Data processing method, device, equipment and computer readable storage medium

Abstract

The present invention discloses a data processing method, device, equipment and computer-readable storage medium, wherein the method comprises: classifying target users based on historical viewing data of target users, and determining candidate time periods corresponding to each classification; determining the number of users who meet preset conditions in each classification corresponding to multiple start times in the candidate time period based on a preset live broadcast duration; determining the live broadcast revenue corresponding to each start time based on a decision weight factor corresponding to each classification and the number of users who meet the preset conditions, and determining the best start time in each start time based on the live broadcast revenue. The present invention can accurately determine the best start time based on historical viewing data by calculating the live broadcast revenue based on the number of users who meet the preset conditions in each classification corresponding to multiple start times and the decision weight factor, and then use the candidate time period with the largest live broadcast revenue as the optimal live broadcast time period, thereby increasing the host revenue within the live broadcast time period.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a data processing method, device, equipment and computer-readable storage medium. Background Art

[0002] Currently, in live video applications, anchor users can attract fan users on the live streaming platform to watch by turning on the live video function.

[0003] Currently, the broadcasting time of anchors is determined by historical experience data or the anchors' personal preferences. For example, operators conduct manual statistical analysis of historical live broadcast data and use simple percentage data as the conclusion to determine the broadcasting time. However, the dimension and strength of manual analysis are limited and there are certain subjective biases, which results in the failure to achieve the maximum profit within the broadcasting time, resulting in low live broadcasting revenue.

[0004] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of the present invention is to provide a data processing method, device, equipment and computer-readable storage medium, aiming to solve the technical problem that the existing method of determining the start time of broadcasting leads to low live broadcast efficiency.

[0006] To achieve the above object, the present invention provides a data processing method, which comprises the following steps:

[0007] Based on the historical viewing data of the target user, the target user is classified and the candidate time periods corresponding to the classifications are determined, wherein the historical viewing data includes the viewing time periods;

[0008] Based on the preset live broadcast duration, determine the number of users meeting the preset conditions in each category corresponding to the multiple broadcast start times in the candidate time period;

[0009] Based on the decision weight factors corresponding to each classification and the number of users who meet the preset conditions, the live broadcast revenue corresponding to each broadcast start time is determined, and based on the live broadcast revenue, the best broadcast start time is determined among the various broadcast start times.

[0010] Furthermore, the step of determining the candidate time periods corresponding to each category includes:

[0011] Based on the viewing time period of each user in the historical viewing data, obtaining a first viewing time period corresponding to each category;

[0012] The candidate time period is determined based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period.

[0013] Further, the candidate time period includes a first candidate time period and a second candidate time period, and the step of determining the candidate time period based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period includes:

[0014] Based on the first number of users, determine a first preset number of first candidate time periods corresponding to each category;

[0015] The second candidate time period is determined based on the second number of users.

[0016] Furthermore, the decision weight factor is obtained according to the following steps:

[0017] Based on the user value information of users in each category, determine the corresponding weight factor in each category;

[0018] Based on the weight factors of users in each category, the decision weight factors corresponding to each category are determined.

[0019] Furthermore, the step of determining the corresponding weight factor in each category based on the user value information of the users in each category includes:

[0020] Based on the number of users in each category, each category is subjected to dimensionality reduction processing to obtain each category after dimensionality reduction;

[0021] Based on the corresponding order of the user value information of the users in each category after dimensionality reduction, the weight factor corresponding to each category is determined.

[0022] Furthermore, the step of classifying the target user based on the historical viewing data of the target user includes:

[0023] Determine user value information corresponding to each target user based on the historical viewing data;

[0024] Clustering the user value information based on a first clustering algorithm to obtain the number of categories corresponding to the clustering result;

[0025] Based on the second clustering algorithm and the number of categories, the target users are clustered to obtain various categories.

[0026] Furthermore, the step of determining the number of users meeting preset conditions in each category corresponding to the plurality of broadcast start times in the candidate time period based on the preset live broadcast duration includes:

[0027] Based on the preset live broadcast duration, determine the number of third users in each category corresponding to multiple broadcast start times in the candidate time period;

[0028] The number of users meeting a preset condition in each category is determined based on the third number of users.

[0029] In addition, to achieve the above object, the present invention further provides a data processing device, the data processing device comprising:

[0030] A first determination module, configured to classify the target user based on the historical viewing data of the target user, and determine candidate time periods corresponding to each classification, wherein the historical viewing data includes viewing time periods;

[0031] A second determination module is used to determine the number of users meeting preset conditions in each category corresponding to multiple broadcast start times in the candidate time period based on a preset live broadcast duration;

[0032] The third determination module is used to determine the live broadcast revenue corresponding to each broadcast start time based on the decision weight factor corresponding to each classification and the number of users who meet the preset conditions, and determine the best broadcast start time among the various broadcast start times based on the live broadcast revenue.

[0033] In addition, to achieve the above-mentioned purpose, the present invention also provides a data processing device, which includes: a memory, a processor, and a data processing program stored in the memory and executable on the processor, and the data processing program implements the steps of the aforementioned data processing method when executed by the processor.

[0034] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a data processing program is stored, and when the data processing program is executed by a processor, the steps of the aforementioned data processing method are implemented.

[0035] The present invention classifies the target users based on their historical viewing data to determine candidate time periods corresponding to the classifications, wherein the historical viewing data includes viewing time periods; then, based on a preset live broadcast duration, determines the number of users who meet preset conditions in each classification corresponding to a plurality of broadcasting moments in the candidate time period; then, based on a decision weight factor corresponding to each classification and the number of users who meet the preset conditions, determines the live broadcast revenue corresponding to each broadcasting moment, and determines the best broadcasting moment in each broadcasting moment based on the live broadcast revenue; by calculating the live broadcast revenue according to the number of users who meet the preset conditions in each classification corresponding to a plurality of broadcasting moments and the decision weight factor, the best broadcasting moment can be accurately determined according to the historical viewing data, and then the candidate time period with the largest live broadcast revenue is used as the optimal time period for live broadcast, thereby increasing the revenue of the host within the live broadcast time period. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1It is a structural diagram of a data processing device in a hardware operating environment involved in an embodiment of the present invention;

[0037] Figure 2 It is a flowchart of a first embodiment of a data processing method of the present invention;

[0038] Figure 3 A schematic diagram of hierarchical viewing time periods according to an embodiment of a data processing method of the present invention;

[0039] Figure 4 It is a schematic diagram of functional modules of a data processing device according to an embodiment of the present invention.

[0040] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0041] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0042] like Figure 1 As shown, Figure 1 It is a structural diagram of a data processing device in a hardware operating environment involved in an embodiment of the present invention.

[0043] The data processing device in the embodiment of the present invention may be a PC. Figure 1 As shown, the data processing device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0044] Optionally, the data processing device may also include a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, etc. Of course, the data processing device may also be configured with other sensors such as a barometer, a hygrometer, a thermometer, an infrared sensor, etc., which will not be described in detail here.

[0045] Those skilled in the art will understand that Figure 1The terminal structure shown in the figure does not constitute a limitation on the data processing device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0046] like Figure 1 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a data processing program.

[0047] exist Figure 1 In the terminal shown, the network interface 1004 is mainly used to connect to the background server and communicate data with the background server; the user interface 1003 is mainly used to connect to the client (user end) and communicate data with the client; and the processor 1001 can be used to call the data processing program stored in the memory 1005.

[0048] In this embodiment, the data processing device includes: a memory 1005, a processor 1001, and a data processing program stored in the memory 1005 and executable on the processor 1001, wherein the processor 1001 calls the data processing program stored in the memory 1005 and executes the steps of the data processing method in each of the following embodiments.

[0049] The present invention also provides a data processing method, referring to Figure 2 , Figure 2 FIG. 1 is a flow chart of a first embodiment of a data processing method according to the present invention.

[0050] In this embodiment, the data processing method includes:

[0051] Step S101, classifying the target user based on the historical viewing data of the target user, and determining candidate time periods corresponding to each classification, wherein the historical viewing data includes viewing time periods;

[0052] In this embodiment, the target user is a user who has watched the live broadcast of a target anchor. The historical viewing data may be the viewing data of all target users within a preset period before the current moment. The historical viewing data includes the viewing time period, that is, the time period during which each target user watched the live broadcast of the target anchor. The historical viewing data may also include the number of comments, the length of comments, the number of gifts, etc.

[0053] The user value information corresponding to each target user is obtained based on the historical viewing data. Specifically, the historical viewing data includes the light viewing data within a month. The core fields of the target user Ui's live broadcast viewing event are A1, A2, A3, ..., An. The data of the target user Ui in the past month forms a two-dimensional array (matrix): [[a 01 ,a 02 ,a 03,...,a 0n ],[a 11 ,a 12 ,a 13 ,...,a 1n ],...,[a d1 ,a d2 ,a d3 ,...,a dn ]], where a dn It represents the value of the An field in the live broadcast watching event generated by the target user Ui on the dth day of a natural month. For example, A1, A2, ..., An include the number of comments, comment length, number of gifts, gift value, etc. Then calculate the value of the live broadcast watching event of the target user Ui in one day. The formula for the value of the dth day is:

[0054] V i d =x1*a d1 +x2*(1+a d2 ) 2 +x3*a d3 +...+x n *a dn ;

[0055] Among them, x1, x2, x n The value contributions of A1, A2, ..., An are respectively. For the end viewing parameter A2, its power level is 2. Determine the corresponding user value information V according to the value level of the target user Ui during the preset period ui , the specific formula is:

[0056]

[0057] V i It is the mean or sum of the value levels of the target user Ui. If the target user Ui watches the live broadcast continuously for more than the preset number of days, for example, more than 7 days, the value of the increase in user value is recorded as S1, where s is the continuous viewing time of the live broadcast - 7; if the target user Ui does not watch it halfway, the user value is reduced according to the number of consecutive days without watching, which is recorded as T1, where t is the number of consecutive days without watching.

[0058] Next, the target users are classified based on their historical viewing data. Specifically, the target users are classified according to user value information corresponding to each target user to obtain various classifications. Each cluster in the clustering result is regarded as a classification.

[0059] In this embodiment, based on the viewing time period of each user in the historical viewing data, the candidate time period corresponding to each category is determined. Specifically, based on the viewing time period of each user in the historical viewing data, the first viewing time period corresponding to each category is obtained. For each user value information in the category, the viewing data of the user corresponding to the user value information is obtained in the historical viewing data. A user portrait analysis is performed on the user based on the viewing data. The user portrait analysis includes the dimensions of viewing time and viewing time period. The first viewing time period corresponding to each category is obtained through the user portrait analysis of each user. Figure 3 The first viewing time period includes the live broadcast time and the user value at each moment in the corresponding category, and the user value is the sum of the user value information of all users in the category at the current moment.

[0060] Based on the number of viewers in the hierarchical viewing time period corresponding to each category, the first viewing time period corresponding to each category is determined. Specifically, since users in the same category belong to a group, but the entry and exit times of each group are different, the number of users currently remaining to watch is a fluctuating value, such as Figure 3 As shown, the most dense time period can be obtained by the point-by-point line drawing method. For example, the sampling interval is 10 minutes, and the historical live broadcast duration is 2 hours, then there are 12 points in total. When the target user Ui watches the live broadcast, there is a start and end time. The time point values ​​in this time period are accumulated, and the final 12 points with the largest median value are the most dense time points. Then, the preset time interval including the most dense time points is used as a first candidate time period, and the first preset number of first candidate time periods corresponding to the classification is obtained. The first preset number can be 2.

[0061] Finally, based on the number of viewers corresponding to the viewing time period, the second candidate time period with the largest number of viewers is determined, wherein the candidate time period includes the first candidate time period and the second candidate time period. Specifically, the second candidate time period is obtained in the same way as the first candidate time period. The second preset number may be 2, thereby obtaining 2*N+2 candidate time periods, where N is the number of categories.

[0062] Step S102, based on the preset live broadcast duration, determining the number of users meeting preset conditions in each category corresponding to the multiple broadcast start times in the candidate time period;

[0063] In this embodiment, when the candidate time period is obtained, the number of users who meet the preset conditions in each category corresponding to the multiple broadcasting times in the candidate time period is determined based on the preset live broadcast duration. Specifically, the number of third users in each category corresponding to the multiple broadcasting times in the candidate time period is first determined. Based on the preset live broadcast duration and each candidate time period, multiple broadcasting times are first determined, and the number of users in each category corresponding to each broadcasting time is obtained. For each category, the number of third users of historical users in the live broadcast time period corresponding to the broadcasting time is determined through user value information and the viewing time period of each user.

[0064] Then, based on the third number of users, the number of users who meet the preset conditions in each category is determined. Specifically, for each category, the number of triggered users who trigger the consideration point event in each category corresponding to each broadcast start time is obtained in the historical viewing data. For example, if the consideration point event is giving gifts, the number of triggered users is the number of users who give gifts. The number of triggered users is the number of users who meet the preset conditions. Finally, the number of users who meet the preset conditions in each category is obtained.

[0065] Step S103, based on the decision weight factors corresponding to the various categories and the number of users who meet the preset conditions, determine the live broadcast revenue corresponding to each broadcast start time, and determine the best broadcast start time among the various broadcast start times based on the live broadcast revenue.

[0066] In this embodiment, based on the decision weight factors corresponding to each category and the number of users who meet the preset conditions, the live broadcast revenue corresponding to each broadcast start time is determined. Specifically, for each broadcast start time, the formula for the corresponding live broadcast revenue is:

[0067]

[0068] Among them, Ei is the live broadcast revenue, N is the number of categories; Pk is the number of users in the current category; cik is the conversion ratio of the current category, and its value is between 0 and 1. For each category, its conversion ratio = the number of triggered users of the category (the number of users who meet the preset conditions) / the number of users of the category; KPk is the decision weight factor of the current category; bi is the offset (reward or punishment) of the consideration point, and bi is a preset value, which can be set reasonably.

[0069] Next, based on the live broadcast revenue, the best broadcast start time is determined among various broadcast start times. Specifically, the broadcast start time with the largest live broadcast revenue can be used as the best broadcast start time, or the broadcast start time with the largest live broadcast revenue per unit time can be used as the best broadcast start time.

[0070] It should be noted that after determining the best start time, you can also obtain the user stock and user increment resulting from live broadcast at the best start time, and then determine ADD and DEL. ADD represents the number of new viewers, and DEL represents the number of old viewers lost. The start time can be further adjusted through ADD and DEL.

[0071] The data processing method proposed in this embodiment classifies the target users based on their historical viewing data to determine candidate time periods corresponding to each category, wherein the historical viewing data includes viewing time periods; then, based on a preset live broadcast duration, determines the number of users who meet preset conditions in each category corresponding to multiple broadcasting moments in the candidate time period; then, based on a decision weight factor corresponding to each category and the number of users who meet the preset conditions, determines the live broadcast revenue corresponding to each broadcasting moment, and determines the best broadcasting moment in each broadcasting moment based on the live broadcast revenue; by calculating the live broadcast revenue based on the number of users who meet the preset conditions in each category corresponding to multiple broadcasting moments and the decision weight factor, the best broadcasting moment can be accurately determined based on the historical viewing data, and then the candidate time period with the largest live broadcast revenue is used as the optimal time period for live broadcast, thereby increasing the host's revenue within the live broadcast time period.

[0072] Based on the first embodiment, a second embodiment of the data processing method of the present invention is proposed. In this embodiment, step S101 includes:

[0073] Step S201, based on the viewing time period of each target user in the historical viewing data, obtaining a first viewing time period corresponding to each category;

[0074] Step S202: determining the candidate time period based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period.

[0075] In this embodiment, firstly, based on the viewing time period of each target user in the historical viewing data, the first viewing time period corresponding to each category is obtained. Specifically, for each user value information in the category, the user viewing data of the user corresponding to the user value information is obtained in the historical viewing data, and a user portrait analysis is performed on the user based on the user viewing data. The user portrait analysis includes the dimensions of viewing time and viewing time period. The first viewing time period corresponding to each category is obtained through the user portrait analysis of each user, and the first viewing time period corresponding to each category is obtained. Figure 3 The first viewing time period includes the live broadcast time and the user value at each moment in the corresponding category, and the user value is the sum of the user value information of all users in the category at the current moment.

[0076] Next, based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period, the candidate time period is determined. Specifically, the candidate time period includes a first candidate time period and a second candidate time period. Step S202 includes:

[0077] S2021, determining a first preset number of first candidate time periods corresponding to each category based on the first number of users;

[0078] S2022: Determine the second candidate time period based on the second number of users.

[0079] In this embodiment, based on the first number of users, a first preset number of first viewing time periods corresponding to each category is determined. Specifically, since users in the same category belong to a group, but the time when each user in the group enters and exits the group is different, the number of users currently remaining to watch is a fluctuating value, such as Figure 3 As shown, the most dense time period can be obtained by the point-by-point line drawing method. For example, the sampling interval is 10 minutes, and the historical live broadcast duration is 2 hours, then there are 12 points in total. When the target user Ui watches the live broadcast, there is a start and end time. The time point values ​​in this time period are accumulated, and the final 12 points with the largest median value are the most dense time points. Then, the preset time interval including the most dense time points is used as a first candidate time period, and the first preset number of first candidate time periods corresponding to the classification is obtained. The first preset number can be 2.

[0080] Next, based on the second number of users, the second candidate time period is determined. Specifically, the second candidate time period is obtained in the same manner as the first candidate time period. The second preset number may be 2, thereby obtaining 2*N+2 candidate time periods, where N is the number of categories, and the candidate time periods include the first candidate time period and the second candidate time period. In this embodiment, several time periods with the largest number of viewers are selected as candidate time periods to maximize the number of users performing live broadcasts in each candidate time period.

[0081] The data processing method proposed in this embodiment obtains the first viewing time period corresponding to each category based on the viewing time period of each user in the historical viewing data; then determines the candidate time period based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period, and accurately determines the candidate time period through the historical viewing data, thereby improving the accuracy of the candidate time period, thereby improving the accuracy of the optimal broadcast start time, and further improving the income of the anchor during the live broadcast time period.

[0082] Based on the first embodiment, a third embodiment of the data processing method of the present invention is proposed. In this embodiment, the decision weight factor is obtained according to the following steps:

[0083] Step S301, determining the corresponding weight factor in each category based on the user value information of the users in each category;

[0084] Step S302: determining a decision weight factor corresponding to each category based on the weight factors of users in each category.

[0085] In this embodiment, based on the user value information of users in each category, the corresponding weight factor in each category is determined. For example, for each category, the user value information of the target users in the category is sorted, and the weight factor corresponding to each target user in the category is determined according to the sorting of the user value information of the category. Of course, since there are a large number of target users in each category, in order to improve processing efficiency, each category can be first subjected to dimensionality reduction processing. Specifically, step S301 includes:

[0086] Step S3011, based on the number of users in each category, perform dimensionality reduction processing on each category to obtain each category after dimensionality reduction;

[0087] Step S3012, determining the weight factor corresponding to each category based on the corresponding order of the user value information of the users in each category after the dimensionality reduction.

[0088] In this embodiment, based on the number of users corresponding to the user value information in each category, the user value information in the category is first reduced in dimension to obtain each category after dimension reduction. For example, the category is divided into three layers with N being 3. The number of target users in each category is set to P1, P2, and P3, and the total number of users is P1+P2+P3. Then, geometric dimension reduction is performed to reduce the number to a multiple of 10. For example, if the total number of users is 1 million, the number can be directly reduced to 10 people. Thus, P1, P2, and P3 are also reduced in dimension. Assume that it is 6+2+2 people. The original P1 is 600,000 people, which becomes 6 people after dimension reduction.

[0089] Next, based on the order of the user value information of the users in each category after dimensionality reduction, the weight factor corresponding to each category is determined. Specifically, the users in each category after dimensionality reduction are sorted according to the value order of the user value information, and each entity in the sorting of each category is mapped to a score to obtain the weight factor corresponding to each entity in each category after dimensionality reduction. For example, a number (for example, 10 to 1) corresponds to each value from high to low, and the weight factors of P1 include 1, 2, 3, 4, 5, and 6. The weight factors of P2 include 7 and 8, and the weight factors of P3 include 9 and 10, thereby obtaining the weight factors corresponding to each category.

[0090] Finally, based on the weight factors of users in each category, the decision weight factors corresponding to each category are determined. Specifically, for each category, the average value of the weight factor corresponding to the category after dimensionality reduction is the corresponding decision weight factor, for example, P1 = (1 + 6) * 6 / 2 / 6 = 3.5, P2 = (7 + 8) / 2 = 7.5, P3 = (9 + 10) / 2 = 9.5. By reducing the dimensionality of the category and calculating the decision weight factor, the efficiency of determining the decision weight factor is improved.

[0091] The data processing method proposed in this embodiment determines the corresponding weight factors in each category based on the user value information of the users in each category; then determines the decision weight factors corresponding to each category based on the weight factors of the users in each category, and calculates the decision weight factors according to the user value information of the users in each category, thereby improving the accuracy of the decision weight factors and further improving the accuracy of the best broadcast start time.

[0092] Based on the first embodiment, a fourth embodiment of the data processing method of the present invention is proposed. In this embodiment, step S101 includes:

[0093] Step S401, determining user value information corresponding to each target user based on the historical viewing data;

[0094] Step S402, clustering the user value information based on a first clustering algorithm to obtain the number of categories corresponding to the clustering result;

[0095] Step S403: clustering the target users based on the second clustering algorithm and the number of categories to obtain various categories.

[0096] In this embodiment, the user value information corresponding to each target user is determined based on the historical viewing data. Specifically, the user value information corresponding to each user is determined based on the viewing parameters in the historical viewing data, wherein the viewing parameters include the number of comments, the length of comments, and the number of gifts. For example, the value of the user Ui's viewing of live events in one day is calculated first, and then the corresponding user value information is determined according to the value of the user Ui during the preset period.

[0097] Next, cluster the user value information based on the first clustering algorithm to obtain the number of categories corresponding to the clustering result. Specifically, the clustering process of the first clustering algorithm is as follows: the sample set (the set of user value information) is D = {x1, x2, ..., xn} with a total of n sample points, and the domain radius min_dr and the minimum number of units MinPts are set. The type number k = 1, and the unvisited sample set NF_D = D. First, traverse the sample set, and by calculating the distance (the distance between the currently traversed sample and other samples), 1 < = i, j < = n, and (i! = j), if the distance between xi (the currently traversed sample) and xj (other samples) is less than min_dr, then add xj to the neighborhood set of xi, and then obtain the neighborhood set N{xi} of xi. If the number of elements in the set N{xi} is not less than MinPts, then add xi to the core object set KeyX until all samples in the sample set are traversed. Then, a core object XO is selected from the core object set KeyX, and XO is added to the current cluster core object queue. The current cluster core object queue CurKey = {XO}, the current cluster sample set Ck = {XO}, k is the type serial number, at this time, the unvisited sample set NF_D = NF_D-{XO}, according to the neighborhood set N{xi}, determine the neighborhood set N{XO'} of XO, take the neighborhood set N{XO'} and the sample Ck as a cluster, remove the elements of N{XO'} and CK from the unvisited sample set to obtain a new unvisited sample set NF_D, and return to the step of selecting a core object XO from the core object set KeyX and adding XO to the current cluster core object queue. If the current cluster core object queue CurKey is empty and if the core object set KeyX is empty, the clustering is completed, and the cluster C = {C1, C2, ..., CK} is obtained. At this time, the number of categories K corresponding to the clustering result of cluster C is obtained.

[0098] Then, based on a second clustering algorithm K-means clustering algorithm and the number of categories, the target users are clustered to obtain a plurality of categories corresponding to the user value information, wherein the second clustering algorithm may be a K-means clustering algorithm.

[0099] The data processing method proposed in this embodiment determines the user value information corresponding to each target user based on the historical viewing data; then clusters the user value information based on a first clustering algorithm to obtain the number of categories corresponding to the clustering result; and then clusters the target users based on a second clustering algorithm and the number of categories to obtain each classification. This can avoid the problem of inaccurate clustering results caused by the instability of the clustering algorithm, improve the accuracy of the classification, and further improve the accuracy of the best broadcast start time.

[0100] Based on the above embodiments, a fifth embodiment of the data processing method of the present invention is proposed. In this embodiment, step S102 includes:

[0101] Step S501, based on a preset live broadcast duration, determining the number of third users in each category corresponding to a plurality of broadcast start times in the candidate time period;

[0102] Step S502: determining the number of users in each category that meet a preset condition based on the third number of users.

[0103] In this embodiment, when the candidate time period is obtained, the number of users who meet the preset conditions in each category corresponding to the multiple broadcasting times in the candidate time period is determined based on the preset live broadcast duration. Specifically, the number of third users in each category corresponding to the multiple broadcasting times in the candidate time period is first determined. Based on the preset live broadcast duration and each candidate time period, multiple broadcasting times are first determined, and the number of users in each category corresponding to each broadcasting time is obtained. For each category, the number of third users of historical users in the live broadcast time period corresponding to the broadcasting time is determined through user value information and the viewing time period of each user.

[0104] Then, based on the third number of users, the number of users who meet the preset conditions in each category is determined. Specifically, for each category, the number of triggered users who trigger the consideration point event in each category corresponding to each broadcast start time is obtained in the historical viewing data. For example, if the consideration point event is giving gifts, the number of triggered users is the number of users who give gifts. The number of triggered users is the number of users who meet the preset conditions. Finally, the number of users who meet the preset conditions in each category is obtained.

[0105] The data processing method proposed in this embodiment determines the number of third users in each category corresponding to multiple broadcasting start times in the candidate time period based on the preset live broadcast duration, and then determines the number of users in each category that meet the preset conditions based on the number of third users. The method can accurately obtain the number of users in each category that meet the preset conditions based on the number of users in each category, thereby improving the accuracy of the optimal broadcasting start time and further increasing the income of the anchor in the live broadcast time period.

[0106] The present invention also provides a data processing device, referring to Figure 4 , the data processing device comprises:

[0107] A first determination module 10 is used to classify the target user based on the historical viewing data of the target user, and determine the candidate time periods corresponding to the respective classifications, wherein the historical viewing data includes the viewing time periods;

[0108] The second determination module 20 is used to determine the number of users meeting the preset conditions in each category corresponding to the multiple broadcast start times in the candidate time period based on the preset live broadcast duration;

[0109] The third determination module 30 is used to determine the live broadcast revenue corresponding to each broadcast start time based on the decision weight factor corresponding to each classification and the number of users who meet the preset conditions, and determine the best broadcast start time among the various broadcast start times based on the live broadcast revenue.

[0110] Optionally, the first determining module 10 is further configured to:

[0111] Based on the viewing time period of each target user in the historical viewing data, obtaining a first viewing time period corresponding to each category;

[0112] The candidate time period is determined based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period.

[0113] Optionally, the first determining module 10 is further configured to:

[0114] Based on the first number of users, determine a first preset number of first candidate time periods corresponding to each category;

[0115] The second candidate time period is determined based on the second number of users.

[0116] Optionally, the data processing device further includes:

[0117] Based on the user value information of users in each category, determine the corresponding weight factor in each category;

[0118] Based on the weight factors of users in each category, the decision weight factors corresponding to each category are determined.

[0119] Optionally, the data processing device further includes:

[0120] Based on the number of users in each category, each category is subjected to dimensionality reduction processing to obtain each category after dimensionality reduction;

[0121] Based on the corresponding order of the user value information of the users in each category after dimensionality reduction, the weight factor corresponding to each category is determined.

[0122] Optionally, the first determining module 10 is further configured to:

[0123] Determine user value information corresponding to each target user based on the historical viewing data;

[0124] Clustering the user value information based on a first clustering algorithm to obtain the number of categories corresponding to the clustering result;

[0125] Based on the second clustering algorithm and the number of categories, the target users are clustered to obtain various categories.

[0126] Optionally, the second determining module 20 is further configured to:

[0127] Based on the preset live broadcast duration, determine the number of third users in each category corresponding to multiple broadcast start times in the candidate time period;

[0128] The number of users meeting a preset condition in each category is determined based on the third number of users.

[0129] The methods executed by the above-mentioned program units can refer to the various embodiments of the data processing method of the present invention, and will not be described in detail here.

[0130] The present invention also provides a computer-readable storage medium.

[0131] The computer-readable storage medium of the present invention stores a data processing program, and when the data processing program is executed by a processor, the steps of the data processing method described above are implemented.

[0132] Among them, the method implemented when the data processing program running on the processor is executed can refer to the various embodiments of the data processing method of the present invention, and will not be repeated here.

[0133] In addition, an embodiment of the present invention further provides a computer program product, which includes a data processing program. When the data processing program is executed by a processor, the steps of the data processing method described above are implemented.

[0134] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0135] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0136] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0137] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A data processing method, characterized in that: The data processing method comprises the following steps: Based on the historical viewing data of the target user, the target user is classified to determine the candidate time periods corresponding to each classification, wherein the historical viewing data includes the viewing time period, wherein the value degree of each target user corresponding to each day is determined based on the live broadcast viewing event corresponding to each target user in the historical viewing data, and the user value information corresponding to each target user is determined based on the value degree and the continuous live broadcast viewing time corresponding to each target user in the historical viewing data; the target user is classified based on the user value information to obtain each classification, and the formula for the value degree of the dth day is: V i d =x1*a d1 +x2*(1+a d2 ) 2 +x3*a d3 +...+x n *a dn ; User value information V Ui The formula is: Among them, x1, x2, x n are the value contributions of A1, A2, ..., An respectively. The core fields of the target user Ui’s live broadcast watching event are A1, A2, A3, ..., An, a dn is the value of the An field in the live broadcast watching event generated by the target user Ui on the dth day of a natural month, V i is the mean or sum of the value of the target user Ui. If the target user Ui watches the live broadcast continuously for the preset number of days, the value of the user value increase is recorded as S1, where s is the continuous viewing time - 7; if the target user Ui does not watch the live broadcast halfway, the user value is reduced according to the number of consecutive days without watching, which is recorded as T1, where t is the number of consecutive days without watching. Based on the preset live broadcast duration, determine the number of users meeting the preset conditions in each category corresponding to the multiple broadcast start times in the candidate time period; Based on the decision weight factors corresponding to each classification and the number of users who meet the preset conditions, the live broadcast revenue corresponding to each broadcast start time is determined, and the best broadcast start time is determined among the broadcast start times based on the live broadcast revenue, wherein the broadcast start time with the largest live broadcast revenue is taken as the best broadcast start time, or the broadcast start time with the largest live broadcast revenue per unit time is taken as the best broadcast start time. The formula for live broadcast revenue is: Among them, Ei is the live broadcast income, N is the number of categories; P k is the number of users in the current category; c ik KP is the conversion ratio of the current category. For each category, the conversion ratio = the number of users who meet the preset conditions / the number of users in the category; k is the decision weight factor of the current classification; bi is the offset of the consideration point, and bi is the preset value.

2. The data processing method according to claim 1, characterized in that: The step of determining the candidate time periods corresponding to each category includes: Based on the viewing time period of each target user in the historical viewing data, obtaining a first viewing time period corresponding to each category; The candidate time period is determined based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period.

3. The data processing method according to claim 2, characterized in that: The candidate time periods include a first candidate time period and a second candidate time period, and the step of determining the candidate time periods based on the number of first users corresponding to the first viewing time period and the number of second users corresponding to the viewing time period includes: Based on the first number of users, determine a first preset number of first candidate time periods corresponding to each category; The second candidate time period is determined based on the second number of users.

4. The data processing method according to claim 1, characterized in that: The decision weight factor is obtained according to the following steps: Based on the user value information of users in each category, determine the corresponding weight factor in each category; Based on the weight factors of users in each category, the decision weight factors corresponding to each category are determined.

5. The data processing method according to claim 4, characterized in that: The step of determining the corresponding weight factor in each category based on the user value information of the users in each category includes: Based on the number of users in each category, each category is subjected to dimensionality reduction processing to obtain each category after dimensionality reduction; Based on the corresponding order of the user value information of the users in each category after dimensionality reduction, the weight factor corresponding to each category is determined.

6. The data processing method according to claim 1, characterized in that: The step of classifying the target users based on the user value information and obtaining each classification comprises: Clustering the user value information based on a first clustering algorithm to obtain the number of categories corresponding to the clustering result; Based on the second clustering algorithm and the number of categories, the target users are clustered to obtain various categories.

7. The data processing method according to any one of claims 1 to 6, characterized in that: The step of determining the number of users meeting preset conditions in each category corresponding to a plurality of broadcast start times in the candidate time period based on the preset live broadcast duration includes: Based on the preset live broadcast duration, determine the number of third users in each category corresponding to multiple broadcast start times in the candidate time period; The number of users meeting a preset condition in each category is determined based on the third number of users.

8. A data processing device, characterized in that: The data processing device comprises: The first determination module is used to classify the target users based on the historical viewing data of the target users, and determine the candidate time periods corresponding to the classifications, wherein the historical viewing data includes viewing time periods, wherein the value levels of the target users corresponding to the days are determined based on the live broadcast viewing events corresponding to the target users in the historical viewing data, and the user value information corresponding to the target users is determined based on the value levels and the continuous live broadcast viewing time corresponding to the target users in the historical viewing data; the target users are classified based on the user value information to obtain the value levels of the classifications, and the formula for the value level of the dth day is: V i d =x1*a d1 +x2*(1+a d2 ) 2 +x3*a d3 +...+x n *a dn ; User value information V ui The formula is: Among them, x1, x2, x n are the value contributions of A1, A2, ..., An respectively. The core fields of the target user Ui’s live broadcast watching event are A1, A2, A3, ..., An, a dn is the value of the An field in the live broadcast watching event generated by the target user Ui on the dth day of a natural month, V i is the mean or sum of the value of the target user Ui. If the target user Ui watches the live broadcast continuously for the preset number of days, the value of the user value increase is recorded as S1, where s is the continuous viewing time - 7; if the target user Ui does not watch the live broadcast halfway, the user value is reduced according to the number of consecutive days without watching, which is recorded as T1, where t is the number of consecutive days without watching. A second determination module is used to determine the number of users meeting preset conditions in each category corresponding to multiple broadcast start times in the candidate time period based on a preset live broadcast duration; The third determination module is used to determine the live broadcast revenue corresponding to each broadcast start time based on the decision weight factor corresponding to each classification and the number of users who meet the preset conditions, and determine the best broadcast start time among the broadcast start times based on the live broadcast revenue, wherein the broadcast start time with the largest live broadcast revenue is taken as the best broadcast start time, or the broadcast start time with the largest live broadcast revenue per unit time is taken as the best broadcast start time, and the formula for the live broadcast revenue is: Among them, Ei is the live broadcast income, N is the number of categories; P k is the number of users in the current category; c ik KP is the conversion ratio of the current category. For each category, the conversion ratio = the number of users who meet the preset conditions / the number of users in the category; k is the decision weight factor of the current classification; bi is the offset of the consideration point, and bi is the preset value.

9. A data processing device, characterized in that: The data processing device comprises: a memory, a processor, and a data processing program stored in the memory and executable on the processor, wherein the data processing program implements the steps of the data processing method according to any one of claims 1 to 7 when executed by the processor.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a data processing program, and when the data processing program is executed by a processor, the steps of the data processing method according to any one of claims 1 to 7 are implemented.

Images