Hot event extraction method based on Internet

Abstract

The invention relates to the technical field of data analysis, and discloses an internet-based hot event extraction method, which comprises the following steps: collecting keywords during user query and query times and sorting, selecting target keywords and forming a user set. Then, selecting users in the user set, obtaining hot events browsed by the users, calculating an average browsing rate, determining the hot events and counting actual browsing times; dividing the keywords into a high-frequency group and a low-frequency group according to actual browsing times; for the low-frequency keywords, the average growth rate of the low-frequency keywords is calculated by setting a sub-time interval, a growth threshold value is set, and the low-frequency keywords meeting the condition are moved into a high-frequency set; and finally, sorting the high-frequency keyword set according to the comprehensive score to represent the popularity of different events, thereby realizing effective extraction and sorting of internet hot events.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a method for extracting hot events based on the Internet. Background Art

[0002] With the rapid development of the Internet and the increasing convenience of information dissemination, the Internet has become the main channel for people to obtain information and communicate and interact. Every day, a huge amount of information is generated and disseminated on the Internet. Among them, hot events, as the focus of social attention, have triggered extensive discussion and participation.

[0003] However, in the vast ocean of information, how to quickly and accurately extract the truly hot events has become an important challenge. Traditional methods for extracting hot events often rely on manual editing or simple statistical indicators, which are difficult to adapt to the characteristics of rapid changes in network information and diversified propagation paths, and easily lead to problems such as information overload and hot spot lag. Therefore, developing a method that can extract hot Internet events in real time, dynamically, and accurately is of great significance for timely understanding of social dynamics, grasping the direction of public opinion, and assisting decision-making.

[0004] The existing methods for extracting hot events have the following limitations. First, traditional methods often use simple keyword statistics, such as keyword search frequency and number of clicks, as indicators to measure the popularity of events, ignoring the user's interest in the event and browsing behavior, which can easily lead to some low-attention events being misjudged as hot spots. Secondly, traditional methods are difficult to capture the dynamic changes in the popularity of events. Internet hot spots are often time-sensitive. An event may break out quickly in a short period of time and then quickly fade away. Traditional static analysis methods are difficult to capture such dynamic changes. In addition, traditional methods often use manual screening and editing, which is inefficient and cannot adapt to the characteristics of rapid updates of Internet information. Furthermore, many methods do not make full use of user behavior data, and it is difficult to evaluate the user's real interest in the event based only on the number of clicks.

[0005] In order to solve the above problems, this technical solution proposes a method for extracting hot events based on the Internet. This method can more accurately extract truly hot events by analyzing the user's query behavior, browsing behavior and the dynamic changes of the event itself, which has important theoretical significance and practical application value. This method can effectively solve many problems existing in the current hot event extraction, and provide strong support for people to timely understand social dynamics, grasp the direction of public opinion, and assist in decision-making. Summary of the invention

[0006] The present invention provides a method for extracting hot events based on the Internet, which is used to promote the solution of the problems mentioned in the above background technology.

[0007] The present invention provides the following technical solution: a method for extracting hot events based on the Internet, comprising:

[0008] S1. Collect keywords used by all users during query, record the number of queries for each keyword, sort by the number of queries, select the sorted keywords as target keywords in turn, and obtain users who query the target keywords to form a user set;

[0009] S2. Select any user in the user set, obtain the popular events browsed by the user, calculate the average browsing rate, determine the popular events according to the average browsing rate of the user and the total duration of the events corresponding to the target keyword, and count the actual browsing times of the popular events corresponding to each target keyword;

[0010] S3, calculating the average browsing times of the keywords according to the actual browsing times, and entering the keywords whose actual browsing times are greater than or equal to the average browsing times into the first keyword set, and entering the keywords whose actual browsing times are less than the average browsing times into the second keyword set;

[0011] S4, setting multiple sub-time intervals, sequentially selecting keywords in the second keyword set as low-frequency keywords, and calculating an average growth rate according to the growth rate of the actual number of browsing times in each sub-time interval;

[0012] S5. Set a growth threshold, move low-frequency keywords with an average growth rate greater than or equal to the growth threshold into the first set, and sort the first keyword set according to the comprehensive score to indicate the popularity of different events.

[0013] Optionally, the collecting of keywords used by all users during querying, recording the number of queries for each keyword, and sorting by the number of queries, sequentially selecting the sorted keywords as target keywords, and obtaining users who query the target keywords to form a user set includes:

[0014] Set the time interval for collecting keywords;

[0015] Get the keywords used by all users when searching within a time interval;

[0016] Each keyword corresponds to multiple events, and all events are presented in the form of videos;

[0017] Record the number of queries for each keyword separately and enter the frequency set;

[0018] Sort the keywords in the frequency set from largest to smallest according to the number of queries;

[0019] Select keywords in the frequency set in turn as target keywords;

[0020] In the time interval, obtain the keyword of each user query, if the keyword queried by the user is the target keyword, enter the user into the user set;

[0021] Each query corresponds to a user.

[0022] Optionally, the selecting of any user in the user set, obtaining the hot events browsed by the user, calculating the average browsing rate thereof, determining the hot events according to the average browsing rate of the user and the total duration of the events corresponding to the target keyword, and counting the actual number of browsing times of the hot events corresponding to each target keyword include:

[0023] Select any user in the user set as the target user;

[0024] Obtain consent from the target user, obtain all popular events that the target user has browsed in the past, and record them in the browsing collection;

[0025] The average browsing rate of target users is calculated by the following formula:

[0026]

[0027] In the above formula, R is the average browsing rate of the target user in the browsing set, n is the total number of elements in the browsing set, and g is i is the actual time taken by the target user to watch the i-th popular event in the browsing set, t i is the total duration of browsing the i-th popular event in the collection;

[0028] Get all events corresponding to the target keyword, determine the hot events of the target keyword through the actual time when all users browse each event and the total duration of the corresponding events, and get the actual number of views of the hot events corresponding to the target keyword.

[0029] Optionally, the step of acquiring all events corresponding to the target keyword, determining the hot events of the target keyword by the actual time when all users browse each event and the total duration of the corresponding events, and acquiring the actual number of views of the hot events corresponding to the target keyword includes:

[0030] Obtain all events corresponding to the target keyword browsed by the target user and enter the event collection;

[0031] Calculate the product of the target user's average browsing rate and the total duration of each event in the event set;

[0032] Mark the browsing results that are greater than or equal to the actual time of the corresponding event as irrelevant;

[0033] Reselect any user in the user set as a new target user and calculate the product of the actual time the target user browses all events corresponding to the target keyword and the average browsing rate of the target user until all users in the user set are traversed.

[0034] Optionally, the step of acquiring all events corresponding to the target keyword, determining the hot events of the target keyword by the actual time when all users browse each event and the total duration of the corresponding events, and acquiring the actual number of views of the hot events corresponding to the target keyword includes:

[0035] Mark the data with a product smaller than the actual time of the corresponding event as related browsing, and obtain the time of the related browsing;

[0036] Reselect any user in the user set as a new target user and calculate the product of the actual time when the target user browses all events corresponding to the target keyword and the average browsing rate of the target user;

[0037] When all users in the user set are traversed, the event with the most related views in the event set is obtained as the hot event of the target keyword;

[0038] The number of relevant views of the target keyword corresponding to the hot event is recorded as the actual number of views;

[0039] Reselect any keyword in the frequency set as a new target keyword, and obtain the actual number of views of the hot event corresponding to the target keyword, until all keywords in the frequency set are traversed.

[0040] Optionally, the step of calculating the average browsing times of keywords according to the actual browsing times, and entering keywords with actual browsing times greater than or equal to the average browsing times into the first keyword set, and entering keywords with actual browsing times less than the average browsing times into the second keyword set, includes:

[0041] If all keywords in the frequency set are traversed, the average number of views of the frequency set is calculated;

[0042] Average number of views = the sum of the actual number of views of all keywords in the frequency set / the number of keywords in the frequency set;

[0043] Enter the keywords in the frequency set whose actual browsing times are greater than or equal to the average browsing times into the first keyword set;

[0044] Keywords in the frequency set whose actual browsing times are less than the average browsing times are entered into the second keyword set.

[0045] Optionally, setting multiple sub-time intervals, selecting keywords in the second keyword set as low-frequency keywords in sequence, and calculating an average growth rate according to the growth rate of the actual number of browsing times in each sub-time interval includes:

[0046] Divide the time interval into multiple sub-time intervals of equal length;

[0047] Select any keyword in the second keyword set as a low-frequency keyword;

[0048] Obtain the number of low-frequency keyword-related views in each sub-time interval in order from morning to night, and enter them into a trend set;

[0049] The average growth rate of low-frequency keywords is calculated using the following formula:

[0050]

[0051] In the above formula, Z is the average growth rate of low-frequency keywords, m is the number of elements in the trend set, and a i+1 is the i+1th element in the trend set corresponding to the low-frequency keyword, a i It is the i-th element in the trend set corresponding to the low-frequency keyword.

[0052] Optionally, the setting of a growth threshold, moving low-frequency keywords with an average growth rate greater than or equal to the growth threshold into the first set, and sorting the first keyword set according to the comprehensive score to indicate the popularity of different events, includes:

[0053] Set the growth threshold used to determine the average growth rate;

[0054] If the average growth rate of the low-frequency keyword is greater than or equal to the growth threshold, the low-frequency keyword is entered into the first keyword set, and any element in the second keyword set is reselected as a new low-frequency keyword to calculate the average growth rate of the low-frequency keyword;

[0055] If the average growth rate of the low-frequency keyword is less than the growth threshold, the low-frequency keyword is removed from the second keyword set, and any element in the second keyword set is reselected as a new low-frequency keyword to calculate the average growth rate of the low-frequency keyword;

[0056] When all keywords in the second keyword set are traversed, the formula Calculate the comprehensive score of each keyword in the first keyword set, where V in the above formula is the actual number of page views of the keyword, and Z is the average growth rate of the keyword;

[0057] The popular events corresponding to all the keywords in the first keyword set are sorted from large to small according to the comprehensive scores of the keywords.

[0058] The present invention has the following beneficial effects:

[0059] 1. A method for extracting hot events based on the Internet, by setting a time interval, is used to collect keywords queried by users within the time interval. All keywords queried by users within the time interval will be recorded. Each keyword corresponds to multiple events presented in the form of videos. Then, the system will count the number of queries for each keyword respectively, and record the keyword and the corresponding query number in a frequency set. The keywords in the frequency set will be sorted from large to small according to the number of queries. Then, the system will select the keywords in the frequency set as target keywords in turn. For each target keyword, the system will search for users who have queried the keyword within the time interval, and add these users to the user set corresponding to the target keyword. It should be noted that each query corresponds to a user; this can effectively screen out high-frequency keywords and cluster users who have queried these keywords together, providing a data basis for subsequent hot event analysis. By sorting keywords according to the number of queries, keywords with high user attention can be prioritized, thereby improving analysis efficiency. Recording the user corresponding to each keyword provides convenience for subsequent analysis of user browsing behavior and evaluation of the user's real interest in events under the keyword. Events presented in the form of videos make it easier to collect user behavior data. This method can quickly find user groups associated with specific keywords, providing strong support for subsequent precise analysis and hot event mining.

[0060] 2. A method for extracting hot events based on the Internet, by selecting a user as a target user from a user set, and after obtaining the user's consent, obtaining all the hot events that the user has browsed in history, and recording these events in a browsing set. Then, the average browsing rate of the target user is calculated by a formula, which divides the actual time the user watches each event by the total time of the event, and calculates the average. Then, according to the actual time when all users browse the events corresponding to the target keyword and the total time of the event, the hot events of the target keyword are preliminarily determined, and the actual number of views of these hot events is counted; in this way, the average browsing habits of the user can be quantified through the user's historical browsing behavior, and a benchmark can be provided for the subsequent judgment of the user's attention to a specific event. The calculation of the average browsing rate takes into account the relationship between the actual time when the user watches each event and the total time of the event, and can more accurately evaluate the user's browsing depth and interest. Through the browsing data of all users, the hot events of the target keyword are preliminarily determined, which provides a direction for the subsequent hot event screening. This method uses the user's historical behavior data, is closer to the user's real interest preferences, and provides more reliable basic data for the subsequent hot event extraction, so that the extracted hot events are more in line with the user's actual focus.

[0061] 3. A method for extracting hot events based on the Internet, for each target user, obtain all events corresponding to the target keyword that the user has browsed and put them into the event set. Then, calculate the product of the user's average browsing rate and the total duration of each event in the event set, and compare the product with the actual time when the user browses the corresponding event. If the product is greater than or equal to the actual browsing time, the browsing is considered to be irrelevant; conversely, if the product is less than the actual browsing time, the browsing is considered to be relevant, and the time when the relevant browsing occurs is recorded. This process will traverse all users in the user set and perform the same calculation for each user. After traversing all users, the system will count the number of relevant browsing times of all events in the event set, and determine the event with the most relevant browsing times as the hot event of the target keyword. Finally, the number of relevant browsing times of the hot event corresponding to the target keyword is recorded as the actual number of browsing times of the hot event, and repeat this process for all target keywords; this can effectively filter out events that users may accidentally or quickly browse, retain those events that users are really interested in and have browsed in depth, so as to more accurately evaluate the popularity of the event. By introducing the average browsing rate of users as a reference, it is possible to more accurately determine whether users have watched an event seriously, rather than just using the number of clicks as a criterion. By dividing browsing behavior into relevant browsing and irrelevant browsing, and selecting the events with the most relevant browsing as hot events, the screening accuracy of hot events can be improved. At the same time, counting the number of relevant browsing times for hot events provides a more reliable data basis for subsequent evaluation of the popularity of keywords, making subsequent hot event extraction more accurate and effective.

[0062] 4. A method for extracting hot events based on the Internet. After the system traverses all the keywords in the frequency set, it calculates the average number of views of all the keywords in the frequency set. The calculation method is to divide the sum of the actual number of views of all the keywords in the frequency set by the number of keywords in the frequency set. Then, the system divides the keywords into two categories according to the comparison results of the actual number of views of the keywords and the average number of views. Keywords with actual views greater than or equal to the average number of views are entered into the first keyword set, while keywords with actual views less than the average number of views are entered into the second keyword set. This can effectively classify the keywords according to the browsing popularity of the keywords, providing a basis for subsequent keyword analysis of different types. By calculating the average number of views, it is possible to quickly determine which keywords have a high degree of attention, so that these high-frequency keywords can be processed first, improving the analysis efficiency. Dividing keywords with views less than the average into the second set provides the possibility of subsequent screening of potential hot events based on the growth rate, avoiding the omission of some events that may come later. This classification method lays the foundation for the next step of dynamic analysis, especially the growth trend analysis of low-frequency keywords.

[0063] 5. A method for extracting hot events based on the Internet, by dividing the entire time interval into multiple sub-time intervals of the same length, and then selecting a keyword from the second keyword set as a low-frequency keyword. Then, in chronological order, the number of relevant views of the low-frequency keyword in each sub-time interval is counted in turn, and these views are recorded in the trend set. Then, the average growth rate of the low-frequency keyword is calculated using a formula, which calculates the growth rate of two adjacent elements in the trend set and calculates the average value. A growth threshold is set. If the average growth rate of a low-frequency keyword is greater than or equal to the threshold, the keyword is moved from the second set to the first set; if it is less than the threshold, the keyword is directly removed from the second set. Repeat the above steps until all keywords in the second set are traversed. Finally, a comprehensive scoring formula is used to mark the events corresponding to all keywords in the first set as hot events, and sort them according to the comprehensive score, so as to indicate the heat of events in different time periods; this can effectively mine potential hot events that may be erupting, making up for the deficiency of traditional methods that only focus on high-frequency keywords. By analyzing the growth rate of low-frequency keywords, we can timely discover those events that have a low current pageview volume but a fast growth rate, so as to more comprehensively evaluate the popularity of the event. Using sub-time intervals can more finely capture the dynamic changes of events. In addition, the average growth rate indicator can quantitatively evaluate the growth trend of events, thus avoiding errors caused by subjective judgment. Finally, using the comprehensive score as the basis for sorting can more comprehensively consider the impact of pageviews and growth rates, making the sorting of hot events more reasonable and closer to user attention. This dynamic analysis method can timely discover potential hot events, and can more accurately evaluate the popularity of events, providing users with more comprehensive and timely information services. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0065] 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.

[0066] Example 1, see Figure 1 , a method for extracting hot events based on the Internet, comprising:

[0067] S1. Collect keywords used by all users during query, record the number of queries for each keyword, sort by the number of queries, select the sorted keywords as target keywords in turn, and obtain users who query the target keywords to form a user set;

[0068] S2. Select any user in the user set, obtain the popular events browsed by the user, calculate the average browsing rate, determine the popular events according to the average browsing rate of the user and the total duration of the events corresponding to the target keyword, and count the actual browsing times of the popular events corresponding to each target keyword;

[0069] S3, calculating the average browsing times of the keywords according to the actual browsing times, and entering the keywords whose actual browsing times are greater than or equal to the average browsing times into the first keyword set, and entering the keywords whose actual browsing times are less than the average browsing times into the second keyword set;

[0070] S4, setting multiple sub-time intervals, sequentially selecting keywords in the second keyword set as low-frequency keywords, and calculating an average growth rate according to the growth rate of the actual number of browsing times in each sub-time interval;

[0071] S5. Set a growth threshold, move low-frequency keywords with an average growth rate greater than or equal to the growth threshold into the first set, and sort the first keyword set according to the comprehensive score to indicate the popularity of different events.

[0072] The method collects keywords used by all users during query, records the query times of each keyword, and sorts them by the query times, selects the sorted keywords as target keywords in turn, and obtains users who query the target keywords to form a user set, including:

[0073] Set the time interval for collecting keywords;

[0074] Get the keywords used by all users when searching within a time interval;

[0075] Each keyword corresponds to multiple events, and all events are presented in the form of videos;

[0076] Record the number of queries for each keyword separately and enter the frequency set;

[0077] Sort the keywords in the frequency set from largest to smallest according to the number of queries;

[0078] Select keywords in the frequency set as target keywords in sequence;

[0079] In the time interval, obtain the keyword of each user query, if the keyword queried by the user is the target keyword, enter the user into the user set;

[0080] Each query corresponds to a user.

[0081] By setting a time interval, the keywords that users searched in the time interval are collected. All keywords that users searched in the time interval will be recorded. Each keyword corresponds to multiple events presented in the form of videos. Then, the system will count the number of queries for each keyword and record the keyword and the corresponding query number in the frequency set. The keywords in the frequency set will be sorted from large to small according to the number of queries. Then, the system will select the keywords in the frequency set as target keywords in turn. For each target keyword, the system will search for users who have searched for the keyword in the time interval and add these users to the user set corresponding to the target keyword. It should be noted that each query corresponds to a user; this can effectively filter out high-frequency keywords and cluster users who have searched for these keywords together, providing a data basis for subsequent hot event analysis. By sorting keywords according to the number of queries, keywords with high user attention can be prioritized, which improves analysis efficiency. Recording the users corresponding to each keyword provides convenience for subsequent analysis of user browsing behavior and evaluation of users' real interest in events under the keyword. Events presented in the form of videos make it easier to collect user behavior data. This method can quickly find user groups associated with specific keywords, providing strong support for subsequent precise analysis and hot event mining.

[0082] The step of selecting any user from the user set, obtaining the hot events browsed by the user, calculating the average browsing rate, determining the hot events according to the average browsing rate of the user and the total duration of the events corresponding to the target keyword, and counting the actual browsing times of the hot events corresponding to each target keyword includes:

[0083] Select any user in the user set as the target user;

[0084] Obtain consent from the target user, obtain all popular events that the target user has browsed in the past, and record them in the browsing collection;

[0085] The average browsing rate of target users is calculated by the following formula:

[0086]

[0087] In the above formula, R is the average browsing rate of the target user in the browsing set, n is the total number of elements in the browsing set, and g is i is the actual time taken by the target user to watch the i-th popular event in the browsing set, t i is the total duration of browsing the i-th popular event in the collection;

[0088] Get all events corresponding to the target keyword, determine the hot events of the target keyword through the actual time when all users browse each event and the total duration of the corresponding events, and get the actual number of views of the hot events corresponding to the target keyword.

[0089] By selecting a user from the user set as the target user, and after obtaining the user's consent, all the hot events that the user has browsed in the past are obtained, and these events are recorded in a browsing set. Then, the average browsing rate of the target user is calculated by a formula, which divides the actual viewing time of each event by the total duration of the event and calculates the average. Then, according to the actual time of all users browsing the events corresponding to the target keyword and the total duration of the event, the hot events of the target keyword are preliminarily determined, and the actual number of views of these hot events is counted; in this way, the average browsing habits of users can be quantified through the historical browsing behavior of users, providing a benchmark for subsequent judgment of the user's attention to specific events. The calculation of the average browsing rate takes into account the relationship between the actual time of users watching each event and the total duration of the event, and can more accurately evaluate the browsing depth and interest of users. Through the browsing data of all users, the hot events of the target keyword are preliminarily determined, providing a direction for the subsequent hot event screening. This method uses the historical behavior data of users, is closer to the real interest preferences of users, and provides more reliable basic data for the subsequent hot event extraction, so that the extracted hot events are more in line with the actual focus of users.

[0090] The step of obtaining all events corresponding to the target keyword, determining the hot events of the target keyword by the actual time when all users browse each event and the total duration of the corresponding events, and obtaining the actual number of views of the hot events corresponding to the target keyword includes:

[0091] Obtain all events corresponding to the target keyword browsed by the target user and enter the event collection;

[0092] Calculate the product of the target user's average browsing rate and the total duration of each event in the event set;

[0093] Mark the browsing results that are greater than or equal to the actual time of the corresponding event as irrelevant;

[0094] Reselect any user in the user set as a new target user and calculate the product of the actual time the target user browses all events corresponding to the target keyword and the average browsing rate of the target user until all users in the user set are traversed.

[0095] The step of obtaining all events corresponding to the target keyword, determining the hot events of the target keyword by the actual time when all users browse each event and the total duration of the corresponding events, and obtaining the actual number of views of the hot events corresponding to the target keyword includes:

[0096] Mark the data with a product smaller than the actual time of the corresponding event as related browsing, and obtain the time of the related browsing;

[0097] Reselect any user in the user set as a new target user and calculate the product of the actual time when the target user browses all events corresponding to the target keyword and the average browsing rate of the target user;

[0098] When all users in the user set are traversed, the event with the most related views in the event set is obtained as the hot event of the target keyword;

[0099] The number of relevant views of the target keyword corresponding to the hot event is recorded as the actual number of views;

[0100] Reselect any keyword in the frequency set as a new target keyword, and obtain the actual number of views of the hot event corresponding to the target keyword, until all keywords in the frequency set are traversed.

[0101] For each target user, all events corresponding to the target keyword that the user has browsed are obtained and put into the event set. Then, the product of the user's average browsing rate and the total duration of each event in the event set is calculated, and the product is compared with the actual time when the user browses the corresponding event. If the product is greater than or equal to the actual browsing time, the browsing is considered to be irrelevant; conversely, if the product is less than the actual browsing time, the browsing is considered to be relevant, and the time when the relevant browsing occurs is recorded. This process will traverse all users in the user set and perform the same calculation for each user. After traversing all users, the system will count the number of relevant browsing times of all events in the event set, and determine the event with the most relevant browsing times as the hot event of the target keyword. Finally, the number of relevant browsing times of the hot event corresponding to the target keyword is recorded as the actual number of browsing times of the hot event, and this process is repeated for all target keywords; this can effectively filter out events that users may accidentally or quickly browse, retain those events that users are really interested in and have browsed in depth, and more accurately evaluate the popularity of events. By introducing the user's average browsing rate as a reference, it is possible to more accurately judge whether a user has watched an event seriously, rather than just based on the number of clicks as a criterion. By dividing browsing behaviors into relevant browsing and irrelevant browsing, and selecting the events with the most relevant browsing as hot events, the accuracy of hot event screening can be improved. At the same time, counting the number of relevant browsing times of hot events provides a more reliable data basis for subsequent evaluation of the popularity of keywords, making subsequent hot event extraction more accurate and effective.

[0102] The method of calculating the average browsing times of keywords according to the actual browsing times, and entering keywords with actual browsing times greater than or equal to the average browsing times into the first keyword set, and entering keywords with actual browsing times less than the average browsing times into the second keyword set, includes:

[0103] If all keywords in the frequency set are traversed, the average number of views of the frequency set is calculated;

[0104] Average number of views = the sum of the actual number of views of all keywords in the frequency set / the number of keywords in the frequency set;

[0105] Enter the keywords in the frequency set whose actual browsing times are greater than or equal to the average browsing times into the first keyword set;

[0106] Keywords in the frequency set whose actual browsing times are less than the average browsing times are entered into the second keyword set.

[0107] After the system has traversed all the keywords in the frequency set, it will calculate the average number of views of all the keywords in the frequency set. The calculation method is to divide the sum of the actual number of views of all the keywords in the frequency set by the number of keywords in the frequency set. Then, according to the comparison results of the actual number of views of the keywords and the average number of views, the system will divide the keywords into two categories. Keywords with actual views greater than or equal to the average number of views are entered into the first keyword set, while keywords with actual views less than the average number of views are entered into the second keyword set. This can effectively classify the keywords according to their browsing popularity, providing a basis for subsequent keyword analysis of different types. By calculating the average number of views, it is possible to quickly determine which keywords have a high degree of attention, so that these high-frequency keywords can be processed first, improving the efficiency of analysis. Dividing keywords with views less than the average into the second set makes it possible to screen potential hot events based on the growth rate in the future, avoiding the omission of some events that may come later. This classification method lays the foundation for the next step of dynamic analysis, especially the growth trend analysis of low-frequency keywords.

[0108] The step of setting a plurality of sub-time intervals, sequentially selecting keywords in the second keyword set as low-frequency keywords, and calculating an average growth rate according to the growth rate of the actual number of browsing times in each sub-time interval includes:

[0109] Divide the time interval into multiple sub-time intervals of equal length;

[0110] Select any keyword in the second keyword set as a low-frequency keyword;

[0111] Obtain the number of low-frequency keyword-related views in each sub-time interval in order from morning to night, and enter them into a trend set;

[0112] The average growth rate of low-frequency keywords is calculated using the following formula:

[0113]

[0114] In the above formula, Z is the average growth rate of low-frequency keywords, m is the number of elements in the trend set, and a i+1 is the i+1th element in the trend set corresponding to the low-frequency keyword, a i It is the i-th element in the trend set corresponding to the low-frequency keyword.

[0115] The setting of the growth threshold, moving the low-frequency keywords with an average growth rate greater than or equal to the growth threshold into the first set, and sorting the first keyword set according to the comprehensive score to indicate the popularity of different events, includes:

[0116] Set the growth threshold used to determine the average growth rate;

[0117] If the average growth rate of the low-frequency keyword is greater than or equal to the growth threshold, the low-frequency keyword is entered into the first keyword set, and any element in the second keyword set is reselected as a new low-frequency keyword to calculate the average growth rate of the low-frequency keyword;

[0118] If the average growth rate of the low-frequency keyword is less than the growth threshold, the low-frequency keyword is removed from the second keyword set, and any element in the second keyword set is reselected as a new low-frequency keyword to calculate the average growth rate of the low-frequency keyword;

[0119] When all keywords in the second keyword set are traversed, the formula Calculate the comprehensive score of each keyword in the first keyword set, where V in the above formula is the actual number of page views of the keyword, and Z is the average growth rate of the keyword;

[0120] The popular events corresponding to all the keywords in the first keyword set are sorted from large to small according to the comprehensive scores of the keywords.

[0121] By dividing the entire time interval into multiple sub-time intervals of the same length, a keyword is selected from the second keyword set as a low-frequency keyword. Then, the number of relevant views of the low-frequency keyword in each sub-time interval is counted in chronological order, and these views are recorded in the trend set. Then, the average growth rate of the low-frequency keyword is calculated using a formula, which calculates the growth rate of two adjacent elements in the trend set and calculates the average value. A growth threshold is set. If the average growth rate of a low-frequency keyword is greater than or equal to the threshold, the keyword is moved from the second set to the first set; if it is less than the threshold, the keyword is directly removed from the second set. Repeat the above steps until all keywords in the second set are traversed. Finally, a comprehensive scoring formula is used to mark the events corresponding to all keywords in the first set as hot events, and sort them according to the comprehensive score, so as to indicate the heat of events in different time periods; this can effectively mine potential hot events that may be erupting, making up for the deficiency of traditional methods that only focus on high-frequency keywords. By analyzing the growth rate of low-frequency keywords, those events that have a low current browsing volume but a fast growth rate can be discovered in time, so as to more comprehensively evaluate the heat of the event. Using sub-time intervals can more finely capture the dynamic changes of events. In addition, the average growth rate indicator can be used to quantitatively evaluate the growth trend of events, thus avoiding errors caused by subjective judgment. Finally, using the comprehensive score as the basis for sorting can more comprehensively consider the impact of the number of views and growth rate, making the sorting of hot events more reasonable and closer to user attention. This dynamic analysis method can timely discover potential hot events and more accurately evaluate the popularity of events, providing users with more comprehensive and timely information services.

[0122] 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.

[0123] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for extracting hot events based on the Internet, characterized in that: include: S1. Collect keywords used by all users during query, record the number of queries for each keyword, sort by the number of queries, select the sorted keywords as target keywords in turn, and obtain users who query the target keywords to form a user set; S2. Select any user in the user set, obtain the popular events browsed by the user, calculate the average browsing rate, determine the popular events according to the average browsing rate of the user and the total duration of the events corresponding to the target keyword, and count the actual browsing times of the popular events corresponding to each target keyword; S3, calculating the average browsing times of the keywords according to the actual browsing times, and entering the keywords whose actual browsing times are greater than or equal to the average browsing times into the first keyword set, and entering the keywords whose actual browsing times are less than the average browsing times into the second keyword set; S4, setting multiple sub-time intervals, sequentially selecting keywords in the second keyword set as low-frequency keywords, and calculating an average growth rate according to the growth rate of the actual number of browsing times in each sub-time interval; S5. Set a growth threshold, move low-frequency keywords with an average growth rate greater than or equal to the growth threshold into the first set, and sort the first keyword set according to the comprehensive score to indicate the popularity of different events.

2. The method for extracting hot events based on the Internet according to claim 1, characterized in that: The method collects keywords used by all users during query, records the query times of each keyword, and sorts them by the query times, selects the sorted keywords as target keywords in turn, and obtains users who query the target keywords to form a user set, including: Set the time interval for collecting keywords; Get the keywords used by all users when searching within a time interval; Each keyword corresponds to multiple events, and all events are presented in the form of videos; Record the number of queries for each keyword separately and enter the frequency set; Sort the keywords in the frequency set from largest to smallest according to the number of queries; Select keywords in the frequency set as target keywords in sequence; In the time interval, obtain the keyword of each user query, if the keyword queried by the user is the target keyword, enter the user into the user set; Each query corresponds to a user.

3. The method for extracting hot events based on the Internet according to claim 1, characterized in that: The step of selecting any user from the user set, obtaining the hot events browsed by the user, calculating the average browsing rate, determining the hot events according to the average browsing rate of the user and the total duration of the events corresponding to the target keyword, and counting the actual browsing times of the hot events corresponding to each target keyword includes: Select any user in the user set as the target user; Obtain consent from the target user, obtain all popular events that the target user has browsed in the past, and record them in the browsing collection; The average browsing rate of target users is calculated by the following formula: In the above formula, R is the average browsing rate of the target user in the browsing set, n is the total number of elements in the browsing set, and g is i is the actual time taken by the target user to watch the i-th popular event in the browsing set, t i is the total duration of browsing the i-th popular event in the collection; Get all events corresponding to the target keyword, determine the hot events of the target keyword through the actual time when all users browse each event and the total duration of the corresponding events, and get the actual number of views of the hot events corresponding to the target keyword.

4. The method for extracting hot events based on the Internet according to claim 3, characterized in that: The step of obtaining all events corresponding to the target keyword, determining the hot events of the target keyword by the actual time when all users browse each event and the total duration of the corresponding events, and obtaining the actual number of views of the hot events corresponding to the target keyword includes: Obtain all events corresponding to the target keyword browsed by the target user and enter the event collection; Calculate the product of the target user's average browsing rate and the total duration of each event in the event set; Mark the browsing results that are greater than or equal to the actual time of the corresponding event as irrelevant; Reselect any user in the user set as a new target user and calculate the product of the actual time the target user browses all events corresponding to the target keyword and the average browsing rate of the target user until all users in the user set are traversed.

5. The method for extracting hot events based on the Internet according to claim 3, characterized in that: The step of obtaining all events corresponding to the target keyword, determining the hot events of the target keyword by the actual time when all users browse each event and the total duration of the corresponding events, and obtaining the actual number of views of the hot events corresponding to the target keyword includes: Mark the data with a product smaller than the actual time of the corresponding event as related browsing, and obtain the time of the related browsing; Reselect any user in the user set as a new target user and calculate the product of the actual time when the target user browses all events corresponding to the target keyword and the average browsing rate of the target user; When all users in the user set are traversed, the event with the most related views in the event set is obtained as the hot event of the target keyword; The number of relevant views of the target keyword corresponding to the hot event is recorded as the actual number of views; Reselect any keyword in the frequency set as a new target keyword, and obtain the actual number of views of the hot event corresponding to the target keyword, until all keywords in the frequency set are traversed.

6. The method for extracting hot events based on the Internet according to claim 1, characterized in that: The method of calculating the average browsing times of keywords according to the actual browsing times, and entering keywords with actual browsing times greater than or equal to the average browsing times into the first keyword set, and entering keywords with actual browsing times less than the average browsing times into the second keyword set, includes: If all keywords in the frequency set are traversed, the average number of views of the frequency set is calculated; Average number of views = the sum of the actual number of views of all keywords in the frequency set / the number of keywords in the frequency set; Enter the keywords in the frequency set whose actual browsing times are greater than or equal to the average browsing times into the first keyword set; Keywords in the frequency set whose actual browsing times are less than the average browsing times are entered into the second keyword set.

7. The method for extracting hot events based on the Internet according to claim 1, characterized in that: The step of setting a plurality of sub-time intervals, sequentially selecting keywords in the second keyword set as low-frequency keywords, and calculating an average growth rate according to the growth rate of the actual number of browsing times in each sub-time interval includes: Divide the time interval into multiple sub-time intervals of equal length; Select any keyword in the second keyword set as a low-frequency keyword; Obtain the number of low-frequency keyword-related views in each sub-time interval in order from morning to night, and enter them into a trend set; The average growth rate of low-frequency keywords is calculated using the following formula: In the above formula, Z is the average growth rate of low-frequency keywords, m is the number of elements in the trend set, and a i+1 is the i+1th element in the trend set corresponding to the low-frequency keyword, a i It is the i-th element in the trend set corresponding to the low-frequency keyword.

8. The method for extracting hot events based on the Internet according to claim 1, characterized in that: The setting of the growth threshold, moving the low-frequency keywords with an average growth rate greater than or equal to the growth threshold into the first set, and sorting the first keyword set according to the comprehensive score to indicate the popularity of different events, includes: Set the growth threshold used to determine the average growth rate; If the average growth rate of the low-frequency keyword is greater than or equal to the growth threshold, the low-frequency keyword is entered into the first keyword set, and any element in the second keyword set is reselected as a new low-frequency keyword to calculate the average growth rate of the low-frequency keyword; If the average growth rate of the low-frequency keyword is less than the growth threshold, the low-frequency keyword is removed from the second keyword set, and any element in the second keyword set is reselected as a new low-frequency keyword to calculate the average growth rate of the low-frequency keyword; When all keywords in the second keyword set are traversed, the formula Calculate the comprehensive score of each keyword in the first keyword set, where V in the above formula is the actual number of page views of the keyword, and Z is the average growth rate of the keyword; The popular events corresponding to all the keywords in the first keyword set are sorted from large to small according to the comprehensive scores of the keywords.

Images