Healthy science popularization text popularity determination method and system for Internet

Abstract

The invention discloses a health science popularization text popularity determination method and system for the Internet, and belongs to the technical field of text popularity processing, and the method specifically comprises the steps: obtaining a plurality of texts of which the release time is in the same time period, marking the texts as reference texts, obtaining the text information of any reference text, obtaining a to-be-determined text, calculating text similarity x and field similarity z between the current to-be-determined text and each reference text, taking the text similarity x as a horizontal coordinate and the field similarity z as a vertical coordinate, constructing a similar coordinate system, generating feature points corresponding to all the reference texts, clustering the reference texts according to the feature points, and obtaining a clustering result; generating a plurality of category clusters; respectively analyzing all the category clusters to obtain a target category cluster, acquiring a reference text corresponding to the target category cluster and calibrating the reference text as a target text, and calculating the popularity value of the undetermined text according to the popularity value of the target text, so that the accuracy of text popularity estimation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of text heat processing, and in particular to a method and system for determining the heat of health popular science texts for the Internet. Background Art

[0002] Internet health science popularization texts refer to health knowledge popularization content for the general public published through Internet platforms (such as websites, social media, blogs, e-books, video platforms, etc.). These texts are usually presented in a concise, easy-to-understand, and highly interactive form, aiming to improve people's health literacy, disseminate scientific health information, and help users develop healthy living habits and improve health management. Due to the widespread dissemination of the Internet, health science popularization content can reach a wider audience, regardless of geographical location, age, or cultural background. Access to relevant health information through the Internet. Internet platforms provide interactive functions such as comments, sharing, and likes, allowing readers to provide timely feedback, ask questions, and interact with experts or other readers. At the same time, the Internet can quickly update health information to ensure the timeliness of the content.

[0003] At present, when pushing hot articles to users, the popularity of each article is determined based on the user behavior data corresponding to each article, and then the most popular article is pushed to the user as the hot article. The user behavior data here includes: the number of reads, the number of likes, and the number of reposts, etc. However, there are often some articles, such as articles with high timeliness, that have not generated user behavior. In the absence of user behavior data, how to determine the popularity of an article is an unresolved problem. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for determining the popularity of health popular science texts on the Internet, so as to solve the following technical problems:

[0005] At present, when pushing hot articles to users, the popularity of each article is determined based on the user behavior data corresponding to each article, and then the most popular article is pushed to the user as the hot article. The user behavior data here includes: the number of reads, the number of likes, and the number of reposts, etc. However, there are often some articles, such as articles with high timeliness, that have not generated user behavior. In the absence of user behavior data, how to determine the popularity of an article is an unresolved problem.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A method for determining the popularity of health popular science texts on the Internet, comprising the following steps:

[0008] S1, obtaining multiple texts published in the same time period and marking them as reference texts, and obtaining text information of any reference text, wherein the text information includes a popularity value and field information;

[0009] S2, obtaining the pending text and calculating the text similarity x and domain similarity z between the current pending text and each reference text respectively, constructing a similarity coordinate system with the text similarity x as the horizontal coordinate and the domain similarity z as the vertical coordinate, and generating feature points corresponding to all reference texts;

[0010] Select any feature point as the center, set the control radius R, calculate the feature point density P within the control radius R, obtain the mean feature point density within the control radius R of all feature points, and mark the mean feature point density as AcgP;

[0011] If there is any feature point whose density P is greater than AcgP within the radius R of the feature point, the feature point is marked as a core point, and a category cluster is generated with the core point as the center; if there is a non-core point whose density P within the control radius R of the feature set is also greater than AcgP, the category cluster generated by the non-core point is merged with the original category cluster to generate several category clusters;

[0012] S3, analyze all category clusters respectively to obtain the target category cluster, obtain the reference text corresponding to the target category cluster and mark it as the target text, and calculate the heat value of the pending text according to the heat value of the target text.

[0013] As a further solution of the present invention: in S2, the specific calculation process of the similarity is:

[0014] S11, respectively converting the pending text and any reference text into two groups of ASCII code values ​​in character order; using the ASCII code values ​​of surrounding characters to represent the features of each character; using an autoencoder to perform dimensionality reduction processing on the features of each character point; using the Jaccard similarity coefficient to calculate the similarity between the two texts for the features after dimensionality reduction;

[0015] The formula for calculating text similarity is:

[0016]

[0017] Among them, A is the feature set of the text to be determined, Bi is the feature set of the i-th reference text, and x is the similarity;

[0018] S12, obtain all fields corresponding to the pending text and any reference text, and then obtain all overlapping fields and mark them as Y1, Y2, ..., YN, and calculate the field similarity z according to the calculation formula z = N / (MN), where M is the number of all fields corresponding to the pending text and any reference text.

[0019] As a further solution of the present invention: in S2, the process of setting the cluster control radius R is:

[0020] Taking any feature point as the center, calculate the Euclidean distance between the feature point and any other feature point, sum each Euclidean distance to get u, and get the control radius R based on the value u. The calculation formula is as follows:

[0021]

[0022] Among them, u is the sum of the Euclidean distance data values ​​of all feature points, and I is the Euclidean distance between any feature points.

[0023] As a further solution of the present invention: in S2, the calculation formula for calculating the density of feature points within the control radius R is:

[0024] P=i / (πR 2 );

[0025] Where i is the number of feature points within the control radius R.

[0026] As a further solution of the present invention: In S3, the specific process of determining the target category cluster is:

[0027] Select any category cluster, and generate several subclusters with R as the radius and any feature point in the category cluster as the center. Calculate the degree of aggregation of each subcluster and mark it as DP; select the subcluster with the highest DP value as the representative cluster, and obtain the representative clusters corresponding to all category clusters. Calculate the mean of each axis coordinate of the feature points corresponding to all representative clusters in turn and calculate the preferred score of any representative cluster according to the calculation formula W=b*z'+d*x', where b and d are preset coefficients, z' is the mean of the horizontal axis coordinate, and x' is the mean of the vertical axis coordinate. Select the category cluster corresponding to the maximum preferred score as the target category cluster.

[0028] As a further solution of the present invention: in S3, the specific calculation process of the heat value is:

[0029]

[0030] Among them, m is the total number of target texts in the target category cluster, and refi is the heat value of the i-th target text.

[0031] A health science popularization text popularity determination system for the Internet, comprising:

[0032] A data acquisition module, used to acquire multiple texts published in the same time period and mark them as reference texts, and acquire text information of any reference text, wherein the text information includes a popularity value and field information;

[0033] A data analysis module is used to obtain a pending text and calculate the text similarity x and field similarity z between the current pending text and each reference text, construct a similarity coordinate system with the text similarity x as the horizontal coordinate and the field similarity z as the vertical coordinate, and generate feature points corresponding to all reference texts;

[0034] Select any feature point as the center, set the control radius R, calculate the feature point density P within the control radius R, obtain the feature point density mean within the control radius R of all feature points, and mark the feature point density mean as AcgP; if there is any feature point whose density P within the radius R is greater than AcgP, then mark the feature point as a core point, and generate a category cluster with the core point as the center; if there is a non-core point feature set whose density P within the control radius R is also greater than AcgP, then merge the category cluster generated by the non-core point with the original category cluster to generate several category clusters;

[0035] The result generation module is used to analyze all category clusters to obtain the target category cluster, obtain the reference text corresponding to the target category cluster and mark it as the target text, and calculate the heat value of the pending text according to the heat value of the target text.

[0036] Beneficial effects of the present invention:

[0037] The present invention first obtains the text with user interaction and published in the same time period as the reference text. It is understandable that the popularity and attention of the text are usually strongly affected by the publishing time, and the popularity of events or topics in different time periods may be very different. For example, a health and safety event may quickly gain a lot of attention in a specific time period, but may be ignored in other time periods. Therefore, in order to ensure the comparability of popularity in subsequent analysis, only texts published in the same time period are likely to be in a similar time background and comparable. It is understandable that the comparison between texts depends largely on the selection of keywords. Since keywords may have multiple meanings, similar texts will be classified into different fields. Therefore, only calculating the heat based on the similarity between the pending text and the target text may lead to inaccurate heat estimation. Therefore, the present invention selects the reference text that is most similar to the pending text by considering text similarity and field similarity, and estimates the heat of the pending text based on the reference text, which can ensure that the pending text and the reference text are compared in the same or similar context, avoiding mistaking texts in unrelated fields for similar texts, thereby improving the accuracy of heat estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention will be further described below in conjunction with the accompanying drawings.

[0039] Figure 1The present invention is a flow chart of a method for determining the popularity of health popular science texts on the Internet. DETAILED DESCRIPTION

[0040] 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.

[0041] See also Figure 1 As shown, the present invention is a method for determining the popularity of health popular science texts for the Internet, comprising the following steps:

[0042] S1, obtaining multiple texts published in the same time period and marking them as reference texts, and obtaining text information of any reference text, wherein the text information includes a popularity value and field information;

[0043] S2, obtaining the pending text and calculating the text similarity x and domain similarity z between the current pending text and each reference text respectively, constructing a similarity coordinate system with the text similarity x as the horizontal coordinate and the domain similarity z as the vertical coordinate, and generating feature points corresponding to all reference texts;

[0044] Select any feature point as the center, set the control radius R, calculate the feature point density P within the control radius R, obtain the mean feature point density within the control radius R of all feature points, and mark the mean feature point density as AcgP;

[0045] If there is any feature point whose density P is greater than AcgP within the radius R of the feature point, the feature point is marked as a core point, and a category cluster is generated with the core point as the center; if there is a non-core point whose density P within the control radius R of the feature set is also greater than AcgP, the category cluster generated by the non-core point is merged with the original category cluster to generate several category clusters;

[0046] S3, analyze all category clusters respectively to obtain the target category cluster, obtain the reference text corresponding to the target category cluster and mark it as the target text, and calculate the heat value of the pending text according to the heat value of the target text.

[0047] The present invention first obtains the text with user interaction and published in the same time period as the reference text. It is understandable that the popularity and attention of the text are usually strongly affected by the publishing time, and the popularity of events or topics in different time periods may be very different. For example, a health and safety event may quickly gain a lot of attention in a specific time period, but may be ignored in other time periods. Therefore, in order to ensure the comparability of popularity in subsequent analysis, only texts published in the same time period are likely to be in a similar time background and comparable. It is understandable that the comparison between texts depends largely on the selection of keywords. Since keywords may have multiple meanings, similar texts will be classified into different fields. Therefore, only calculating the heat based on the similarity between the pending text and the target text may lead to inaccurate heat estimation. Therefore, the present invention selects the reference text that is most similar to the pending text by considering text similarity and field similarity, and estimates the heat of the pending text based on the reference text, which can ensure that the pending text and the reference text are compared in the same or similar context. It avoids mistaking texts in unrelated fields for similar texts, thereby improving the accuracy of heat estimation.

[0048] In a preferred embodiment of the present invention, in S2, the specific calculation process of the similarity is:

[0049] S11, respectively converting the pending text and any reference text into two groups of ASCII code values ​​in character order; using the ASCII code values ​​of surrounding characters to represent the features of each character; using an autoencoder to perform dimensionality reduction processing on the features of each character point; using the Jaccard similarity coefficient to calculate the similarity between the two texts for the features after dimensionality reduction;

[0050] The formula for calculating text similarity is:

[0051]

[0052] Among them, A is the feature set of the text to be determined, Bi is the feature set of the i-th reference text, and x is the similarity;

[0053] S12, obtain all fields corresponding to the pending text and any reference text, and then obtain all overlapping fields and mark them as Y1, Y2, ..., YN, and calculate the field similarity z according to the calculation formula z = N / (MN), where M is the number of all fields corresponding to the pending text and any reference text.

[0054] In another preferred embodiment of the present invention, in S2, the process of setting the cluster control radius R is:

[0055] Taking any feature point as the center, calculate the Euclidean distance between the feature point and any other feature point, sum each Euclidean distance to get u, and get the control radius R based on the value u. The calculation formula is as follows:

[0056]

[0057] Among them, u is the sum of the Euclidean distance data values ​​of all feature points, and I is the Euclidean distance between any feature points.

[0058] In another preferred embodiment of the present invention, in S2, the calculation formula for calculating the density of feature points within the control radius R is:

[0059] P=i / (πR 2 );

[0060] Where i is the number of feature points within the control radius R.

[0061] In another preferred embodiment of the present invention, in S3, the specific process of determining the target category cluster is:

[0062] Select any category cluster, and generate several subclusters with R as the radius and any feature point in the category cluster as the center. Calculate the degree of aggregation of each subcluster and mark it as DP; select the subcluster with the highest DP value as the representative cluster, and obtain the representative clusters corresponding to all category clusters. Calculate the mean of each axis coordinate of the feature points corresponding to all representative clusters in turn and calculate the preferred score of any representative cluster according to the calculation formula W=b*z'+d*x', where b and d are preset coefficients, z' is the mean of the horizontal axis coordinate, and x' is the mean of the vertical axis coordinate. Select the category cluster corresponding to the maximum preferred score as the target category cluster.

[0063] In another preferred embodiment of the present invention, in S3, the specific calculation process of the heat value is:

[0064]

[0065] Among them, m is the total number of target texts in the target category cluster, and refi is the heat value of the i-th target text.

[0066] A health science popularization text popularity determination system for the Internet, comprising:

[0067] A data acquisition module, used to acquire multiple texts published in the same time period and mark them as reference texts, and acquire text information of any reference text, wherein the text information includes a popularity value and field information;

[0068] A data analysis module is used to obtain a pending text and calculate the text similarity x and field similarity z between the current pending text and each reference text, construct a similarity coordinate system with the text similarity x as the horizontal coordinate and the field similarity z as the vertical coordinate, and generate feature points corresponding to all reference texts;

[0069] Select any feature point as the center, set the control radius R, calculate the feature point density P within the control radius R, obtain the feature point density mean within the control radius R of all feature points, and mark the feature point density mean as AcgP; if there is any feature point whose density P within the radius R is greater than AcgP, then mark the feature point as a core point, and generate a category cluster with the core point as the center; if there is a non-core point feature set whose density P within the control radius R is also greater than AcgP, then merge the category cluster generated by the non-core point with the original category cluster to generate several category clusters;

[0070] The result generation module is used to analyze all category clusters to obtain the target category cluster, obtain the reference text corresponding to the target category cluster and mark it as the target text, and calculate the heat value of the pending text according to the heat value of the target text.

[0071] 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 method for determining the popularity of health popular science texts on the Internet, characterized in that: The following steps are involved: S1, obtaining multiple texts published in the same time period and marking them as reference texts, and obtaining text information of any reference text, wherein the text information includes a popularity value and field information; S2, obtaining the pending text and calculating the text similarity x and domain similarity z between the current pending text and each reference text respectively, constructing a similarity coordinate system with the text similarity x as the horizontal coordinate and the domain similarity z as the vertical coordinate, and generating feature points corresponding to all reference texts; Select any feature point as the center, set the control radius R, calculate the feature point density P within the control radius R, obtain the mean feature point density within the control radius R of all feature points, and mark the mean feature point density as AcgP; If there is any feature point whose density P is greater than AcgP within the radius R of the feature point, the feature point is marked as a core point, and a category cluster is generated with the core point as the center; if there is a non-core point whose density P within the control radius R of the feature set is also greater than AcgP, the category cluster generated by the non-core point is merged with the original category cluster to generate several category clusters; S3, analyze all category clusters respectively to obtain the target category cluster, obtain the reference text corresponding to the target category cluster and mark it as the target text, and calculate the heat value of the pending text according to the heat value of the target text.

2. A method for determining the popularity of health popular science texts for the Internet according to claim 1, characterized in that: In S2, the specific calculation process of similarity is: S11, converting the pending text and any reference text into two groups of ASCII code values ​​in character order; for each character, the ASCII code values ​​of surrounding characters are used to represent its features; The autoencoder is used to reduce the dimension of each character point; the Jaccard similarity coefficient is used to calculate the similarity between two texts after the dimension reduction. The formula for calculating text similarity is: Among them, A is the feature set of the text to be determined, Bi is the feature set of the i-th reference text, and x is the similarity; S12, obtain all fields corresponding to the pending text and any reference text, and then obtain all overlapping fields and mark them as Y1, Y2, ..., YN, and calculate the field similarity z according to the calculation formula z = N / (MN), where M is the number of all fields corresponding to the pending text and any reference text.

3. A method for determining the popularity of health popular science texts for the Internet according to claim 1, characterized in that: In S2, the process of setting the cluster control radius R is: Taking any feature point as the center, calculate the Euclidean distance between the feature point and any other feature point, sum each Euclidean distance to get u, and get the control radius R based on the value u. The calculation formula is as follows: Among them, u is the sum of the Euclidean distance data values ​​of all feature points, and I is the Euclidean distance between any feature points.

4. A method for determining the popularity of health popular science texts for the Internet according to claim 1, characterized in that: In S2, the calculation formula for calculating the density of feature points within the control radius R is: P=i / (πR 2 ); Where i is the number of feature points within the control radius R.

5. The method for determining the popularity of health popular science texts for the Internet according to claim 1, characterized in that: In S3, the specific process of determining the target category cluster is as follows: Select any category cluster, and generate several subclusters with any feature point in the category cluster as the center and R as the radius, calculate the degree of aggregation of each subcluster and mark it as DP; The subcluster with the highest DP value is selected as the representative cluster to obtain the representative cluster corresponding to all category clusters. The mean of each axis coordinate of the feature points corresponding to all representative clusters is calculated in turn, and the preferred score of any representative cluster is calculated according to the calculation formula W=b*z'+d*x', where b and d are preset coefficients, z' is the mean of the horizontal axis coordinate, and x' is the mean of the vertical axis coordinate. The category cluster corresponding to the maximum preferred score is selected as the target category cluster.

6. A method for determining the popularity of health popular science texts for the Internet according to claim 5, characterized in that: In S3, the specific calculation process of the heat value is: Among them, m is the total number of target texts in the target category cluster, and refi is the heat value of the i-th target text.

7. A system for determining the popularity of health popular science texts on the Internet, characterized in that: include: A data acquisition module, used to acquire multiple texts published in the same time period and mark them as reference texts, and acquire text information of any reference text, wherein the text information includes a popularity value and field information; A data analysis module is used to obtain a pending text and calculate the text similarity x and field similarity z between the current pending text and each reference text, construct a similarity coordinate system with the text similarity x as the horizontal coordinate and the field similarity z as the vertical coordinate, and generate feature points corresponding to all reference texts; Select any feature point as the center, set the control radius R, calculate the feature point density P within the control radius R, obtain the feature point density mean within the control radius R of all feature points, and mark the feature point density mean as AcgP; if there is any feature point whose density P within the radius R is greater than AcgP, then mark the feature point as a core point, and generate a category cluster with the core point as the center; if there is a non-core point feature set whose density P within the control radius R is also greater than AcgP, then merge the category cluster generated by the non-core point with the original category cluster to generate several category clusters; The result generation module is used to analyze all category clusters to obtain the target category cluster, obtain the reference text corresponding to the target category cluster and mark it as the target text, and calculate the heat value of the pending text according to the heat value of the target text.

Images