Live broadcast platform potential user identification method and system
By analyzing user behavior data and feature information, combined with Sigmoid function scoring and keyword extraction, potential users are identified and personalized live streaming content is pushed, solving the problems of user identification and retention in live streaming platforms and improving user experience and retention.
Patent Information
- Application Number
- CN202411433049.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing live streaming platforms struggle to accurately identify potential users and cannot push suitable live streaming content to different users, resulting in low user retention.
By acquiring user behavior data and characteristic information on the platform, using the Sigmoid function for scoring, and combining user interaction data on relevant platforms and online survey results, keywords are extracted and live streaming content is adjusted to identify potential users and personalize push notifications.
It improved the accuracy of identifying potential users and user retention, ensuring that the pushed live content better meets user preferences and expectations, thus enhancing the user experience.
Smart Images

Figure CN118945428A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of live broadcast information technology, and in particular relates to a method and system for identifying potential users of a live broadcast platform. Background Art
[0002] At present, in commercial online service platforms, it is crucial for manufacturers to accurately obtain potential user groups. With the rise of live broadcasting, many large-scale real-person interactive video live broadcast communities have emerged, referred to as live broadcast platforms. With the continuous development of live broadcast platform business, live broadcast platforms are generally part of the main site, and the main site is manifested in the following forms: video playback platform, music playback platform, portal website, etc.
[0003] Taking the video playback platform as an example, the live broadcast platform can be a live broadcast section in the video playback platform. There are a large number of users in the main site of the live broadcast platform, among which many users have the possibility of watching the live broadcast platform, and they are likely to become potential users of the live broadcast platform.
[0004] The existing method of mining potential users of live broadcast platforms uses simple screening rules to screen potential users of live broadcast platforms from the massive users of the live broadcast platform main site. However, it is unable to accurately extract potential users of live broadcast platforms from the massive users, and is unable to increase the possibility of potential users by pushing different types of live broadcasts to different users. Summary of the invention
[0005] The purpose of the embodiment of the present invention is to provide a method for identifying potential users of a live broadcast platform, aiming to solve the problem raised in the third part of the background technology.
[0006] The embodiment of the present invention is implemented as follows: a method for identifying potential users of a live broadcast platform, the method comprising: Obtaining user behavior data on the platform, including viewing time, comments and sharing data, and obtaining user feature information; Scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value; Obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcasts based on the keywords; The online survey result is obtained, and the online survey is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, and the push live broadcast is adjusted according to the online survey. The suitability of the push live broadcast is judged according to the adjusted live broadcast score.
[0007] Preferably, the step of scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value specifically includes: Determine the user's activity level based on the behavior data and feature information, where the activity level is calculated by a scoring function to calculate the probability of a potential user; Obtaining users' viewing habits of different types of content, including the channels, hosts or specific types of live broadcasts they watch, and supplementing activity levels based on preferences; Scoring is performed based on activity to obtain a first score, and the first score is compared with a first preset value. If the first score is higher than the first preset value, the user is determined to be a potential user.
[0008] Preferably, the step of obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcast according to the keywords specifically includes: Sending a permission request to the user terminal, the permission request is used to obtain platform data related to the live broadcast platform, receiving the permission to obtain the user's live broadcast related platform data; Count the frequency of word occurrences, set a keyword threshold, compare the frequency of word occurrences with the keyword threshold, and if the frequency of word occurrences is greater than the keyword threshold, determine it as a keyword and extract the keyword through relevant platform data; Obtain tags corresponding to live content, match tags with keywords, and push corresponding live content through keywords, where the tags are a summary of the live content.
[0009] Preferably, the step of obtaining online survey results, wherein the online survey is sent to a user terminal, wherein the online survey is used to obtain the user's preferences, viewing habits and expectations, adjusting the push live broadcast according to the online survey, and judging the suitability of the push live broadcast according to the adjusted live broadcast score, specifically includes: Send survey data to user terminals, obtain online survey results, and determine live broadcast satisfaction based on the survey results; Adjust push live broadcast according to live broadcast satisfaction, and obtain a second score according to the adjusted live broadcast score; Compare the first score with the second score to determine the suitability of pushing the live broadcast.
[0010] Preferably, the characteristic information includes age group, gender and geographical location.
[0011] Another object of an embodiment of the present invention is to provide a system for identifying potential users of a live broadcast platform, the system comprising: A data acquisition module is used to acquire user behavior data on the platform, including viewing time, comments and sharing data, and to acquire user feature information; A first scoring module scores the behavior data to obtain a first score, compares the first score with a first preset value, and determines the user as a potential user if the first score is higher than the first preset value; The threshold module obtains the user's live broadcast related platform data, extracts keywords through the related platform data, and the keywords are words with a probability exceeding the threshold in the related platform data, and pushes the live broadcast according to the keywords; The second scoring module obtains the results of an online survey, which is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, adjust the push live broadcast according to the online survey, and judge the suitability of the push live broadcast according to the adjusted live broadcast score.
[0012] Preferably, the first scoring module includes: An activity unit, which determines the activity of a user based on the behavior data and feature information, wherein the activity is calculated by a scoring function to calculate the probability of a potential user; A supplementing unit, which obtains the user's viewing habits of different types of content, wherein the viewing habits include the channels, anchors or specific types of live broadcasts watched, and supplements the activity level according to the preferences; The activity is scored through the first scoring unit to obtain a first score, and the first score is compared with a first preset value. If the first score is higher than the first preset value, the user is determined to be a potential user.
[0013] Preferably, the threshold module includes: The permission unit sends a permission request to the user terminal, wherein the permission request is used to obtain platform data related to the live broadcast platform, receives the permission, and obtains the user's live broadcast related platform data; Threshold unit, counts the frequency of word occurrence, sets the keyword threshold, compares the word frequency with the keyword threshold, and if the word frequency is greater than the keyword threshold, it is determined to be a keyword, and the keyword is extracted through relevant platform data; The matching unit obtains tags corresponding to the live content, matches the tags with keywords, and pushes corresponding live content through the keywords, wherein the tags are a summary of the live content.
[0014] Preferably, the second scoring module includes: A survey unit sends survey data to a user terminal, obtains online survey results, and determines live broadcast satisfaction based on the survey results; The second scoring unit adjusts the push of live broadcast according to the live broadcast satisfaction, and scores the adjusted live broadcast to obtain a second score; The comparison unit compares the first score with the second score to determine the suitability of pushing the live broadcast.
[0015] Preferably, the characteristic information includes age group, gender and geographical location.
[0016] A method for identifying potential users of a live broadcast platform provided by an embodiment of the present invention obtains behavioral data of users on the platform, collects behavioral data of users, including viewing time, viewing frequency, likes and comments, and can predict the activity and participation of each user on the platform based on the user's viewing time, viewing frequency, likes and comments, and preference for viewing content, obtain user characteristic information, know the user's age group, gender and geographical location, score according to the behavioral data, obtain a first score through a Sigmoid function, compare the first score with a first preset value, and when the first score is higher than the first preset value, the user is determined to be a potential user, and when the first score is lower than the first preset value, the user is determined to be a potential user. They are judged as non-potential users, and data on user live broadcast-related platforms are obtained. User comments and barrage data in the live broadcast of the game are collected in real time, keywords are extracted, and word frequency statistics are performed to find out the keywords that users discuss most frequently in the live broadcast, such as team names, scores, and tactics, etc. Live broadcasts are pushed based on keywords to improve the retention of potential users, survey data is sent to user terminals, online survey results are obtained, and live broadcast satisfaction is determined based on the survey results. This solves the problems of not being able to accurately extract potential users of the live broadcast platform from a large number of users, and not being able to push different types of live broadcasts to different users to increase the possibility of potential users. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A flowchart of a method for identifying potential users of a live broadcast platform provided by an embodiment of the present invention; Figure 2 A flowchart of the steps of scoring according to the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value, provided in an embodiment of the present invention; Figure 3 A flowchart of the steps of obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcasts according to the keywords, provided in an embodiment of the present invention; Figure 4 A flowchart of the steps of obtaining online survey results provided by an embodiment of the present invention, wherein the online survey is sent to a user terminal, the online survey is used to obtain the user's preferences, viewing habits and expectations, adjusting the push live broadcast according to the online survey, scoring the adjusted live broadcast, and judging the suitability of the push live broadcast; Figure 5 An architectural diagram of a system for identifying potential users of a live broadcast platform provided by an embodiment of the present invention; Figure 6 An architectural diagram of a first scoring module provided in an embodiment of the present invention; Figure 7 An architectural diagram of a threshold module provided in an embodiment of the present invention; Figure 8 This is an architecture diagram of the second scoring module provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0018] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0019] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.
[0020] like Figure 1 As shown, a method for identifying potential users of a live broadcast platform provided by an embodiment of the present invention includes: S100, obtaining user behavior data on the platform, wherein the behavior data includes viewing time, comments and sharing data, and obtaining user feature information.
[0021] In this step, the user's behavior data on the platform is obtained and collected, including viewing time, viewing frequency, likes and comments. The activity and participation of each user on the platform can be predicted based on the user's viewing time, viewing frequency, likes and comments, and viewing content preferences, and the user's characteristic information is obtained to know the user's age group, gender and geographical location; Behavioral data and feature information can be used to understand user preferences and assess the user's activity and participation in the platform. For example, if user A has a longer average viewing time and a higher frequency of likes and shares, the system can assign a higher behavioral score to him or her, indicating that he or she may be a potential user who actively participates in the platform. By analyzing basic information such as age, gender, and geographic location, combined with their activity and influence on social media, if user B is an influential opinion leader or has characteristics consistent with the platform's target users (such as being young and active in the Internet community), the system will give a higher score and identify them as having greater potential.
[0022] S200, scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value.
[0023] In this step, the behavioral data is scored, and the first score is obtained by the Sigmoid function. The first score is compared with the first preset value. When the first score is higher than the first preset value, it is determined to be a potential user. When the first score is lower than the first preset value, it is determined to be a non-potential user, because it can be achieved by analyzing the user's comment content, likes and sharing behaviors. If the user frequently participates in a specific type of live broadcast or shows a positive emotional attitude, it may indicate that he has a high attraction and participation in the platform content, and is more likely to become a potential user; Sigmoid functions are often used to map input values to a range between 0 and 1, which is suitable for scoring probabilities or potentials. For example: Score= Among them, z is the weighted sum of the input or the model output, and the output of the Sigmoid function can be interpreted as the probability or potential score of the user becoming a potential user.
[0024] S300, obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcast according to the keywords.
[0025] In this step, we obtain the platform data related to user live broadcast. The live broadcast platform usually provides an API interface for developers to obtain user data, including comments, bullet screens, likes, etc. To obtain this data, you need to apply for access rights, comply with the platform's usage agreement and data use policy, mine and analyze the collected user text data, extract keywords through relevant platform data, and calculate the probability of occurrence of words. The probability of occurrence refers to the frequency of a certain word appearing in the entire data set. The calculation formula is: Occurrence probability = number of word occurrences / total number of words / probability of occurrence / total number of words / number of word occurrences When the probability of occurrence exceeds the threshold, it is determined to be a keyword, and the live broadcast is pushed based on the keyword to achieve the effect of improving user retention.
[0026] S400, obtaining online survey results, wherein the online survey is sent to a user terminal, wherein the online survey is used to obtain the user's preferences, viewing habits and expectations, and the push live broadcast is adjusted according to the online survey, and the suitability of the push live broadcast is determined according to the adjusted live broadcast score.
[0027] In this step, the online survey results are obtained. In order to collect user satisfaction with the live broadcast pushed by the platform, a detailed questionnaire can be designed. The questionnaire should include questions in multiple dimensions to comprehensively evaluate user satisfaction, obtain user preferences, viewing habits and expectations, and obtain user satisfaction with the current live broadcast push. Adjustments are made based on user satisfaction, and the live broadcast is pushed again. Based on the adjusted live broadcast score, the suitability of the live broadcast push is judged, and the Sigmoid function is used again to score the probability or potential.
[0028] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of scoring according to the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value, specifically includes: S201, judging the user's activity level based on the behavior data and feature information, wherein the activity level is calculated by a scoring function to calculate the probability of a potential user.
[0029] In this step, the user's activity is judged based on behavioral data and feature information, and the average or cumulative viewing time of each live broadcast is evaluated. Long-term viewing usually means that the user has a high interest and recognition in the content, and the user's interactive behaviors such as likes, comments, and sharing. Frequent interaction indicates that the user is willing to participate in community communication and has a high level of investment and interaction in the platform content. The user's viewing frequency and continuous usage time are analyzed. Users who frequently return to the platform and continue to consume content are more likely to become potential users; When users have a high level of interest and recognition in content, a high level of investment and interaction in platform content, and a high viewing frequency and continuous usage time, the user's activity is determined to be high. The probability of a potential user can be calculated through activity. The Sigmoid function is used to map the input value to a range between 0 and 1. Score→1 indicates a high probability of becoming a potential user, and Score→0 indicates a low probability of becoming a potential user.
[0030] S202, obtaining the user's viewing habits for different types of content, wherein the viewing habits include the channels, hosts, or specific types of live broadcasts watched, and supplementing the activity level according to the preferences.
[0031] In this step, we obtain users’ viewing habits of different types of content, evaluate the frequency of users’ likes and sharing of live content, which reflects users’ recognition of content and willingness to promote it. We also analyze users’ viewing history, including the channels, hosts or specific types of live broadcasts they watch, to understand users’ preferred content types and viewing habits, which helps predict their future engagement. At the same time, users of different age groups and genders have different acceptance and participation methods for content. Geographical location affects users' attention to and willingness to participate in live events in a specific area. By analyzing users' historical behavior patterns and trends and predicting their future activity and willingness to participate on the platform, the accuracy and real-time performance of the scoring function can be further optimized.
[0032] S203, scoring based on activity to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value.
[0033] In this step, it is a common and effective method to score live broadcast platform users by activity and use Sigmoid function to judge whether they are potential users. Sigmoid function can map the input value to a range between 0 and 1, which can be understood as the probability or score of the user becoming a potential user. Score=
[0034] When z→∞, e -z →0, then Score→1, indicating a high probability of becoming a potential user; When z→-∞, e -z →∞, then Score→0, indicating a low probability of becoming a potential user; When z=0, Score=0.5, indicating a medium probability of becoming a potential user.
[0035] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of obtaining user live broadcast related platform data, extracting keywords through the related platform data, the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcast according to the keywords specifically includes: S301, sending a permission request to a user terminal, wherein the permission request is used to obtain platform data related to a live broadcast platform, receiving a permission to obtain user live broadcast related platform data.
[0036] In this step, a permission request is sent to the user terminal, and the API interface provided by the live broadcast platform allows developers to obtain user comments, bullet screens, likes and other data. By using the API, user interaction data can be obtained in real time and further analyzed; Send a permission acquisition request to the user terminal, receive the permission acquisition permission, obtain the user's live broadcast related platform data, and ensure that when obtaining and processing user data, relevant laws, regulations and platform regulations are complied with, especially the legal requirements regarding user privacy protection. For data involving personal identity, appropriate data anonymization measures are taken, and the security of data during transmission and storage is ensured to prevent data leakage and abuse.
[0037] S302, count the frequency of word occurrences, set a keyword threshold, compare the frequency of word occurrences with the keyword threshold. If the frequency of word occurrences is greater than the keyword threshold, it is determined as a keyword, and keywords are extracted through relevant platform data.
[0038] In this step, count the frequency of word occurrences, obtain the user's interaction data on the live broadcast platform, including comments, bullet screens, likes, shares and follows, remove stop words such as "de", "le" and "shi", etc., which have no practical significance for keyword extraction, count the frequency of each word in the data, generate a word frequency table, set an occurrence probability threshold, for example, 0.05 (that is, a word appears in 5% of the data), and words exceeding the threshold will be regarded as keywords.
[0039] The occurrence probability refers to the frequency of a certain word appearing in the entire dataset. The calculation formula is: Occurrence probability = (Number of times the word appears / Total number of words) The threshold is a pre-set standard for screening out important keywords. The set threshold can be 5%, that is, a certain word must appear in at least 5% of the records in the data to be considered a meaningful keyword.
[0040] For example, within a week, 1000 user comments are collected. The following are some words and their occurrence times: "LOL" appears 150 times "New hero" appears 80 times "Skin" appears 50 times Calculate their occurrence probabilities: The occurrence probability of "LOL": 150 / 1000 = 0.15 (15%) The occurrence probability of "New hero": 80 / 1000 = 0.08 (8%) The occurrence probability of "Skin": 50 / 1000 = 0.05 (5%) If the set threshold is 5%, then these words all exceed the threshold, so they can be used as keywords for recommending live broadcast content.
[0041] S303, obtaining a tag corresponding to the live content, matching the tag with a keyword, and pushing the corresponding live content through the keyword, wherein the tag is a summary of the live content.
[0042] In this step, obtain the corresponding label of the live content, match the extracted keywords with the keywords of the live content, find out the highly relevant live content, and use the content-based recommendation algorithm to push the live broadcast according to the keyword similarity. In the collected user data, you cleaned up irrelevant information and stop words, and performed word segmentation. You found that the words frequently used by users include "LOL", "new hero", "skin", "match", "teammate", etc. After determining LOL, new hero and skin as keywords, match the keywords with the live content, and find that there is a recent live broadcast event about "LOL new hero" on the platform. The label with LOL new hero is the corresponding label of the live content. According to the user's keywords, push this live broadcast to relevant users.
[0043] like Figure 4 As shown, as a preferred embodiment of the present invention, the step of obtaining online survey results, sending the online survey to a user terminal, and obtaining the user's preferences, viewing habits, and expectations, adjusting the push live broadcast according to the online survey, and judging the suitability of the push live broadcast according to the adjusted live broadcast score, specifically includes: S401, sending survey data to the user terminal, obtaining online survey results, and determining the live broadcast satisfaction according to the survey results.
[0044] In this step, the survey data is sent to the user terminal, and the user's rating data for each question is collected and organized into a table. Each question corresponds to a score, and the average score of each question is calculated. The online survey results are obtained to obtain the satisfaction score of each dimension, and the live broadcast satisfaction is determined based on the survey results.
[0045] For example, if a question uses a 1-5 rating and the average score is 4.2, it means that users are generally satisfied on that dimension.
[0046] S402, adjusting the push of the live broadcast according to the live broadcast satisfaction, and obtaining a second score according to the adjusted live broadcast score.
[0047] In this step, the push live broadcast is adjusted according to the live broadcast satisfaction. Based on the adjusted live broadcast score, it is a common and effective method to use the Sigmoid function to score the live broadcast platform users to determine whether they become potential users. The Sigmoid function can map the input value to a range between 0 and 1, which can be understood as the probability or rating score of the user becoming a potential user. Score=
[0048] When z→∞, e -z →0, then Score→1, indicating a high probability of becoming a potential user; When z→-∞, e -z →∞, then Score→0, indicating a low probability of becoming a potential user; When z=0, Score=0.5, indicating a medium probability of becoming a potential user; A second score may be obtained to determine the suitability of the re-pushed live broadcast.
[0049] S403, comparing the first score with the second score to determine the suitability of pushing the live broadcast.
[0050] In this step, the first score is compared with the second score. If the second score is higher than the first score, it means that the re-pushed live broadcast is more suitable. If the second score is higher than the first score, it means that the re-pushed live broadcast is less suitable, which means that the re-pushed live broadcast is unqualified.
[0051] like Figure 5 As shown, a system for identifying potential users of a live broadcast platform provided by an embodiment of the present invention includes: The data acquisition module 100 is used to acquire the user's behavior data on the platform, the behavior data including viewing time, comments and sharing data, and acquire user feature information.
[0052] In this system, the data acquisition module 100 acquires the user's behavior data on the platform, collects the user's behavior data, including viewing time, viewing frequency, likes and comments, and can predict each user's activity and participation in the platform based on the user's viewing time, viewing frequency, likes and comments, and viewing content preferences, acquire user feature information, and know the user's age group, gender and geographical location; Behavioral data and feature information can be used to understand user preferences and assess the user's activity and participation in the platform. For example, if user A has a longer average viewing time and a higher frequency of likes and shares, the system can assign a higher behavioral score to him or her, indicating that he or she may be a potential user who actively participates in the platform. By analyzing basic information such as age, gender, and geographic location, combined with their activity and influence on social media, if user B is an influential opinion leader or has characteristics consistent with the platform's target users (such as being young and active in the Internet community), the system will give a higher score and identify them as having greater potential.
[0053] The first scoring module 200 is used to score according to the behavior data to obtain a first score, compare the first score with a first preset value, and determine the user as a potential user if the first score is higher than the first preset value.
[0054] In this system, the first scoring module 200 scores according to the behavior data to obtain a first score, compares the first score with a first preset value, and determines the user as a potential user if the first score is higher than the first preset value.
[0055] In this step, the behavioral data is scored, and the first score is obtained by the Sigmoid function. The first score is compared with the first preset value. When the first score is higher than the first preset value, it is determined to be a potential user. When the first score is lower than the first preset value, it is determined to be a non-potential user, because it can be achieved by analyzing the user's comment content, likes and sharing behaviors. If the user frequently participates in a specific type of live broadcast or shows a positive emotional attitude, it may indicate that he has a high attraction and participation in the platform content, and is more likely to become a potential user; Sigmoid functions are often used to map input values to a range between 0 and 1, which is suitable for scoring probabilities or potentials. For example: Score= Among them, z is the weighted sum of the input or the model output, and the output of the Sigmoid function can be interpreted as the probability or potential score of the user becoming a potential user.
[0056] The threshold module 300 is used to obtain user live broadcast related platform data, extract keywords through the related platform data, and the keywords are words with a probability exceeding a threshold in the related platform data, and push live broadcasts based on the keywords.
[0057] In this system, the threshold module 300 obtains user live broadcast related platform data. The live broadcast platform usually provides an API interface for developers to obtain user data, including comments, bullet screens, likes, etc. To obtain this data, you need to apply for access rights, comply with the platform's usage agreement and data usage policy, mine and analyze the collected user text data, extract keywords through relevant platform data, and calculate the probability of occurrence of words. The probability of occurrence refers to the frequency of a certain word appearing in the entire data set. The calculation formula is: Occurrence probability = number of word occurrences / total number of words / probability of occurrence / total number of words / number of word occurrences When the probability of occurrence exceeds the threshold, it is determined to be a keyword, and the live broadcast is pushed based on the keyword to achieve the effect of improving user retention.
[0058] The second scoring module 400 is used to obtain online survey results. The online survey is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, adjust the push live broadcast according to the online survey, and judge the suitability of the push live broadcast according to the adjusted live broadcast score.
[0059] In this system, the second scoring module 400 obtains the online survey results. In order to collect the user's satisfaction with the live broadcast pushed by the platform, a detailed questionnaire can be designed. The questionnaire should include questions of multiple dimensions to comprehensively evaluate the user's satisfaction, and can obtain the user's preferences, viewing habits and expectations, and obtain the user's satisfaction with the current live broadcast push. Adjustments are made according to the user's satisfaction, and the live broadcast is pushed again. According to the adjusted live broadcast score, the suitability of the live broadcast push is judged, and the Sigmoid function is used again to score the probability or potential.
[0060] like Figure 6 As shown, as a preferred embodiment of the present invention, the first scoring module 200 includes: The activity unit 201 is used to determine the activity of the user according to the behavior data and feature information, and the activity is calculated by a scoring function to calculate the probability of a potential user.
[0061] In this module, the activity unit 201 determines the user's activity based on the behavior data and feature information, and evaluates the average duration or cumulative viewing time of each live broadcast watched by the user. Long-term viewing usually means that the user has a high interest and recognition in the content, and the user's interactive behaviors such as likes, comments, and sharing. Frequent interaction indicates that the user is willing to participate in community communication, has a high degree of investment and interaction in the platform content, and analyzes the user's viewing frequency and continuous usage time. Users who frequently return to the platform and continue to consume content are more likely to become potential users; When users have a high level of interest and recognition in content, a high level of investment and interaction in platform content, and a high viewing frequency and continuous usage time, the user's activity is determined to be high. The probability of a potential user can be calculated through activity. The Sigmoid function is used to map the input value to a range between 0 and 1. Score→1 indicates a high probability of becoming a potential user, and Score→0 indicates a low probability of becoming a potential user.
[0062] The supplementation unit 202 is used to obtain the user's viewing habits for different types of content, where the viewing habits include the channels, hosts or specific types of live broadcasts watched, and to supplement the activity level according to preferences.
[0063] In this module, the supplementary unit 202 obtains the user's viewing habits of different types of content, evaluates the user's likes and sharing frequency of live content, which reflects the user's recognition of the content and willingness to promote it, analyzes the user's viewing history, including the channels, anchors or specific types of live broadcasts watched, and understands the user's preferred content types and viewing habits, which helps predict their future engagement; At the same time, users of different age groups and genders have different acceptance and participation methods for content. Geographical location affects users' attention to and willingness to participate in live events in a specific area. By analyzing users' historical behavior patterns and trends and predicting their future activity and willingness to participate on the platform, the accuracy and real-time performance of the scoring function can be further optimized.
[0064] The first scoring unit 203 is used to score according to the activity level to obtain a first score, compare the first score with a first preset value, and determine the user as a potential user if the first score is higher than the first preset value.
[0065] In this module, the first scoring unit 203 scores based on activity. It is a common and effective method to use the Sigmoid function to score live platform users to determine whether they are potential users. The Sigmoid function can map the input value to a range between 0 and 1, which can be understood as the probability or score of the user becoming a potential user. Score=
[0066] When z→∞, e -z →0, then Score→1, indicating a high probability of becoming a potential user; When z→-∞, e -z →∞, then Score→0, indicating a low probability of becoming a potential user; When z=0, Score=0.5, indicating a medium probability of becoming a potential user.
[0067] like Figure 7 As shown, as a preferred embodiment of the present invention, the threshold module 300 includes: The permission unit 301 is used to send a permission request to the user terminal, wherein the permission request is used to obtain platform data related to the live broadcast platform, receive permission to obtain user live broadcast related platform data.
[0068] In this module, the permission unit 301 sends a permission request to the user terminal, and provides an API interface through the live broadcast platform, allowing developers to obtain user comments, bullet screens, likes and other data. By using the API, user interaction data can be obtained in real time and further analyzed; Send permission requests to user terminals, receive permission, obtain user live broadcast related platform data, ensure compliance with relevant laws, regulations and platform regulations when obtaining and processing user data, especially legal requirements related to user privacy protection, take appropriate data anonymization measures for data involving personal identity, and ensure the security of data during transmission and storage to prevent data leakage and abuse.
[0069] A threshold unit 302 is used to count the frequency of word occurrences, set a keyword threshold, compare the word occurrence frequency with the keyword threshold. If the word occurrence frequency is greater than the keyword threshold, it is determined as a keyword, and keywords are extracted through relevant platform data.
[0070] In this module, the threshold unit 302 counts the frequency of word occurrences, obtains the interaction data of users on the live platform, including comments, bullet screens, likes, shares, and follows, removes stop words such as "de", "le", and "shi", etc., which have no practical significance for keyword extraction, counts the frequency of each word in the data, generates a word frequency table, sets an occurrence probability threshold, for example, 0.05 (that is, a word appears in 5% of the data), and words exceeding the threshold will be regarded as keywords.
[0071] The occurrence probability refers to the frequency of a certain word appearing in the entire dataset. The calculation formula is: Occurrence probability = (Number of times the word appears / Total number of words) The threshold is a pre-set standard used to screen out important keywords. The set threshold can be 5%, that is, a certain word is considered a meaningful keyword only when it appears in at least 5% of the records in the data.
[0072] For example, 1000 user comments are collected in a week. The following are some words and their occurrence times: "LOL" appears 150 times "New hero" appears 80 times "Skin" appears 50 times Calculate their occurrence probabilities: The occurrence probability of "LOL": 150 / 1000 = 0.15 (15%) The occurrence probability of "New hero": 80 / 1000 = 0.08 (8%) The occurrence probability of "Skin": 50 / 1000 = 0.05 (5%) If the set threshold is 5%, then these words all exceed the threshold, so they can be used as keywords for recommending live content.
[0073] A matching unit 303 is used to obtain the corresponding tags of the live content, match the tags with the keywords, and push the corresponding live content through the keywords. The tags are the summaries of the live content.
[0074] In this module, the matching unit 303 obtains the corresponding label of the live content, matches the extracted keywords with the keywords of the live content, finds the live content with high relevance, and can use a content-based recommendation algorithm to push the live broadcast according to the keyword similarity. In the collected user data, you cleaned up irrelevant information and stop words, and performed word segmentation. You found that the words frequently used by users include "LOL", "new hero", "skin", "match", "teammate", etc. After determining LOL, new hero and skin as keywords, the keywords are matched with the live content. It is found that there is a live broadcast event about "LOL new hero" on the platform recently, and the label with LOL new hero is the corresponding label of the live content. According to the user's keywords, this live broadcast is pushed to relevant users.
[0075] like Figure 8 As shown, as a preferred embodiment of the present invention, the second scoring module 400 includes: The survey unit 401 is used to send survey data to the user terminal, obtain online survey results, and determine the live broadcast satisfaction according to the survey results.
[0076] In this module, the survey unit 401 sends survey data to the user terminal, collects the user's rating data for each question, and organizes it into a table. Each question corresponds to a score, and the average score of each question is calculated to obtain the online survey results, obtain the satisfaction score of each dimension, and determine the live broadcast satisfaction based on the survey results.
[0077] For example, if a question uses a 1-5 rating and the average score is 4.2, it means that users are generally satisfied on that dimension.
[0078] The second scoring unit 402 is used to adjust the push of live broadcast according to the live broadcast satisfaction, and obtain a second score according to the adjusted live broadcast score.
[0079] In this module, the second scoring unit 402 adjusts the push of live broadcast according to the live broadcast satisfaction. According to the adjusted live broadcast score, it is a common and effective method to score the live broadcast platform users using the Sigmoid function to determine whether they become potential users. The Sigmoid function can map the input value to a range between 0 and 1, which can be understood as the probability or score of the user becoming a potential user. Score=
[0080] When z→∞, e -z →0, then Score→1, indicating a high probability of becoming a potential user; When z→-∞, e -z →∞, then Score→0, indicating a low probability of becoming a potential user; When z=0, Score=0.5, indicating a medium probability of becoming a potential user; A second score may be obtained to determine the suitability of the re-pushed live broadcast.
[0081] The comparison unit 403 is used to compare the first score with the second score to determine the suitability of pushing the live broadcast.
[0082] In this module, the comparison unit 403 compares the first score with the second score. If the second score is higher than the first score, it means that the re-pushed live broadcast has a higher suitability. If the second score is higher than the first score, it means that the re-pushed live broadcast has a lower suitability, which means that the re-pushed live broadcast is unqualified.
[0083] In one embodiment, a computer device is provided, the computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented: Obtaining user behavior data on the platform, including viewing time, comments and sharing data, and obtaining user feature information; Scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value; Obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcasts based on the keywords; The online survey result is obtained, and the online survey is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, and the push live broadcast is adjusted according to the online survey. The suitability of the push live broadcast is judged according to the adjusted live broadcast score.
[0084] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the processor performs the following steps: Obtaining user behavior data on the platform, including viewing time, comments and sharing data, and obtaining user feature information; Scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value; Obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcasts based on the keywords; The online survey result is obtained, and the online survey is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, and the push live broadcast is adjusted according to the online survey. The suitability of the push live broadcast is judged according to the adjusted live broadcast score.
[0085] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.
[0086] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0087] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0088] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.
[0089] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for identifying potential users of a live broadcast platform, characterized in that: The method comprises: Obtaining user behavior data on the platform, including viewing time, comments and sharing data, and obtaining user feature information; Scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value; Obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcasts based on the keywords; The online survey result is obtained, and the online survey is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, and the push live broadcast is adjusted according to the online survey. The suitability of the push live broadcast is judged according to the adjusted live broadcast score.
2. A method for identifying potential users of a live broadcast platform according to claim 1, characterized in that: The step of scoring the behavior data to obtain a first score, comparing the first score with a first preset value, and determining the user as a potential user if the first score is higher than the first preset value, specifically includes: Determine the user's activity level based on the behavior data and feature information, where the activity level is calculated by a scoring function to calculate the probability of a potential user; Obtaining users' viewing habits of different types of content, including the channels, hosts or specific types of live broadcasts they watch, and supplementing activity levels based on preferences; Scoring is performed based on activity to obtain a first score, and the first score is compared with a first preset value. If the first score is higher than the first preset value, the user is determined to be a potential user.
3. A method for identifying potential users of a live broadcast platform according to claim 1, characterized in that: The step of obtaining user live broadcast related platform data, extracting keywords through the related platform data, wherein the keywords are words with a probability exceeding a threshold in the related platform data, and pushing live broadcast according to the keywords specifically includes: Sending a permission request to the user terminal, the permission request is used to obtain platform data related to the live broadcast platform, receiving the permission to obtain the user's live broadcast related platform data; Count the frequency of word occurrences, set a keyword threshold, compare the frequency of word occurrences with the keyword threshold, and if the frequency of word occurrences is greater than the keyword threshold, determine it as a keyword and extract the keyword through relevant platform data; Obtain tags corresponding to live content, match tags with keywords, and push corresponding live content through keywords, where the tags are a summary of the live content.
4. A method for identifying potential users of a live broadcast platform according to claim 1, characterized in that: The step of obtaining the online survey result, wherein the online survey is sent to the user terminal, wherein the online survey is used to obtain the user's preferences, viewing habits and expectations, adjusting the push live broadcast according to the online survey, and judging the suitability of the push live broadcast according to the adjusted live broadcast score specifically includes: Send survey data to user terminals, obtain online survey results, and determine live broadcast satisfaction based on the survey results; Adjust push live broadcast according to live broadcast satisfaction, and obtain a second score according to the adjusted live broadcast score; Compare the first score with the second score to determine the suitability of pushing the live broadcast.
5. The method for identifying potential users of a live broadcast platform according to claim 1, characterized in that: The characteristic information includes age group, gender and geographical location.
6. A system for identifying potential users of a live broadcast platform, characterized in that: The system comprises: A data acquisition module is used to acquire user behavior data on the platform, including viewing time, comments and sharing data, and to acquire user feature information; A first scoring module scores the behavior data to obtain a first score, compares the first score with a first preset value, and determines the user as a potential user if the first score is higher than the first preset value; The threshold module obtains the user's live broadcast related platform data, extracts keywords through the related platform data, and the keywords are words with a probability exceeding the threshold in the related platform data, and pushes the live broadcast according to the keywords; The second scoring module obtains the results of an online survey, which is sent to the user terminal. The online survey is used to obtain the user's preferences, viewing habits and expectations, adjust the push live broadcast according to the online survey, and judge the suitability of the push live broadcast according to the adjusted live broadcast score.
7. A system for identifying potential users of a live broadcast platform according to claim 6, characterized in that: The first scoring module includes: An activity unit, which determines the activity of a user based on the behavior data and feature information, wherein the activity is calculated by a scoring function to calculate the probability of a potential user; A supplementation unit, which obtains the user's viewing habits of different types of content, wherein the viewing habits include the channels, anchors or specific types of live broadcasts watched, and supplements the activity level according to the preferences; The first scoring unit scores the activity to obtain a first score, compares the first score with a first preset value, and determines the user as a potential user if the first score is higher than the first preset value.
8. A system for identifying potential users of a live broadcast platform according to claim 7, characterized in that: The threshold module comprises: The permission unit sends a permission request to the user terminal, wherein the permission request is used to obtain platform data related to the live broadcast platform, receives the permission, and obtains the user's live broadcast related platform data; Threshold unit, counts the frequency of word occurrence, sets the keyword threshold, compares the word frequency with the keyword threshold, and if the word frequency is greater than the keyword threshold, it is determined to be a keyword, and the keyword is extracted through relevant platform data; The matching unit obtains tags corresponding to the live content, matches the tags with keywords, and pushes corresponding live content through the keywords, wherein the tags are a summary of the live content.
9. A system for identifying potential users of a live broadcast platform according to claim 8, characterized in that: The second scoring module includes: A survey unit sends survey data to a user terminal, obtains online survey results, and determines live broadcast satisfaction based on the survey results; The second scoring unit adjusts the push of live broadcast according to the live broadcast satisfaction, and scores the adjusted live broadcast to obtain a second score; The comparison unit compares the first score with the second score to determine the suitability of pushing the live broadcast.
10. A system for identifying potential users of a live broadcast platform according to claim 9, characterized in that: The characteristic information includes age group, gender and geographical location.
Citation Information
Patent Citations
Method and equipment for recommending content of Internet based on user browse behavior
CN101968802A
Video recommendation method and device
CN106231428A
A method and a device for extracting potential users of a live broadcast platform
CN108683949A
Apparatus and method for analyzing user reviews
KR102505204B1
