0.5 hours, and the content type BC is calculated by substituting it into the formula a (Y t ) in History Online Activity Y t The proportion of time in That is, the duration accounts for 50%.

[0059] Furthermore, through the user's historical online activities Y t The duration ratio of each content type browsed during the browsing time is calculated, and a content type-duration ratio coordinate system is constructed.

[0060] The horizontal axis of the content type-duration ratio coordinate system is the user's historical online activity Y t The vertical axis of the content type-duration ratio coordinate system is the user's historical online activity Y t The duration ratio of each content type.

[0061] Further, the coordinate points of the content types and the duration proportions corresponding to the content types in the content type-duration proportion coordinate system are obtained; all the coordinate points are obtained and connected in sequence to form a fluctuation curve, and the fluctuation curve is recorded as the online activity Y t Content ratio fluctuation curve CRC(Y t ).

[0062] Get the content ratio fluctuation curve CRC(Y t ) is recorded as the high-frequency content category BC a,top [CRC(Y t )], among which BC a,top [CRC(Y t )] represents the content ratio fluctuation curve CRC(Y t ) is the ath content category; let t = t + 1, obtain all high-frequency content categories of the user during the total T online activities in history, and construct a high-frequency content category set, denoted as HCC = {BC a,top [CRC(Y t )]|a∈[1,A],t∈[1,T]}.

[0063] Step S4: Calculate the probability of occurrence of the high-frequency content type in the user's next online activity, preset a threshold, analyze and recommend content types.

[0064] Specifically, based on the high-frequency content category set HCC = {BC a,top [CRC(Y t )]|a∈[1,A],t∈[1,T]}, calculate the high-frequency content category BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence is calculated as follows:

[0065]

[0066] Among them, P(BC a,top [CRC(Y T+1 )]|HCC) indicates high frequency content category BC a,top [CRC(Y t )] During the user's next online activity Y T+1The probability of occurrence when P(BC a,top [CRC(Y T+1 )]∩HCC) represents the high-frequency content category BC a,top [CRC(Y t )] represents the probability of appearing in the high-frequency content category set HCC, and P(HCC) represents the probability of appearing in the high-frequency content category set HCC.

[0067] Furthermore, a probability threshold is preset, if the high-frequency content type BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence P(BC a,top [CRC(Y T+1 )]|HCC) is greater than the probability threshold, then in the user's next online activity Y T+1 When the user clicks the ath content type, the ath content type is recommended to the user first.

[0068] See also Figure 2 In the second embodiment of the present invention, an online self-media interactive management system based on artificial intelligence is provided, which includes: a user activity data collection module, a historical activity information processing module, a content analysis and feature extraction module, and a content recommendation and optimization module.

[0069] The user activity data collection module obtains the user's network log in the online self-media after authorization by the user, extracts the start time, end time and browsing content type of the user's historical online activities, and constructs a browsing content type set and activity information association library.

[0070] The historical activity information processing module is used to initialize the user's online activity history time sequence and add time tags to the start time, end time and browsing content type.

[0071] The content analysis and feature extraction module calculates the duration ratio of the content type in historical online activities, and constructs a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtains high-frequency content types corresponding to the peaks of the content ratio fluctuation curve, and constructs a high-frequency content type set.

[0072] The content recommendation and optimization module calculates the probability of occurrence of the high-frequency content type during the user's next online activity, presets a threshold, and analyzes and recommends content types.

[0073] Furthermore, the user activity data collection module includes a log extraction unit and an activity information association library generation unit.

[0074] The log extraction unit: after authorization by the user, obtains the network log of the user's online activities in the online self-media, extracts the start time and end time of the user's single online activity from the network log, and uses natural language processing technology to extract all the browsing content types of the user during the single online activity, and constructs a browsing content type set.

[0075] The activity information association library generating unit: associates the start time, end time and browsed content type set of the user during the single online activity, and constructs an activity file generated by the user during the single online activity, wherein the activity file records the associated start time, end time and browsed content type set of the user during the single online activity; and constructs an activity information association library, wherein the activity information association library records the activity files generated by the user during each online activity in chronological order.

[0076] Furthermore, the historical activity information processing module includes an encoding unit.

[0077] The encoding unit initializes the user's online activity history time series, extracts the activity files generated by the user's historical online activities based on the activity information association library, and encodes the browsing content category set recorded in the activity files.

[0078] Furthermore, the content analysis and feature extraction module includes a coordinate system construction unit and a high-frequency content category set construction unit.

[0079] The coordinate system construction unit: calculates the time proportion of the content type in the historical online activity based on the content type browsed by the user during the historical online activity; constructs a content type-time proportion coordinate system through the time proportion of each content type browsed by the user during the historical online activity; the horizontal axis of the content type-time proportion coordinate system is the content type browsed by the user during the historical online activity, and the vertical axis of the content type-time proportion coordinate system is the time proportion corresponding to the content type during the user's historical online activity.

[0080] The high-frequency content category set construction unit: obtains the coordinate points of the content category and the duration ratio corresponding to the content category in the content category-duration ratio coordinate system; obtains all the coordinate points and connects them in sequence to form a fluctuation curve, and records the fluctuation curve as the content ratio fluctuation curve during online activities; obtains the content category corresponding to the peak of the content ratio fluctuation curve, and records it as the high-frequency content category; obtains all the high-frequency content categories of the user in all historical online activities, and constructs a high-frequency content category set.

[0081] Furthermore, the probability calculation and content recommendation module includes a probability calculation unit and a content recommendation unit.

[0082] The probability calculation unit calculates the probability of occurrence of the high-frequency content category during the user's next online activity based on the high-frequency content category set.

[0083] The content recommendation unit presets a probability threshold, and if the probability of a high-frequency content type appearing in the user's next online activity is greater than the probability threshold, the high-frequency content type is preferentially recommended to the user in the user's next online activity.

[0084] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0085] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An online self-media interactive management method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1: After the user's authorization, obtain the user's network log in the online self-media, extract the start time, end time and browsing content type of the user's historical online activities; build a browsing content type set and activity information association library; Step S2: Initialize the user's online activity history time series, and add time tags to the start time, end time, and browsing content type; Step S3: Calculate the duration ratio of the content type in the historical online activities, and construct a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtain the high-frequency content types corresponding to the peaks of the content ratio fluctuation curve, and construct a high-frequency content type set; Step S4: Calculate the probability of occurrence of the high-frequency content type in the user's next online activity, preset a threshold, analyze and recommend content types.

2. According to the artificial intelligence-based online self-media interactive management method of claim 1, it is characterized in that: The specific implementation process of step S1 includes: After the user's authorization, the network log of the user's online activities in the online self-media is obtained, the start time and end time of the user's single online activity are extracted from the network log, and all the types of content browsed by the user during the single online activity are extracted using natural language processing technology to construct a set of browsed content types; The start time, end time and browsed content category set of the user during the single online activity are associated to construct an activity profile generated by the user during the single online activity, wherein the activity profile records the associated start time, end time and browsed content category set of the user during the single online activity; and an activity information association library is constructed, wherein the activity information association library records the activity profile generated by the user during each online activity in chronological order.

3. The online self-media interactive management method based on artificial intelligence according to claim 2 is characterized in that: The specific implementation process of step S2 includes: Initialize the user's online activity history time series, denoted as TI = {Y t |t∈[1,T]}, where Y t represents the user's t-th historical online activity, and T represents the total number of historical online activities of the user; Based on the activity information association library, extract the user's historical online activity Y t The activity file generated when the browsing content type set recorded in the activity file is recorded as Q[ST(Y t )→ET(Y t )], where ST(Y t ) represents the user's historical online activity Y t The starting time of ET(Y t ) represents the user's historical online activity Y t The end time of Q[ST(Y t )→ET(Y t )]={BC a (Y t )|a∈[1,A]}, where BC a (Y t ) represents the user's historical online activity Y t The ath content category browsed at the time, A represents the user's historical online activity Y t The total number of content types viewed at that time.

4. The online self-media interactive management method based on artificial intelligence according to claim 3 is characterized in that: The specific implementation process of step S3 includes: Based on the user's historical online activities Y t Content types viewed at BC a (Y t ), calculate the content type BC a (Y t ) in History Online Activity Y t The calculation formula is as follows: Among them, P a (Y t ) indicates the content type BC a (Y t ) in History Online Activity Y t The proportion of time in Represents the user's historical online activities Y t Browse content type BC a (Y t ) duration; Through the user's historical online activities Y t The proportion of time spent browsing each content type during the browsing period, and the content type-time proportion coordinate system are constructed; The horizontal axis of the content type-duration ratio coordinate system is the user's historical online activity Y t The vertical axis of the content type-duration ratio coordinate system is the user's historical online activity Y t The proportion of duration corresponding to the content type; Obtain the coordinate points of the content types and the duration ratios corresponding to the content types in the content type-duration ratio coordinate system; obtain all the coordinate points, and connect them in sequence to form a fluctuation curve, and record the fluctuation curve as the online activity Y t Content ratio fluctuation curve CRC(Y t ); Get the content ratio fluctuation curve CRC(Y t ) is recorded as the high-frequency content category BC a,top [CRC(Y t )], among which BC a,top [CRC(Y t )] represents the content ratio fluctuation curve CRC(Y t ) is the ath content category; let t = t + 1, obtain all high-frequency content categories of the user during the total T online activities in history, and construct a high-frequency content category set, denoted as HCC = {BC a,top [CRC(Y t )]|a∈[1,A],t∈[1,T]}.

5. The online self-media interactive management method based on artificial intelligence according to claim 4 is characterized in that: The specific implementation process of step S4 includes: Based on the high-frequency content category set HCC = {BC a,top [CRC(Y t )]|a∈[1,A],t∈[1,T]}, calculate the high-frequency content category BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence is calculated as follows: Among them, P(BC a,top [CRC(Y T+1 )]|HCC) indicates high frequency content category BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence when a,top [CRC(Y T+1 )]∩HCC) represents the high-frequency content category BC a,top [CRC(Y t )] in the high-frequency content category set HCC, P(HCC) represents the probability of the high-frequency content category set HCC appearing; Preset probability threshold, if high-frequency content type BC a,top [CRC(Y t )] During the user's next online activity Y T+1 The probability of occurrence P(BC a,top [CRC(Y T+1 )]|HCC) is greater than the probability threshold, then in the user's next online activity Y T+1 When the user clicks the ath content type, the ath content type is recommended to the user first.

6. An online self-media interaction management system based on artificial intelligence, executing an online self-media interaction management method based on artificial intelligence as described in any one of claims 1 to 5, characterized in that: The system includes: a user activity data collection module, a historical activity information processing module, a content analysis and feature extraction module, and a content recommendation and optimization module; The user activity data collection module: after authorization by the user, obtains the user's network log in the online self-media, extracts the start time, end time and browsed content type of the user's historical online activities; and constructs a browsed content type set and activity information association library; The historical activity information processing module initializes the user's online activity history time sequence and adds time tags to the start time, end time and browsing content category; The content analysis and feature extraction module calculates the duration ratio of the content type in historical online activities, and constructs a content type-duration ratio coordinate system and a content ratio fluctuation curve; obtains high-frequency content types corresponding to the peaks of the content ratio fluctuation curve, and constructs a high-frequency content type set; The content recommendation and optimization module calculates the probability of occurrence of the high-frequency content type during the user's next online activity, presets a threshold, and analyzes and recommends content types.

7. The online self-media interactive management system based on artificial intelligence according to claim 6 is characterized by: The user activity data collection module includes a log extraction unit and an activity information association library generation unit; The log extraction unit: after authorization by the user, obtains the network log of the user's online activities in the online self-media, extracts the start time and end time of the user's single online activity from the network log, and uses natural language processing technology to extract all the types of browsed content of the user during the single online activity, and constructs a browsed content type set; The activity information association library generating unit: associates the start time, end time and browsed content type set of the user during the single online activity, and constructs an activity file generated by the user during the single online activity, wherein the activity file records the associated start time, end time and browsed content type set of the user during the single online activity; and constructs an activity information association library, wherein the activity information association library records the activity files generated by the user during each online activity in chronological order.

8. The online self-media interactive management system based on artificial intelligence according to claim 7, characterized in that: The historical activity information processing module includes an encoding unit; The encoding unit initializes the user's online activity history time series, extracts the activity files generated by the user's historical online activities based on the activity information association library, and encodes the browsing content category set recorded in the activity files.

9. The online self-media interactive management system based on artificial intelligence according to claim 8, characterized in that: The content analysis and feature extraction module includes a coordinate system construction unit and a high-frequency content category set construction unit; The coordinate system construction unit: calculates the duration ratio of the content type in the historical online activities based on the content type browsed by the user during the historical online activities; constructs a content type-duration ratio coordinate system according to the duration ratio of each content type browsed by the user during the historical online activities; the horizontal coordinate of the content type-duration ratio coordinate system is the content type browsed by the user during the historical online activities, and the vertical coordinate of the content type-duration ratio coordinate system is the duration ratio corresponding to the content type during the historical online activities of the user; The high-frequency content category set construction unit: obtains coordinate points of content categories and duration ratios corresponding to the content categories in the content category-duration ratio coordinate system; Obtain all coordinate points, and connect them in sequence to form a fluctuation curve, and record the fluctuation curve as the content proportion fluctuation curve during online activities; obtain the content type corresponding to the peak of the content proportion fluctuation curve, and record it as the high-frequency content type; Obtain all high-frequency content categories of the user in all historical online activities and build a high-frequency content category set.

10. The online self-media interactive management system based on artificial intelligence according to claim 9, characterized in that: The probability calculation and content recommendation module includes a probability calculation unit and a content recommendation unit; The probability calculation unit calculates the probability of occurrence of the high-frequency content category in the next online activity of the user based on the high-frequency content category set; The content recommendation unit: presets a probability threshold, and if the probability of a high-frequency content type appearing in the user's next online activity is greater than the probability threshold, the high-frequency content type is preferentially recommended to the user in the user's next online activity.

Images