Online live broadcasting room data analysis decision system and method

Abstract

The invention discloses an online live broadcast room data analysis decision system, which relates to a live broadcast data analysis technology, and comprises the following steps: collecting live broadcast data of a live broadcast room from the outside of the system, and carrying out normalization processing on the live broadcast data to obtain normalized feature data; performing analysis and calculation processing on the normalized feature data to obtain a live broadcast comprehensive score; carrying out intelligent decision making on a specific live broadcast mode according to the live broadcast comprehensive score so as to select a target live broadcast mode; and performing live broadcast layout on a new media platform according to the target live broadcast mode. The invention also discloses an online live broadcast room data analysis decision method. The system has the advantages of being objective, high in efficiency, accurate in data, standardized, high in anti-interference performance and the like. The system can guide merchants to better operate the live broadcast matrix, thereby giving full play to respective advantages of real persons and digital persons, and obtaining maximum benefits.

Description

Technical Field

[0001] The present invention relates to live broadcast data analysis technology, and more specifically, to an online live broadcast room data analysis decision system and method. Background Art

[0002] Online live streaming is a new e-commerce marketing model that has emerged in recent years. It refers to a sales model in which the anchor shows and recommends products to the audience through live broadcast in the live broadcast room, and the audience can watch and buy the products in real time. Online live streaming can be done in real person, digital person, or a combination of real person and digital person. Live broadcast by real person is costly, fatigue-inducing, and the location of live broadcast is also limited. Live broadcast by digital person is low-cost and can be broadcast at any time and place, and can be broadcast uninterruptedly. For 24 hours a day, the benefits of live broadcast are different in different time periods. Therefore, in order to maximize profits, merchants should arrange the live broadcast method reasonably. However, there is no ready-made technology to comprehensively evaluate the benefits of live broadcast through technical means and make decisions on the selection of live broadcast methods based on the benefits of live broadcast.

[0003] In general, many platforms are currently able to provide various live broadcast data (number of viewers, number of likes, number of orders, etc.) in real time, but they still mainly rely on manual methods to judge the effectiveness of live broadcasts and make decisions on live broadcast methods. There is no standardized software or system.

[0004] There are many disadvantages in judging the benefits of live broadcasts manually, including subjectivity and bias, inefficiency, inaccurate data, lack of standardization, and susceptibility to external interference. In terms of subjectivity and bias, manual judgment is easily affected by subjective factors such as personal preferences, experience, and emotions, resulting in a lack of objectivity and fairness in the judgment results; different judges may have different views and evaluation criteria on the benefits of the same live broadcast, resulting in inconsistent judgment results. In terms of efficiency, manual judgment requires a lot of time and manpower costs, especially for large-scale live broadcast activities, the judgment process may become very cumbersome and time-consuming; in addition, the judge needs to watch the entire live broadcast process and make detailed records and analysis, which increases the difficulty and complexity of the judgment. In terms of data, manual judgment may not be able to accurately capture and record all the key data in the live broadcast process, resulting in inaccurate judgment results; the judge may not be able to track and update the data in real time, resulting in data delays or omissions. In terms of standardization, manual judgment lacks unified standards and processes, resulting in reduced comparability and credibility of the judgment results; different judges may use different judgment criteria and methods, making it difficult to compare and comprehensively analyze the judgment results. In terms of interference, manual judgment is easily interfered by external factors, such as the judge's personal relationship, conflict of interest, etc., which may affect the fairness and objectivity of the judgment. In addition, relying on manual methods to make live broadcast decisions also has disadvantages such as subjectivity and experience dependence, low decision-making efficiency, insufficient data utilization, lack of flexibility and adaptability, and susceptibility to emotions and pressure. Therefore, in practice, if automated tools, data analysis and other technologies can be used to judge the benefits of live broadcasts and make decisions on live broadcast methods, the accuracy and efficiency of live broadcast decisions can be greatly improved. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide an online live broadcast room data analysis decision system and method in view of the shortcomings of the existing technology, so as to effectively help merchants choose the most suitable live broadcast method in different time periods.

[0006] The online live broadcast room data analysis and decision-making system described in the present invention includes:

[0007] The live broadcast data input module is used to collect the live broadcast data of the live broadcast room from outside the system;

[0008] A live data acquisition module, used for receiving the live data sent by the live data input module, and performing normalization processing on the live data to obtain normalized feature data;

[0009] A live broadcast data analysis module, used to analyze and calculate the normalized feature data to obtain a comprehensive live broadcast score;

[0010] A live broadcast intelligent decision module, used to make an intelligent decision on a specific live broadcast method according to the live broadcast comprehensive score, so as to select a target live broadcast method;

[0011] The live broadcast matrix module is used to arrange the live broadcast on the new media platform according to the target live broadcast mode.

[0012] As a further improvement, the live broadcast data is normalized by the following formula:

[0013]

[0014] Among them, X is the characteristic data in the live broadcast data; X min and X max are the minimum and maximum values ​​of the feature data respectively; Y is the normalized feature data.

[0015] As a further improvement, the comprehensive score of the live broadcast is calculated by the following formula:

[0016] S=(C*W c )+(G*W g )+(I*W i )+(D*W d );

[0017] Among them, S is the comprehensive score of the live broadcast; C is the normalized traffic feature data; G is the normalized product feature data; I is the normalized interaction feature data; D is the normalized transaction feature data; W c is the weight of the traffic feature data; W g is the weight of the product feature data; W i is the weight of the interactive feature data; W d is the weight of the product feature data.

[0018] Furthermore, the weight W of the traffic feature data c , the weight of product feature data W g , the weight of the interactive feature data W i , the weight of product feature data W d The sum is equal to 1.

[0019] As a further improvement, the specific decision-making method of the live broadcast intelligent decision-making module is:

[0020] Assume a time period, and set a scoring threshold range for all the live broadcast modes; obtain the cumulative duration of the comprehensive live broadcast score within each scoring threshold range within the time period, and take the live broadcast mode corresponding to the scoring threshold range with the longest cumulative duration as the target live broadcast mode.

[0021] Furthermore, the live broadcasting methods include live broadcasting by sales experts, live broadcasting by real people / digital people mixed, and live broadcasting by digital people.

[0022] As a further improvement, the live broadcast data input module includes a traffic data channel, a product data channel, an interactive data channel, and a transaction data channel.

[0023] A data analysis and decision-making method for an online live broadcast room, the method comprising:

[0024] The live broadcast data of the live broadcast room is collected from outside the system, and the live broadcast data is normalized to obtain normalized feature data; the normalized feature data is analyzed and calculated to obtain a comprehensive live broadcast score; an intelligent decision is made on a specific live broadcast method according to the comprehensive live broadcast score to select a target live broadcast method; and a live broadcast layout is carried out on a new media platform according to the target live broadcast method.

[0025] As a further improvement, the live broadcast data is normalized by the following formula:

[0026]

[0027] Among them, X is the characteristic data in the live broadcast data; X min and X max are the minimum and maximum values ​​of the feature data respectively; Y is the normalized feature data.

[0028] As a further improvement, the comprehensive score of the live broadcast is calculated by the following formula:

[0029] S=(C*W c )+(G*W g )+(I*W i )+(D*W d );

[0030] Among them, S is the comprehensive score of the live broadcast; C is the normalized traffic feature data; G is the normalized product feature data; I is the normalized interaction feature data; D is the normalized transaction feature data; W c is the weight of the traffic feature data; W g is the weight of the product feature data; W i is the weight of the interactive feature data; W d is the weight of the product feature data.

[0031] Beneficial Effects

[0032] The advantages of the present invention are: the system adopts a modular design and can automatically collect a large amount of live broadcast data; the live broadcast benefits are comprehensively evaluated through standardized calculation methods, and intelligent decisions are made on the specific methods of live broadcast according to the comprehensive benefit scores of live broadcast, so as to help merchants choose the most suitable live broadcast methods in different time periods. The system has the advantages of objectivity, high efficiency, accurate data, standardization and strong anti-interference. The system can guide merchants to better operate the live broadcast matrix, so as to give full play to the respective advantages of real people and digital people and achieve maximum benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A schematic diagram of the results of an online live broadcast room data analysis and decision-making system of the present invention;

[0034] Figure 2 A schematic diagram of a daily curve of the comprehensive score of live broadcast according to the present invention;

[0035] Figure 3 A schematic diagram of a live broadcast matrix of a live broadcast channel of the present invention. DETAILED DESCRIPTION

[0036] The present invention will be further described below in conjunction with the embodiments, but it does not constitute any limitation to the present invention. Any limited number of modifications made by anyone within the scope of the claims of the present invention are still within the scope of the claims of the present invention.

[0037] A data analysis and decision-making method for an online live broadcast room, the method comprising:

[0038] The live broadcast data of the live broadcast room is collected from outside the system, and the live broadcast data is normalized to obtain normalized feature data; the normalized feature data is analyzed and calculated to obtain a comprehensive live broadcast score; an intelligent decision is made on a specific live broadcast method according to the comprehensive live broadcast score to select a target live broadcast method; and a live broadcast layout is carried out on a new media platform according to the target live broadcast method.

[0039] The present invention comprehensively evaluates the benefits of live broadcasting through a standardized calculation method, and makes intelligent decisions on the specific methods of live broadcasting according to the comprehensive benefit scores of live broadcasting, helping merchants to choose the most suitable methods of live broadcasting in different time periods. It has the advantages of objectivity, high efficiency, accurate data, standardization and strong anti-interference, and can guide merchants to better operate the live broadcast matrix, thereby giving full play to the respective advantages of real people and digital people and achieving maximum benefits.

[0040] like Figure 1-Figure 3As shown, based on the above-mentioned online live broadcast room data analysis and decision-making method, the present invention also provides an online live broadcast room data analysis and decision-making system, including a live broadcast data input module, a live broadcast data acquisition module, a live broadcast data analysis module, a live broadcast intelligent decision-making module and a live broadcast matrix module. The following will introduce these five modules in detail based on the static and dynamic structural characteristics of the system.

[0041] About static structural characteristics.

[0042] The live broadcast data input module is mainly responsible for inputting various data of the live broadcast room into the system, including traffic data channel, product data channel, interactive data channel, and transaction data channel.

[0043] The traffic data channel is responsible for inputting traffic feature data. Traffic feature data reflects the attractiveness and exposure of the live broadcast room and is an important indicator for evaluating the popularity of live broadcasts. It mainly includes data such as the number of viewers and viewing time.

[0044] The product data channel is responsible for inputting product feature data. Product feature data mainly reflects the popularity of products in the live broadcast room. It mainly includes data such as the number of product clicks and click-through rate.

[0045] The interactive data channel is responsible for inputting interactive feature data. Interactive feature data reflects the atmosphere of the live broadcast room and the user's participation, and is an important indicator for evaluating the interactive effect of the live broadcast. It mainly includes data such as the number of new fans, the number of likes, and the number of rewards.

[0046] The transaction data channel is responsible for inputting transaction feature data. Transaction feature data directly reflects the sales capacity and commercial value of live streaming. It mainly includes data such as order volume and sales volume.

[0047] The live broadcast data collection module is responsible for collecting live broadcast data related to the live broadcast from the live broadcast data input module, and normalizing the data to provide a basis for subsequent data analysis and decision-making.

[0048] The main function of the live broadcast data analysis module is to analyze and calculate the normalized feature data to obtain a comprehensive score for the live broadcast.

[0049] The main function of the live broadcast intelligent decision module is to make intelligent decisions on specific live broadcast methods based on the comprehensive live broadcast score, so as to select the target live broadcast method. The live broadcast methods include live broadcast by leading people, live broadcast by real people / digital people, and live broadcast by digital people.

[0050] The main function of the live broadcast matrix module is to arrange the live broadcast of several programs on different new media platforms (such as Douyin, Kuaishou, Taobao Live, WeChat Video Account, etc.) to achieve diversified content distribution and wider user coverage. Figure 3As shown, it is assumed that the live broadcast matrix has M rows and N columns, with a total of M*N channels. The live broadcast matrix can adopt three live broadcast modes: live broadcast by leading people, live broadcast by real people / digital people, and live broadcast by digital people.

[0051] The three live broadcasting methods in the live broadcast matrix module are explained as follows: live broadcasting by real-person experts means that all programs on all platforms in the matrix are broadcast live by real-person experts; mixed live broadcasting by real-person / digital humans means that some platforms and programs in the matrix are broadcast live by real people, while other platforms and programs are broadcast live by digital humans; live broadcasting by digital humans means that all programs on all platforms in the matrix are broadcast live by digital humans.

[0052] The dynamic structural characteristics of the system will be introduced below.

[0053] Once the system is powered on, each module will enter normal working state.

[0054] The first step is that the live broadcast data input module collects live broadcast related data, namely live broadcast data, from outside the system (such as live broadcast platform, server, etc.).

[0055] Specifically, the live broadcast data input module will split and extract the live broadcast data. Among them, the traffic data channel will extract the traffic data in the data, the product data channel will extract the product data in the data, the interactive data channel will extract the interactive data in the data, and the transaction data channel will extract the transaction data in the data.

[0056] After the data collected by each channel in the live data input module is extracted, it is sent to the live data acquisition module.

[0057] The second step is that the live data acquisition module performs normalization processing according to the data types of different channels, aiming to scale the data to a specific interval [0,1]. After normalization, the dimensional influence of each live data will be eliminated, so that different data have the same weight in the model, and the comparability of the data will be improved. In the present invention, this method can be used for different feature data in the live data, such as the number of viewers, the number of likes, the number of orders, etc. The specific formula of the normalization algorithm of the present invention is:

[0058]

[0059] Among them, X is the characteristic data in the live broadcast data; X min and X max are the minimum and maximum values ​​of the feature data respectively; Y is the normalized feature data.

[0060] After the live data acquisition module processes the data of the four channels, it sends the normalized feature data to the live data analysis module.

[0061] Step 3: The live broadcast data analysis module receives the normalized feature data sent by the live broadcast data collection module and calculates the comprehensive score of the live broadcast according to the following formula:

[0062] S=(C*W c )+(G*W g )+(I*W i )+(D*W d ).

[0063] Among them, S is the comprehensive score of the live broadcast; C is the normalized traffic feature data; G is the normalized product feature data; I is the normalized interaction feature data; D is the normalized transaction feature data; W c is the weight of the traffic feature data; W g is the weight of the product feature data; W i is the weight of the interactive feature data; W d is the weight of the product feature data. Merchants can adjust the weights according to actual conditions, but W c +W g +W i +W d =1.

[0064] After the live broadcast data analysis module calculates the comprehensive score of the live broadcast, it sends the result to the live broadcast decision module.

[0065] Step 4: The live data decision module continuously receives the live comprehensive score data sent by the live data analysis module and records it in two-dimensional coordinates. The vertical axis represents the live comprehensive score, and the horizontal axis divides 1 day 24 hours into several time periods such as t1, t2, t3, etc., thus forming the following Figure 2 The daily curve of the comprehensive score of live broadcast is shown, which represents the live broadcast situation on a certain day.

[0066] from Figure 2 It can be seen that the daily curve of the comprehensive score of live broadcast reflects the busyness or idleness of live broadcast at different time periods of the day. Two dividing lines are drawn according to the score, namely the busy line and the idle line. The area above the busy line is defined as the busy area, the area between the busy line and the idle line is defined as the general area, and the area below the idle line is defined as the idle area.

[0067] The live data decision module counts the time duration of the curve in each area within a certain time period tx, and calculates which area has the longest duration according to the following formula. The characteristic value F of the time period is assigned to a (prime time), b (normal time), and c (idle time).

[0068]

[0069] If F=a, the live broadcast data decision module sends the live broadcast instruction of the goods influencer to the live broadcast matrix module; if F=b, the live broadcast data decision module sends the real person / digital human mixed live broadcast instruction to the live broadcast matrix module; if F=c, the live broadcast data decision module sends the digital human live broadcast instruction to the live broadcast matrix module.

[0070] Therefore, live broadcast decision tables for different time periods are stored inside the live broadcast data decision module, as shown in Table 1, which is a live broadcast decision table formed on a certain day.

[0071] Time period t1 t2 t3 t4 …… tn Live broadcast method a a b c a

[0072] Table 1

[0073] In addition, the system can customize and store live broadcast decision tables for multiple typical dates. For example, the 365 days of a year can be divided into three typical dates: weekdays, weekends, and holidays. The live broadcast decision tables for different typical dates will serve as the basis for the merchant's live broadcast method in each time period of the next typical date.

[0074] Step 5. The live broadcast matrix module receives instructions from the live broadcast data decision module. If a live broadcast instruction is received from a sales expert, it indicates that the time period is the prime time. In this case, all channels should arrange for live broadcasts by sales experts to maximize the conversion rate. If a mixed live broadcast instruction of real people / digital people is received, it indicates that the time period is a general time period. Some channels should be arranged for live broadcasts by real people (the number of channels is R), and the other channels should be arranged for live broadcasts by digital people (M*NR), taking into account both cost and benefit. If a live broadcast instruction of a digital person is received, it indicates that the time period is an idle time period. In this way, all channels should arrange for live broadcasts by digital people, saving the labor cost of the anchor while making full use of the idle traffic to continuously "plant grass".

[0075] Merchants arrange the specific live broadcasting methods of each live broadcasting platform (outside the system) according to the live broadcasting matrix module, which can be arranged automatically by the system or manually by the merchant. The live broadcasting data can be used as the system input for the next typical working day to continue to improve the data of the entire system.

[0076] It should be noted that the present invention is only a specific application case in the e-commerce field. The online live broadcast room data analysis decision system and method of the present invention are applicable to various new media platforms and various types of live broadcasts (entertainment live broadcast, education live broadcast, game live broadcast, etc.).

[0077] The above is only a preferred embodiment of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the structure of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. An online live broadcast room data analysis and decision-making system, It is characterized in that include: The live broadcast data input module is used to collect the live broadcast data of the live broadcast room from outside the system; A live data acquisition module, used for receiving the live data sent by the live data input module, and performing normalization processing on the live data to obtain normalized feature data; A live broadcast data analysis module, used to analyze and calculate the normalized feature data to obtain a comprehensive live broadcast score; A live broadcast intelligent decision module, used to make an intelligent decision on a specific live broadcast method according to the live broadcast comprehensive score, so as to select a target live broadcast method; The live broadcast matrix module is used to arrange the live broadcast on the new media platform according to the target live broadcast mode.

2. According to claim 1, an online live broadcast room data analysis and decision-making system, It is characterized in that The live broadcast data is normalized by the following formula: Among them, X is the characteristic data in the live broadcast data; X min and X max are the minimum and maximum values ​​of the feature data respectively; Y is the normalized feature data.

3. According to claim 1, an online live broadcast room data analysis and decision-making system, It is characterized in that The comprehensive score of the live broadcast is calculated by the following formula: S=(C*W c )+(G*W g )+(I*W i )+(D*W d ); Among them, S is the comprehensive score of the live broadcast; C is the normalized traffic feature data; G is the normalized product feature data; I is the normalized interaction feature data; D is the normalized transaction feature data; W c is the weight of the traffic feature data; W g is the weight of the product feature data; W i is the weight of the interactive feature data; W d is the weight of the product feature data.

4. According to claim 3, an online live broadcast room data analysis and decision-making system, It is characterized in that The weight W of the traffic feature data c , the weight of product feature data W g , the weight of the interactive feature data W i , the weight of product feature data W d The sum is equal to 1.

5. According to claim 1, an online live broadcast room data analysis and decision-making system, It is characterized in that The specific decision-making method of the live broadcast intelligent decision-making module is: Assume a time period, and set a scoring threshold range for all the live broadcast modes; obtain the cumulative duration of the comprehensive live broadcast score within each scoring threshold range within the time period, and take the live broadcast mode corresponding to the scoring threshold range with the longest cumulative duration as the target live broadcast mode.

6. According to claim 5, an online live broadcast room data analysis and decision-making system, It is characterized in that The live broadcast methods include live broadcast by sales experts, live broadcast by real people / digital people mixed, and live broadcast by digital people.

7. According to claim 1, an online live broadcast room data analysis and decision-making system, It is characterized in that The live broadcast data input module includes a traffic data channel, a commodity data channel, an interactive data channel, and a transaction data channel.

8. A data analysis and decision-making method for online live broadcast rooms. It is characterized in that The method is: Collecting live broadcast data of the live broadcast room from outside the system, normalizing the live broadcast data to obtain normalized feature data; analyzing and calculating the normalized feature data to obtain a comprehensive live broadcast score; An intelligent decision is made on a specific live broadcast method according to the comprehensive live broadcast score to select a target live broadcast method; and a live broadcast layout is performed on a new media platform according to the target live broadcast method.

9. According to claim 8, an online live broadcast room data analysis and decision-making system, It is characterized in that The live broadcast data is normalized by the following formula: Among them, X is the characteristic data in the live broadcast data; X min and X max are the minimum and maximum values ​​of the feature data respectively; Y is the normalized feature data.

10. The online live broadcast room data analysis and decision-making system according to claim 8, It is characterized in that The comprehensive score of the live broadcast is calculated by the following formula: S=(C*W c )+(G*W g )+(I*W i )+(D*W d ); Among them, S is the comprehensive score of the live broadcast; C is the normalized traffic feature data; G is the normalized product feature data; I is the normalized interaction feature data; D is the normalized transaction feature data; W c is the weight of the traffic feature data; W g is the weight of the product feature data; W i is the weight of the interactive feature data; W d is the weight of the product feature data.

Images