Match simulation prediction method and system based on user interaction behavior

Abstract

The present invention discloses a method and system for simulating prediction of events based on user interaction behavior, and relates to the technical field of sports event broadcasting, which includes S1, data cleaning, which includes: performing data cleaning on original text data to obtain text data one; S2, simulation trigger analysis, which includes: performing feature extraction on text data one, and judging whether each sentence meets the predefined argument scoring conditions, and if so, performing scoring; counting the scores of audience speech data in previous argument events to obtain a total score; if the total score is greater than a preset simulation trigger threshold, performing event simulation prediction; S3, event simulation prediction. The present application has the effect of helping audiences analyze and predict the results of event behaviors, and reducing audiences' quarrels about event development and results.

Description

Technical Field

[0001] The present application relates to the technical field of sports event broadcasting, and in particular to a method and system for event simulation prediction based on user interaction behavior. Background Art

[0002] During the broadcast and live broadcast of sports events, the audience, especially the corresponding senior sports enthusiasts, often express their opinions in areas such as the barrage. The number of spectators for a sports event ranges from tens of thousands to hundreds of thousands or even millions. Therefore, it is inevitable that sometimes the barrage will be divided into several main arguments because of what each party thinks is the correct behavior and strategy for the athletes to adopt, and even because of the inability to convince each other, uncivilized behavior such as verbal abuse will occur.

[0003] Therefore, it is necessary to propose a new technical solution to help spectators analyze and predict the results of event behaviors, reduce arguments among spectators about event development and results, and build a civilized environment for viewing and discussing events. Summary of the invention

[0004] In order to help the audience analyze and predict the results of event behaviors and reduce the audience's arguments about the development and results of the event, the present application provides an event simulation prediction method and system based on user interactive behavior.

[0005] In a first aspect, the present application provides a method for simulating and predicting an event based on user interaction behavior, which adopts the following technical solution:

[0006] A method for simulating and predicting a competition based on user interaction behavior, comprising:

[0007] S1. Data cleaning, which includes:

[0008] Access permissions to obtain audience speech data on social platforms, and perform data consistency processing to obtain original text data;

[0009] Perform data cleaning on the original text data to obtain text data 1;

[0010] S2. Simulation trigger analysis, which includes:

[0011] Extract features from the text data and determine whether each sentence meets the predefined argument scoring conditions. If yes, score it;

[0012] Count the scores of audience speech data in all previous debate events to get the total score;

[0013] If the total score is greater than the preset simulation trigger threshold, a match simulation prediction is performed;

[0014] S3. Event simulation prediction, which includes:

[0015] Output the start prompt of the event simulation to the social interface of the social platform, and request to establish a data connection with the preset simulation platform;

[0016] Perform natural language processing on the real-time text data 1 to obtain text data 2;

[0017] Perform simulation control behavior recognition on text data 2, and classify and count the number of different types of simulation control behaviors and calculate the proportion;

[0018] The simulation control behavior with the largest proportion is taken as a condition, a preset simulation instruction set is retrieved, and a matching simulation instruction is called;

[0019] Send simulation instructions to the simulation platform and obtain corresponding simulation prediction results;

[0020] Output simulation prediction results to the social interface of the social platform.

[0021] Optionally, the determination of the dispute event includes:

[0022] locate the position of the presupposed argument-triggering vocabulary;

[0023] Extract the speech data of the same audience in the previous time based on the position of the debate triggering words;

[0024] If the speech data meets the preset debate conditions, it is determined to be a debate event, and the time when the second or more audience members have different opinions in the same event is defined as the start time of the debate event.

[0025] Optionally, also include:

[0026] Obtain the athlete's identity information, physical indicators and playing time, and record them as athlete attributes;

[0027] Send the player attributes to the simulation platform.

[0028] Optionally, also include:

[0029] Establish an independent database matching the contracting parties based on the athlete's identity information and pre-set contracting records;

[0030] Define the access permissions of the independent database for specified users and simulation platforms;

[0031] If other users outside the simulation platform access the independent database, a data abnormality reading prompt will be sent to the designated user of the contracting party;

[0032] If the independent database is accessed, the access record is recorded and a report is sent to the designated user of the contracting party.

[0033] Optionally, also include:

[0034] If the athlete's physical indicators cannot be obtained, the athlete's historical game records are called up to perform similar scene analysis;

[0035] Adjust the simulation prediction logic of the simulation platform based on the similarity scenario analysis results.

[0036] Optionally, also include:

[0037] Identify the account that triggered the dispute event and the account that participated in the dispute event, and identify abnormal accounts;

[0038] If an account is judged as an abnormal account, it will no longer be scored in the dispute event and a summary report of the abnormal account will be sent to the designated management party.

[0039] Optionally, also include:

[0040] If the event simulation prediction is triggered, a preset authentication request will be initiated to the abnormal account;

[0041] If the authentication is successful, the account permissions are restored and the speech is simulated and controlled.

[0042] In the second aspect, the present application provides a game simulation prediction system based on user interaction behavior, which adopts the following technical solutions:

[0043] A system for simulating and predicting a match based on user interaction behavior comprises a memory and a processor, wherein the memory stores a computer program that can be loaded and executed by the processor to implement any of the above-mentioned methods for simulating and predicting a match based on user interaction behavior.

[0044] In summary, the present application includes the following beneficial technical effects: simulation technology can be used to interact with the audience, convert the audience's speeches into simulation control instructions, simulate the audience's views, provide support for some views through simulation prediction results, and reduce the audience's arguments about the development and results of the event. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the main process of the method of this application. DETAILED DESCRIPTION

[0046] The following is combined with Figure 1 This application is described in further detail.

[0047] The embodiment of the present application discloses a method for event simulation prediction based on user interaction behavior.

[0048] Reference Figure 1 ,The match simulation prediction method based on user interaction behavior includes:

[0049] S1. Data cleaning;

[0050] S2, simulation trigger analysis;

[0051] S3. Event simulation prediction.

[0052] This method aims to use simulation technology to interact with the audience when they have disagreements and cannot convince each other, convert the audience's speech into simulation control instructions, simulate the audience's views, and provide support for some views through simulation prediction results, reduce the audience's arguments about the development and results of the event, and build a civilized environment for viewing and discussing the event. The specific explanation is as follows.

[0053] Regarding data cleaning, it includes:

[0054] 1) Access permissions, obtain audience speech data from the social platform, and perform data consistency processing to obtain the original text data.

[0055] It is understandable that crawler technology can collect relevant data on the Internet, but if there is no relevant platform or user authorization, it is illegal; therefore, this method first needs to obtain relevant authorization to collect and monitor the audience's speeches.

[0056] Data consistency processing: Audiences can speak in voice, text, or graphics, which hinders machine learning analysis of speech content. Therefore, it is necessary to unify them, such as unifying them into text. Therefore, this step includes audio-to-text conversion and image recognition.

[0057] 2) Clean the original text data to obtain text data 1;

[0058] It is understandable that social platforms are full of various audience comments, many of which are irrelevant to the problem to be solved by this application. Such information is useless and therefore needs to be removed; text cleaning example: staff predefine stop words and delete them; delete repeated comments by the same user within a short period of time (within 1 minute); delete punctuation marks.

[0059] Regarding simulation trigger analysis, it includes:

[0060] 1) Extract features from text data and determine whether each sentence meets the predefined argument scoring conditions. If so, assign points.

[0061] In this embodiment, taking football and basketball as examples, the argument sentence is such as "Whenever XX passes the ball to XXX", the feature here refers to the keywords, "whenever", "ball", and "give", and the argument scoring condition is that at least "whenever", "ball" or "whenever" and "ball" appear at the same time.

[0062] Better than the above, AI parses the synonyms of the features to be extracted, and then performs feature extraction and analysis. The AI ​​can be kimi, iFlytek Spark, ChatGPT, etc.; this setting can effectively improve the accuracy and flexibility of scoring.

[0063] 2) Count the scores of audience speeches in all previous debates and get the total score;

[0064] If the total score is greater than the preset simulation trigger threshold, a match simulation prediction is performed.

[0065] Regarding the simulation prediction of the event, it includes:

[0066] 1) Output the start prompt of the event simulation to the social interface of the social platform, and request to establish a data connection with the preset simulation platform.

[0067] In this method, behavioral action simulation is needed to predict the results of the game. Therefore, the staff needs to establish a platform for action simulation in advance. An example of a simulation platform is NBA2K14. The staff imports the athlete's data to generate the corresponding characters, providing a basis for the behavioral simulation of the game.

[0068] It should be noted that the implementation of the simulation requires real-time tracking and updating of the game situation. For example, when performing a passing simulation, the positions of each athlete need to be arranged accordingly.

[0069] 2) Perform natural language processing on the real-time text data 1 to obtain text data 2.

[0070] Natural language processing is an existing technology like the aforementioned image recognition and audio-to-text conversion, so I will not go into details. Natural language processing includes part-of-speech tagging, entity recognition, etc.; the purpose of this link is to provide a basis for extracting information for analysis later. Another form: vectorize text information and assign binary vectors to each word.

[0071] 3) Perform simulated control behavior recognition on text data 2, and classify and count the number of speeches of different types of control behaviors and calculate the proportion.

[0072] Taking passing as an example, athlete A1 passes the ball to athlete A2, who is then intercepted and loses the ball. The audience holds various opinions in the debate: the first is that the ball should be passed to athlete A3; the second is that the ball should be passed to athlete A4. After the debate triggers the simulation, some audience members express the view that the ball should be passed to athlete A3; some audience members express the view that the ball should be passed to athlete A4.

[0073] In the above example, the text data 2 is subjected to simulation control behavior recognition, that is, the above-mentioned pass to athletes A3 and A4 is recognized, the number of speeches of the two is calculated respectively, and then the proportion of the total is calculated.

[0074] 4) Take the simulation control behavior with the largest proportion as the condition, retrieve the preset simulation instruction set, and call the matching simulation instruction.

[0075] The simulation instruction set corresponds to the above-mentioned simulation platform. After the control behavior is selected, the corresponding matching is searched, such as: searching for the pass control instruction, and defining the objects of the pass to be A3 and A4 athletes.

[0076] 5) Send simulation instructions to the simulation platform and obtain the corresponding simulation prediction results.

[0077] In the above example, the simulation result may be text information corresponding to whether the players A3 and A4 get the ball, or an animation of the entire ball passing process.

[0078] 6) Output the simulation prediction results to the social interface of the social platform.

[0079] Based on the above, this method can use simulation technology to interact with the audience, convert the audience's speech into simulation control instructions, simulate the audience's views, provide support for some views through simulation prediction results, and reduce the audience's arguments about the development and results of the event.

[0080] In one embodiment of the method, the determination of the above dispute event includes:

[0081] 1) Locate the position of the preset debate trigger words; for example, define uncivilized words as debate trigger words, and after identifying the uncivilized words, locate the order in which they appear in the barrage.

[0082] 2) Extract the speech data of the same audience in the previous time based on the position of the debate triggering words.

[0083] 3) If the speech data meets the preset debate conditions, it is determined to be a debate event, and the time when the second or more audience members have different opinions in the same event is defined as the start time of the debate event.

[0084] According to the above content, this method does not count all the disputes in the whole game to trigger the simulation, but divides it into each dispute, and statistically analyzes each time separately, so that the simulation is triggered relatively necessary.

[0085] In another embodiment of the present application, the method further comprises:

[0086] Obtain the athlete's identity information, physical indicators and playing time, and record them as athlete attributes;

[0087] Send the player attributes to the simulation platform.

[0088] Among them, the athlete's physical indicators can be the athlete wearing a smart wearable device, such as a smart ring or smart bracelet; the smart wearable device reports its data to the corresponding health management platform and obtains data from the health management platform. The athlete's playing time can be used to evaluate his physical endurance and provide a basis for action simulation.

[0089] According to the above settings, the simulation prediction results can be made more accurate and more convincing when supporting the audience's opinions. It should be noted that the athlete attributes are not only used by the simulation platform, but also can provide reference for the corresponding coaches to reasonably plan the competition strategy.

[0090] Since most sports events are conducted through teamwork, if the physical indicators of a player on the field are obtained by the coach of the opposing team, it is easy to formulate offensive and defensive strategies in a targeted manner, which will affect the results of the game. This will cause the players to resist using the above content of this method. Therefore, this method is further configured as follows:

[0091] Establish an independent database matching the contracting parties based on the athlete's identity information and the pre-set contracting records (i.e., information indicating the team or organization to which the athlete belongs);

[0092] Define the access rights of the independent database to designated users and simulation platforms; designated users include the coach of the team;

[0093] If other users outside the simulation platform access the independent database, a data abnormality reading prompt will be sent to the designated user of the contracting party;

[0094] If the independent database is accessed, the access record is recorded and a report is sent to the designated user of the contracting party; the information in the report may be the access time, number of times, and visitor.

[0095] According to the above settings, the probability of athlete information being leaked can be reduced, and relevant personnel can receive prompts in time when data is read, so that strategies can be formulated in time when data leakage occurs accidentally, such as: asking athletes to take off smart wearable devices to avoid athletes being targeted and strategies being formulated, so that the game can proceed normally. At the same time, the above also records any access to prevent the identity information of authorized users from being stolen and leaking athlete attributes.

[0096] Based on the above, whether the athlete takes the initiative to take off the smart wearable device, or the coach finds that the independent database is abnormally accessed and there is a risk of information leakage and asks the athlete to take off the smart wearable device, the simulation platform will no longer be able to update the athlete's attributes in real time, affecting the accuracy of the simulation. Therefore, this method is set as follows:

[0097] If the athlete's physical indicators cannot be obtained, the athlete's historical game records are called to perform similar scene analysis. The historical game records can be the athlete's historical competition video data and the game graphic broadcast data. The data is pre-processed with behavior recognition, scene recognition, and portrait recognition technology to achieve dynamic tracking of each athlete's game and obtain recognition results; the recognition results are cleaned and classified and stored, for example: passing records are classified into one category, and records of defending B1 athletes are classified into another category.

[0098] Regarding similarity scene analysis, for example: after the record is converted into text, it is processed by the aforementioned vectorization process, and then the similarity is calculated; or, the image record is retained, and the current event image is obtained for image similarity analysis.

[0099] This method: adjusts the simulation prediction logic of the simulation platform based on the similarity scenario analysis results.

[0100] Example: Take the four most similar historical game records, and assign weight coefficients to the results according to preset rules; the simulation platform continues to simulate the behavior and action without updating the attributes of the athletes, and assigns weight calculations; weight calculation, take the most likely result as the simulation prediction result. It is understandable that the simulation platform can simulate once or multiple times.

[0101] According to the above settings, when it is impossible to obtain the athlete's physical indicators in real time to update the data of the simulation platform, this method can also call the athlete's historical game records to correct the simulation prediction results, making the simulation prediction results more accurate and convincing; under certain conditions, such as when the simulation software crashes and stops responding, on the basis of the above, this method can also use historical game records to complete predictions to meet certain usage requirements.

[0102] In another embodiment of the method, in order to prevent some individuals and organizations from illegally using the simulation prediction mechanism of the method, such as using a large number of puppet accounts to control the speech tendency of social platforms, triggering simulations for illegal purposes to infer the physical indicators of certain athletes, etc., the method is further configured as follows:

[0103] Identify the account that triggered the dispute event and the account that participated in the dispute event, and identify abnormal accounts;

[0104] If an account is judged as an abnormal account, it will no longer be scored in the dispute event and a summary report of the abnormal account will be sent to the designated management party.

[0105] According to the above settings, the probability of this method being illegally used can be reduced, and at the same time, it can also provide a basis for the operation and management of the social platform to clean up puppet accounts in a timely manner, reduce malicious guidance of speech, "setting the pace", and build a healthy viewing environment.

[0106] Regarding abnormal account identification, example:

[0107] 1) Identify the account creation time. If the user creation time is less than the preset lower limit, and the proportion of controversial and bad speeches in the published speech exceeds the preset abnormal speech trigger threshold, it is judged as an abnormal account;

[0108] 2) If the account has not undergone real-name authentication, and the proportion of repetitive speeches and meaningless speeches irrelevant to the event exceeds the preset judgment threshold, it will be judged as an abnormal account.

[0109] The above-mentioned identification of abnormal accounts is only an example. It is better to establish relevant learning models or AI identification to improve the accuracy of identification.

[0110] In another embodiment of the method, in order to reduce the impact of puppet account identification errors, it is further configured as follows:

[0111] If the event simulation prediction is triggered, a preset authentication request will be initiated to the abnormal account;

[0112] If the authentication is successful, the account permissions are restored and the speech is simulated and controlled.

[0113] According to the above settings, if the event simulation prediction is still triggered after checking for abnormal accounts, this method will initiate authentication for the account previously identified as abnormal to verify whether it is a puppet, and after the authentication is passed, its biased speech will be re-included in the calculation of the crowd control simulation process, reducing the impact of the judgment error while improving the user experience. Relatively speaking, the number of authentications required at this time is relatively small, which puts less pressure on the server implementing the method and has a low negative impact.

[0114] It is understandable that the above-mentioned authentication can be unique biometric information authentication such as face, fingerprint, etc.

[0115] The embodiment of the present application also discloses a game simulation prediction system based on user interaction behavior.

[0116] A match simulation prediction system based on user interaction behavior includes a memory and a processor, wherein the memory stores a computer program that can be loaded and executed by the processor to implement any of the above-mentioned match simulation prediction methods based on user interaction behavior.

[0117] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for simulating and predicting events based on user interaction behavior, characterized in that: include: S1. Data cleaning, which includes: Access permissions to obtain audience speech data on social platforms, and perform data consistency processing to obtain original text data; Perform data cleaning on the original text data to obtain text data 1; S2. Simulation trigger analysis, which includes: Perform feature extraction on the text data 1, and determine whether each sentence meets the predefined argument scoring conditions. If so, assign points. If the event corresponds to football or basketball, the argument scoring condition is that at least the keywords: any, ball or the keywords: any, ball appear at the same time. Count the scores of audience speech data in all previous debate events to get the total score; If the total score is greater than the preset simulation trigger threshold, a match simulation prediction is performed; S3. Event simulation prediction, which includes: Output the start prompt of the event simulation to the social interface of the social platform, and request to establish a data connection with the preset simulation platform; Perform natural language processing on the real-time text data 1 to obtain text data 2; Perform simulation control behavior recognition on text data 2, and classify and count the number of different types of simulation control behaviors and calculate the proportion; simulation control behaviors include passing the ball to a certain athlete; The simulation control behavior with the largest proportion is taken as a condition, a preset simulation instruction set is retrieved, and a matching simulation instruction is called; Send simulation instructions to the simulation platform and obtain corresponding simulation prediction results; Output simulation prediction results to the social interface of the social platform.

2. The method for simulating and predicting a match based on user interaction behavior according to claim 1, characterized in that: The determination of the dispute event includes: locate the position of the presupposed argument-triggering vocabulary; Extract the speech data of the same audience in the previous time based on the position of the debate triggering words; If the speech data meets the preset dispute conditions, it is determined to be a dispute event.

3. The event simulation prediction method based on user interaction behavior according to claim 1 is characterized in that: Also includes: Obtain the athlete's identity information, physical indicators and playing time, and record them as athlete attributes; Send the player attributes to the simulation platform.

4. The event simulation prediction method based on user interaction behavior according to claim 3 is characterized in that: Also includes: Establish an independent database matching the contracting parties based on the athlete's identity information and pre-set contracting records; Define the access permissions of the independent database for specified users and simulation platforms; If other users outside the simulation platform access the independent database, a data abnormality reading prompt will be sent to the designated user of the contracting party; If the independent database is accessed, the access record is recorded and a report is sent to the designated user of the contracting party.

5. The method for simulating and predicting a match based on user interaction behavior according to claim 4, characterized in that: Also includes: If the athlete's physical indicators cannot be obtained, the athlete's historical game records are called up to perform similar scene analysis; Adjust the simulation prediction logic of the simulation platform based on the similarity scenario analysis results.

6. The method for simulating and predicting a match based on user interaction behavior according to claim 4, characterized in that: Also includes: Identify the account that triggered the dispute event and the account that participated in the dispute event, and identify abnormal accounts; If an account is judged as an abnormal account, it will no longer be scored in the dispute event and a summary report of the abnormal account will be sent to the designated management party.

7. The method for simulating and predicting a match based on user interaction behavior according to claim 6, characterized in that: Also includes: If the event simulation prediction is triggered, a preset authentication request will be initiated to the abnormal account; If the authentication is successful, the account permissions are restored and the speech is simulated and controlled.

8. A match simulation prediction system based on user interaction behavior, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program that can be loaded and executed by the processor to implement the event simulation prediction method based on user interaction behavior as described in any one of claims 1-7.

Images