Science popularization content personalized recommendation system and method based on big data

Abstract

The invention relates to the technical field of personalized recommendation, and particularly discloses a popular science content personalized recommendation system and method based on big data, and the method comprises the following steps: S1, obtaining a registration date and a current date, calculating a use date, obtaining a browsing number, calculating a daily browsing amount, and screening out key users according to the browsing amount; s2, obtaining the total time consumed by the key users to browse the popular science content, calculating the average time cost, and calculating the browsing intention value and the attention value of the key users; and S3, obtaining the total number of times of browsing the popular science content and the number of times of forwarding, calculating the quality coefficient of the popular science content, obtaining a recommendation value, and screening out the recommendation content. According to the method, the recommendation strategy and method are adjusted according to the interests and hobbies of the user, so that the science popularization content which is high in quality and conforms to the interests of the user is screened out, the purpose of accurate pushing is achieved, and the user obtains better use experience.

Description

Technical Field

[0001] The present invention relates to the field of personalized recommendation technology, and in particular to a popular science content personalized recommendation system based on big data and a method thereof. Background Art

[0002] With the rapid development of the Internet, major websites have launched personalized science content services, aiming to recommend the most likely science content to users based on their past visit records on the website. However, although these personalized recommendation services have improved the user experience to a certain extent, the recommended data they provide is often very sparse and single.

[0003] Due to the limitations of user access records, the recommendation system may only be able to analyze and recommend based on limited data. This means that if the user's past access behavior is relatively single or concentrated in a specific field, it is difficult for the recommendation system to provide comprehensive and diverse popular science content recommendations. In addition, the sparsity of recommended data may also lead to a decrease in the accuracy of recommendation results. Due to the limited amount of data, the recommendation system may find it difficult to accurately judge the user's real interests and needs, and thus give some recommendations that do not meet the user's expectations.

[0004] In the prior art, although personalized popular science content services have improved user experience to a certain extent, the content recommended by popular science websites to users usually has great limitations. They can only characterize and understand a user from a few fewer dimensions, and it is difficult to accurately judge the user's interests and attributes. Such recommendation results are often not accurate enough. In addition, the user's interests and hobbies may gradually change over time. Therefore, it is necessary to continuously adjust the recommendation strategy and method according to the user's interests and hobbies, so as to achieve the purpose of accurate push and give users a better user experience. Summary of the invention

[0005] The purpose of the present invention is to provide a personalized recommendation system for popular science content based on big data and a method thereof to solve the above technical problems.

[0006] The purpose of the present invention can be achieved by the following technical solutions: A method for personalized recommendation of popular science content based on big data, comprising the following steps: S1: Get the user's registration date D on the popular science website L and the current date D N , calculate the usage date D=D N -D L , get the user's registration date D L To current dateD NThe number of views of the popular science content in the content is N, and the popular science content includes popular science articles and popular science videos. The daily views are calculated as R = N / D, and the daily views R>R sta The users are recorded as key users, among which R sta is the preset judgment value; S2: Obtain the total time t spent by key users browsing popular science content and calculate the average time spent t ave =t / N, calculate the browsing intention value of key users for popular science content , where ts represents the duration of the popular science video, A 1 , A 2 is the preset first and second intention coefficients, A 1 <0, B is the preset intention threshold; The popular science content belonging to the same section in the popular science website is recorded as the same type of content. The section refers to the classification directory in the navigation bar of the popular science website. The attention value of key users to the same type of content is calculated. , where λ represents the preset adjustment coefficient and n represents the number of contents of the same type; S3: Obtain the ranking of the number of views of the popular science website, obtain the total number of views S and forwarding times Sp of the popular science content in the preset time period T in the ranking of the number of views, and calculate the quality coefficient of the popular science content , where γ 1 , γ 2 Represent the preset first and second quality coefficients respectively, 0<γ 1 <γ 2 , P represents the percentage of completed science content, η represents the preset amplification factor; calculate the recommended value of science content , where G s The attention value corresponding to the same type of content to which the popular science content belongs is calculated. The popular science content is sorted from high to low according to the recommendation value, and the top 10% of the popular science content is selected as the recommended content.

[0007] As a further solution of the present invention: in the step S1, if the number of popular science content browsed by the user N<N min Then stop the subsequent operation for this user, where N min Represents the preset minimum number of page views.

[0008] As a further solution of the present invention: in step S2, the user browsing time t<t min The popular science content will be removed and will not be included in the registration date D L To current dateD N The number of views of popular science content in the article is N, where t min Represents the preset minimum browsing time.

[0009] As a further solution of the present invention: in the step S2, the popular science content that the user has browsed is marked as duplicate content, and the recommendation value Rec=0.5*Re of the duplicate content is calculated. As a further solution of the present invention: in the step S3, if there are popular science contents with equal recommendation values, the popular science content with higher attention value is ranked first.

[0010] As a further solution of the present invention: in step S1, the daily page views R≤R sta The users are recorded as non-key users, and the popular science contents are sorted from high to low according to the quality coefficient, and the top 10% of the popular science contents are selected as the recommended contents for non-key users.

[0011] As a further solution of the present invention: in step S2, the maximum time t of key users browsing the popular science content is obtained. max , if the duration of the popular science video is ts>2*t max , it is not recommended content.

[0012] A personalized recommendation system for popular science content based on big data, comprising: Filter module: Get the user's registration date on the popular science website D L and the current date D N , calculate the usage date D=D N -D L , get the user's registration date D L To current dateD N The number of views of the popular science content in the content is N, and the popular science content includes popular science articles and popular science videos. The daily views are calculated as R = N / D, and the daily views R>R sta The users are recorded as key users, among which R sta is the preset judgment value; Calculation module: Obtain the total time t spent by key users browsing popular science content and calculate the average time spent t ave =t / N, calculate the browsing intention value of key users for popular science content , where ts represents the duration of the popular science video, A 1 , A 2 is the preset first and second intention coefficients, A 1 <0, B is the preset intention threshold; The popular science contents belonging to the same section in the popular science website are recorded as the same type of content. The section refers to the classification directory in the navigation bar of the popular science website. The attention value of key users to the same type of content is calculated as G=λ*Y*n, where λ represents the preset adjustment coefficient and n represents the number of the same type of content; Recommendation module: obtain the ranking of the number of views of the popular science website, obtain the total number of views S and forwarding times Sp of the popular science content in the preset time period T in the ranking of the number of views, and calculate the quality coefficient of the popular science content , where γ 1 , γ 2 Represent the preset first and second quality coefficients respectively, 0<γ 1 <γ 2 , P represents the percentage of completed science content, and η represents the preset amplification factor; Calculate the recommended value of popular science content , where G s The attention value corresponding to the same type of content to which the popular science content belongs is calculated. The popular science content is sorted from high to low according to the recommendation value, and the top 10% of the popular science content is selected as the recommended content.

[0013] Beneficial effects of the present invention: In the present invention, firstly, in the process of operation management and user behavior analysis of the popular science website, in order to more accurately understand the user's activity and consumption of popular science content, a series of detailed data collection and calculation work needs to be carried out.

[0014] First, we need to obtain the registration date of each user on the popular science website and mark this registration date as DL. This date is the starting point for the user to establish contact with the popular science website, and it is of great significance for the subsequent analysis of the user's behavior trajectory and participation. At the same time, we also need to obtain the current date in real time and mark it as DN. The current date represents the time node for our data analysis. By comparing it with the registration date, we can clearly understand the user's duration on the website.

[0015] Next, we calculate the usage date D, where D reflects the time span from the user's registration to the current moment. This time span is an important basis for subsequent analysis of user browsing behavior.

[0016] Then, count the number of views N of popular science content that users have viewed from the registration date to the current date. It should be noted that the popular science content mentioned here covers a wide variety of forms, including popular science articles and popular science videos. By counting the number of views of various popular science content by users during this period, we can get a comprehensive number of views N, and then get the average number of popular science content that users view per day. This value can intuitively reflect the user's attention to popular science content and the frequency of use.

[0017] Finally, in order to screen out users with high interest and participation in popular science content, a preset judgment value is set. The calculated daily page views R are compared with Rsta. If R is greater than Rsta, the user is recorded as a key user. These key users may have greater potential and value in the dissemination and promotion of popular science knowledge and their own in-depth study of popular science content. Their behavior and needs are of great reference significance for the operation and development of popular science websites.

[0018] In the current network environment, users' time has become increasingly valuable, and the time they can devote to popular science content is relatively limited. This feature requires us to consider the time factor more carefully when analyzing and processing popular science content. Specifically, it is necessary to accurately assess the average time they spend on popular science content by digging deep into users' historical records. This not only includes the time users spend browsing popular science articles and watching popular science videos, but also covers multiple dimensions such as users' participation in popular science interactions, comments and feedback. Through this comprehensive and detailed analysis, we can obtain a more accurate portrait of users' time spending.

[0019] The reason why we emphasize that users have limited time to spend on popular science content is that if popular science content takes up too much time for users, it may cause users to feel bored, which in turn reduces their interest in browsing popular science content. This is undoubtedly a problem that popular science platforms need to avoid. Therefore, we must always pay attention to the time spent by users and ensure that popular science content can provide the greatest value in a limited time while maintaining users' browsing interest and participation.

[0020] In addition, we need to further understand the user's preference for the popular science content type that the user browses. This not only helps to understand the user's interest preferences more deeply, but also provides an important basis for subsequent popular science content recommendations. Through accurate recommendation algorithms, we can push popular science content that users are interested in to improve user satisfaction and loyalty.

[0021] It should be noted that in the present invention, there is no specific requirement for the division of popular science content. Instead, it is based on the division rules of the popular science platform itself. This means that no matter how the popular science platform classifies and manages popular science content, our method and system can flexibly adapt and provide users with personalized popular science content recommendations and services. This flexibility not only improves the adaptability of our system, but also reduces the difficulty of integration with different popular science platforms, providing strong support for the wide application of our method and system.

[0022] In order to more comprehensively evaluate the influence and user appeal of the content on a popular science website, we first need to obtain the website's playback rankings. This ranking is sorted by the playback volume of popular science content, and can intuitively show which content is most popular with users. The playback volume is not only an important indicator for measuring the popularity of popular science content, but also reflects the quality of the content from the side. Generally speaking, content with a high playback volume tends to have higher quality, stronger appeal, or wider topic coverage.

[0023] After obtaining the ranking of views, we need to further count the specific number of views and reposts of each popular science content in the ranking. Views reflect the user's interest and attention to the content, while reposts reflect the user's recognition of the content and willingness to share. These two data together constitute an important basis for evaluating the quality of popular science content. By comprehensively analyzing views and reposts, we can more accurately determine which content is truly popular and loved by users.

[0024] Next, the quality of popular science content is evaluated based on the calculated quality coefficient. The quality coefficient is a comprehensive indicator that takes into account multiple factors such as the number of views, the number of reposts, etc., and can more comprehensively reflect the quality of popular science content. In order to more accurately evaluate the quality of content, the user's attention value also needs to be considered. The attention value reflects the user's attention and interest in a specific topic or field, and it is an important indicator for measuring the matching degree between content and users.

[0025] Finally, the recommendation value of the popular science content is calculated according to the calculation formula of the recommendation value Re. In this formula, we use the quality coefficient Z as the base and the attention value G as the exponent. This calculation method can well show the comprehensive comparison between the quality of popular science content and the user's attention value. Only when the quality coefficient and the attention value are both high, the recommendation value will increase accordingly. This means that only content that meets both the conditions of high quality and high user attention can obtain a higher recommendation value. In this way, we can screen out popular science content that is both high-quality and in line with user interests, and recommend the most suitable content to users.

[0026] In summary, the present invention realizes a strategy and method for adjusting recommendations according to the interests of users, thereby screening out popular science content that is both high-quality and in line with the interests of users, achieving the purpose of accurate push, and enabling users to obtain a better user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention will be further described below in conjunction with the accompanying drawings.

[0028] Figure 1 It is a flow chart of a personalized recommendation system for popular science content based on big data and a method thereof of the present invention. DETAILED DESCRIPTION

[0029] 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.

[0030] See also Figure 1 As shown, the present invention is a method for personalized recommendation of popular science content based on big data, comprising the following steps: S1: Get the user's registration date D on the popular science website L and the current date D N , calculate the usage date D=D N -D L , get the user's registration date D L To current dateD N The number of views of the popular science content in the content is N, and the popular science content includes popular science articles and popular science videos. The daily views are calculated as R = N / D, and the daily views R>R sta The users are recorded as key users, among which R sta is the preset judgment value; S2: Obtain the total time t spent by key users browsing popular science content and calculate the average time spent t ave =t / N, calculate the browsing intention value of key users for popular science content , where ts represents the duration of the popular science video, A 1 , A 2 is the preset first and second intention coefficients, A 1 <0, B is the preset intention threshold; The popular science contents belonging to the same section in the popular science website are recorded as the same type of content. The section refers to the classification directory in the navigation bar of the popular science website. The attention value of key users to the same type of content is calculated as G=λ*Y*n, where λ represents the preset adjustment coefficient and n represents the number of the same type of content; S3: Obtain the ranking of the number of views of the popular science website, obtain the total number of views S and forwarding times Sp of the popular science content in the preset time period T in the ranking of the number of views, and calculate the quality coefficient of the popular science content , where γ 1 , γ 2 Represent the preset first and second quality coefficients respectively, 0<γ 1 <γ 2 , P represents the percentage of completed science content, and η represents the preset amplification factor; Calculate the recommended value of popular science content , where Gs The attention value corresponding to the same type of content to which the popular science content belongs is calculated. The popular science content is sorted from high to low according to the recommendation value, and the top 10% of the popular science content is selected as the recommended content.

[0031] It should be noted that, first of all, in the process of operation management and user behavior analysis of the popular science website, in order to more accurately understand the user's activity and consumption of popular science content, a series of detailed data collection and calculation work is needed.

[0032] First, we need to obtain the registration date of each user on the popular science website and mark this registration date as DL. This date is the starting point for the user to establish contact with the popular science website, and it is of great significance for the subsequent analysis of the user's behavior trajectory and participation. At the same time, we also need to obtain the current date in real time and mark it as DN. The current date represents the time node for our data analysis. By comparing it with the registration date, we can clearly understand the user's duration on the website.

[0033] Next, we calculate the usage date D, where D reflects the time span from the user's registration to the current moment. This time span is an important basis for subsequent analysis of user browsing behavior.

[0034] Then, count the number of views N of popular science content that users have viewed from the registration date to the current date. It should be noted that the popular science content mentioned here covers a wide variety of forms, including popular science articles and popular science videos. By counting the number of views of various popular science content by users during this period, we can get a comprehensive number of views N, and then get the average number of popular science content that users view per day. This value can intuitively reflect the user's attention to popular science content and the frequency of use.

[0035] Finally, in order to screen out users with high interest and participation in popular science content, a preset judgment value is set. The calculated daily page views R are compared with Rsta. If R is greater than Rsta, the user is recorded as a key user. These key users may have greater potential and value in the dissemination and promotion of popular science knowledge and their own in-depth study of popular science content. Their behavior and needs are of great reference significance for the operation and development of popular science websites.

[0036] In the current network environment, users' time has become increasingly valuable, and the time they can devote to popular science content is relatively limited. This feature requires us to consider the time factor more carefully when analyzing and processing popular science content. Specifically, it is necessary to accurately assess the average time they spend on popular science content by digging deep into users' historical records. This not only includes the time users spend browsing popular science articles and watching popular science videos, but also covers multiple dimensions such as users' participation in popular science interactions, comments and feedback. Through this comprehensive and detailed analysis, we can obtain a more accurate portrait of users' time spending.

[0037] The reason why we emphasize that users have limited time to spend on popular science content is that if popular science content takes up too much time for users, it may cause users to feel bored, which in turn reduces their interest in browsing popular science content. This is undoubtedly a problem that popular science platforms need to avoid. Therefore, we must always pay attention to the time spent by users and ensure that popular science content can provide the greatest value in a limited time while maintaining users' browsing interest and participation.

[0038] In addition, we need to further understand the user's preference for the popular science content type that the user browses. This not only helps to understand the user's interest preferences more deeply, but also provides an important basis for subsequent popular science content recommendations. Through accurate recommendation algorithms, we can push popular science content that users are interested in to improve user satisfaction and loyalty.

[0039] It should be noted that in the present invention, there is no specific requirement for the division of popular science content, but it is mainly based on the division rules of the popular science platform itself, and the types of popular science content are based on the popular science content being divided into specific navigation bars. This means that no matter how the popular science platform classifies and manages popular science content, our method and system can flexibly adapt and provide users with personalized popular science content recommendations and services. This flexibility not only improves the adaptability of our system, but also reduces the difficulty of integration with different popular science platforms, providing strong support for the wide application of our method and system.

[0040] In order to more comprehensively evaluate the influence and user appeal of the content on a popular science website, we first need to obtain the website's playback rankings. This ranking is sorted by the playback volume of popular science content, and can intuitively show which content is most popular with users. The playback volume is not only an important indicator for measuring the popularity of popular science content, but also reflects the quality of the content from the side. Generally speaking, content with a high playback volume tends to have higher quality, stronger appeal, or wider topic coverage.

[0041] After obtaining the ranking of views, we need to further count the specific number of views and reposts of each popular science content in the ranking. Views reflect the user's interest and attention to the content, while reposts reflect the user's recognition of the content and willingness to share. These two data together constitute an important basis for evaluating the quality of popular science content. By comprehensively analyzing views and reposts, we can more accurately determine which content is truly popular and loved by users.

[0042] Next, the quality of popular science content is evaluated based on the calculated quality coefficient. The quality coefficient is a comprehensive indicator that takes into account multiple factors such as the number of views, the number of reposts, etc., and can more comprehensively reflect the quality of popular science content. In order to more accurately evaluate the quality of content, the user's attention value also needs to be considered. The attention value reflects the user's attention and interest in a specific topic or field, and it is an important indicator for measuring the matching degree between content and users.

[0043] Finally, the recommendation value of the popular science content is calculated according to the calculation formula of the recommendation value Re. In this formula, we use the quality coefficient Z as the base and the attention value G as the exponent. This calculation method can well show the comprehensive comparison between the quality of popular science content and the user's attention value. Only when the quality coefficient and the attention value are both high, the recommendation value will increase accordingly. This means that only content that meets both the conditions of high quality and high user attention can obtain a higher recommendation value. In this way, we can screen out popular science content that is both high-quality and in line with user interests, and recommend the most suitable content to users.

[0044] In another preferred embodiment of the present invention, if the number of popular science content browsed by the user N<N min Then stop the subsequent operation for this user, where N min Represents the preset minimum number of page views.

[0045] It is worth noting that in the process of data analysis and processing, the size of the data has a crucial impact on the reliability and accuracy of the results. In particular, in the statistics of key indicators such as the number of views, views and forwarding of popular science content, if the available data is too small, then these indicators may not be able to fully and accurately reflect the actual quality of popular science content and the real feedback of users.

[0046] Therefore, we chose not to proceed in this case to avoid misleading conclusions due to insufficient data.

[0047] In another preferred embodiment of the present invention, the user browsing time t<t min The popular science content will be removed and will not be included in the registration date D L To current dateD NThe number of views of popular science content in the article is N, where t min Represents the preset minimum browsing time.

[0048] It is understandable that in actual operation, we found that the browsing time of some popular science content was too short, which may not be a reflection of the user's true intention, but caused by misoperation.

[0049] By eliminating these abnormal data, we can more accurately assess users’ interest and participation in popular science content, and provide more accurate and reliable basis for subsequent analysis and recommendation. This not only helps improve the quality of data analysis, but also ensures that the content recommended to users is more in line with their real needs and interests, thereby enhancing user satisfaction with popular science websites.

[0050] In another preferred embodiment of the present invention, the popular science contents browsed by the user are marked as duplicate contents, and the recommendation value Rec=0.5*Re of the duplicate contents is calculated.

[0051] It is important to reduce the probability of duplicate content being recommended to users, and avoid users being recommended the same popular science content repeatedly, thereby reducing users' favorability towards the popular science website.

[0052] In another preferred embodiment of the present invention, if there are popular science contents with equal recommendation values, the popular science contents with higher attention values ​​are ranked higher.

[0053] It should be noted that popular science content with a higher attention value is content that users like more. Therefore, when there are popular science content with equal recommendation values, the popular science content that users like more should be recommended to them.

[0054] In another preferred embodiment of the present invention, the daily page views R≤R sta The users are recorded as non-key users, and the popular science contents are sorted from high to low according to the quality coefficient, and the top 10% of the popular science contents are selected as the recommended contents for non-key users.

[0055] It is understandable that for users with fewer views, it is difficult to determine their hobbies based on their browsing history. Therefore, we can only make recommendations based on the quality of the popular science content. We can recommend some high-quality popular science content. After the daily views reach a certain range, we can filter out the popular science content that users like based on their browsing preferences.

[0056] In another preferred embodiment of the present invention, the maximum time t for key users to browse popular science content is obtained. max , if the duration of the popular science video is ts>2*t max , it is not recommended content.

[0057] It is worth noting that if the duration of a video is much longer than what the user can accept, in order to avoid invalid recommendations, popular science videos with longer duration will be removed and not recommended.

[0058] A personalized recommendation system for popular science content based on big data, comprising: Filter module: Get the user's registration date on the popular science website D L and the current date D N , calculate the usage date D=D N -D L , get the user's registration date D L To current dateD N The number of views of the popular science content in the content is N, and the popular science content includes popular science articles and popular science videos. The daily views are calculated as R = N / D, and the daily views R>R sta The users are recorded as key users, among which R sta is the preset judgment value; Calculation module: Obtain the total time t spent by key users browsing popular science content and calculate the average time spent t ave =t / N, calculate the browsing intention value of key users for popular science content , where ts represents the duration of the popular science video, A 1 , A 2 is the preset first and second intention coefficients, A 1 <0, B is the preset intention threshold; The popular science contents belonging to the same section in the popular science website are recorded as the same type of content. The section refers to the classification directory in the navigation bar of the popular science website. The attention value of key users to the same type of content is calculated as G=λ*Y*n, where λ represents the preset adjustment coefficient and n represents the number of the same type of content; Recommendation module: obtain the ranking of the number of views of the popular science website, obtain the total number of views S and forwarding times Sp of the popular science content in the preset time period T in the ranking of the number of views, and calculate the quality coefficient of the popular science content , where γ 1 , γ 2 Represent the preset first and second quality coefficients respectively, 0<γ 1 <γ 2 , P represents the percentage of completed science content, and η represents the preset amplification factor; Calculate the recommended value of popular science content , where G s The attention value corresponding to the same type of content to which the popular science content belongs is calculated. The popular science content is sorted from high to low according to the recommendation value, and the top 10% of the popular science content is selected as the recommended content.

[0059] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A personalized recommendation method for popular science content based on big data, characterized in that: The following steps are involved: S1: Get the user's registration date D on the popular science website L and the current date D N , calculate the usage date D=D N -D L , get the user's registration date D L To current dateD N The number of views of the popular science content in the content is N, and the popular science content includes popular science articles and popular science videos. The daily views are calculated as R=N / D, and the daily views R>R sta The users are recorded as key users, among which R sta is the preset judgment value; S2: Obtain the total time t spent by key users browsing popular science content and calculate the average time spent t ave =t / N, calculate the browsing intention value of key users for popular science content , where ts represents the duration of the popular science video, A1 and A2 are the preset first and second intention coefficients, A1 < 0, and B is the preset intention threshold; The popular science contents belonging to the same section in the popular science website are recorded as the same type of content, where the section refers to the classification directory in the navigation bar of the popular science website, and the attention value of key users to the same type of content is calculated as G=λ*Y*n, where λ represents the preset adjustment coefficient and n represents the number of the same type of content; S3: Obtain the ranking of the number of views of the popular science website, obtain the total number of views S and forwarding times Sp of the popular science content in the preset time period T in the ranking of the number of views, and calculate the quality coefficient of the popular science content , where γ1 and γ2 represent the preset first and second quality coefficients respectively, 0<γ1<γ2, P represents the completion ratio of popular science content, and η represents the preset amplification factor; Calculate the recommended value of popular science content , where G s The attention value corresponding to the same type of content to which the popular science content belongs is calculated. The popular science content is sorted from high to low according to the recommendation value, and the top 10% of the popular science content is selected as the recommended content.

2. According to the method of claim 1, the method is characterized in that: In the step S1, if the number of popular science content browsed by the user N<N min Then stop the subsequent operation for this user, where N min Represents the preset minimum number of page views.

3. The method for personalized recommendation of popular science content based on big data according to claim 1, characterized in that: In step S2, the user browsing time t<t min The popular science content will be removed and will not be included in the registration date D L To current dateD N The number of views of popular science content in the article is N, where t min Represents the preset minimum browsing time.

4. The method for personalized recommendation of popular science content based on big data according to claim 1, characterized in that: In the step S2, the popular science content that the user has browsed is marked as duplicate content, and the recommendation value Rec=0.5*Re of the duplicate content is calculated.

5. The method for personalized recommendation of popular science content based on big data according to claim 1, characterized in that: In the step S3, if there are popular science contents with equal recommendation values, the popular science contents with higher attention values ​​are ranked higher.

6. The method for personalized recommendation of popular science content based on big data according to claim 1, characterized in that: In step S1, the daily page views R≤R sta The users are recorded as non-key users, and the popular science contents are sorted from high to low according to the quality coefficient, and the top 10% of the popular science contents are selected as the recommended contents for non-key users.

7. The method for personalized recommendation of popular science content based on big data according to claim 1, characterized in that: In step S2, the maximum time t for key users to browse popular science content is obtained. max , if the duration of the popular science video is ts>2*t max , it is not recommended content.

8. A personalized recommendation system for popular science content based on big data, characterized in that: include: Filter module: Get the user's registration date on the popular science website D L and the current date D N , calculate the usage date D=D N -D L , get the user's registration date D L To current dateD N The number of views of the popular science content in the content is N, and the popular science content includes popular science articles and popular science videos. The daily views are calculated as R = N / D, and the daily views R>R sta The users are recorded as key users, among which R sta is the preset judgment value; Calculation module: Obtain the total time t spent by key users browsing popular science content and calculate the average time spent t ave =t / N, calculate the browsing intention value of key users for popular science content , where ts represents the duration of the popular science video, A1 and A2 are the preset first and second intention coefficients, A1 < 0, and B is the preset intention threshold; The popular science contents belonging to the same section in the popular science website are recorded as the same type of content, where the section refers to the classification directory in the navigation bar of the popular science website, and the attention value of key users to the same type of content is calculated as G=λ*Y*n, where λ represents the preset adjustment coefficient and n represents the number of the same type of content; Recommendation module: obtain the ranking of the number of views of the popular science website, obtain the total number of views S and forwarding times Sp of the popular science content in the preset time period T in the ranking of the number of views, and calculate the quality coefficient of the popular science content , where γ1 and γ2 represent the preset first and second quality coefficients respectively, 0<γ1<γ2, P represents the completion ratio of popular science content, and η represents the preset amplification factor; Calculate the recommended value of popular science content , where Gs is the attention value corresponding to the same type of content to which the popular science content belongs. The popular science content is sorted from high to low according to the recommendation value, and the top 10% of the popular science content is selected as the recommended content.

Images