Cross-platform television program tracking system

Abstract

The invention relates to the technical field of program tracking, in particular to a cross-platform television program tracking system, which comprises an equipment prediction module for collecting equipment use data of a user, constructing a Bayesian network prediction model, analyzing a switching mode of the user among various equipment, and combining a watching progress and a content type of a television program, and generating an equipment switching prediction result through statistical analysis and prediction of the time for the user to replace the equipment. According to the method and the device, the switching mode of the user among various devices is analyzed and combined with the watching progress and the content type of the television program, so that the user experience is optimized, and the seamless switching capability of program watching is enhanced. Besides, environmental factors such as position information, equipment electric quantity and network state are collected and analyzed in real time, so that the system can adjust the prediction model according to the key environmental factors, and more accurate video cache adjustment is realized in evaluation of network speed and current buffer size.

Description

Technical Field

[0001] The present invention relates to the technical field of program tracking, and in particular to a cross-platform television program tracking system. Background Art

[0002] The field of program tracking technology involves the development and implementation of systems and software for monitoring, recording, and analyzing television program viewing behavior. This is done by tracking users' viewing habits on different devices and analyzing this data to identify viewing trends. This includes identifying when a user starts and ends watching a program, and whether the entire program has been watched. The system usually needs to handle a large amount of data input and be able to provide accurate tracking and synchronization services across multiple platforms and devices.

[0003] Among them, the cross-platform TV program tracking system is used to manage and maintain the consistency of users' TV program viewing progress on multiple playback devices. The system allows users to start watching a program on one device and continue watching it on another device without manually searching for the program and adjusting to the correct playback position. Cloud technology or user account information is used to synchronize viewing progress across platforms to ensure that no matter which device the user switches to, they can automatically return to the place where they stopped watching last time. This system theme is designed to improve the consistency and convenience of the user experience, while collecting cross-platform viewing data for content providers.

[0004] In the prior art, although it is possible to synchronize the viewing progress of users on different devices, it mainly focuses on basic progress synchronization, and it is difficult to deeply analyze and predict user behavior. As a result, when users frequently change viewing devices or watch in different network environments, it is impossible to effectively adapt to the actual needs of users. For example, when a user switches from a high-speed network environment at home to a mobile network environment used when going out, it may be difficult for the system to adjust the buffering strategy in time. In this case, users may experience frequent buffering and playback interruptions, affecting their viewing experience. In addition, it is difficult for the system to predict the user's device switching pattern. For example, the habit of switching from a mobile phone to a TV to watch at night may cause the user to repeatedly search for programs and adjust to the appropriate viewing point, increasing the complexity and time consumption of the operation. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a cross-platform television program tracking system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A cross-platform TV program tracking system comprises:

[0007] The device prediction module collects user device usage data, builds a Bayesian network prediction model, analyzes the user's switching pattern between multiple devices, combines the viewing progress and content type of TV programs, predicts the time when users change devices through statistical analysis, and generates device switching prediction results;

[0008] The environmental data analysis module collects environmental factor data associated with the user device, including user location information, device power status, and device network status, identifies key environmental factors that affect the user's device selection based on the device switching prediction results, adjusts the Bayesian network prediction model, and generates environmental factor analysis results;

[0009] The buffer management module evaluates the network speed and the current buffer size through the Markov decision process according to the environmental factor analysis results, dynamically adjusts the video cache in combination with the current network status, and generates a buffer parameter optimization result;

[0010] The real-time adjustment module monitors the viewing progress of the user on multiple devices based on the buffer parameter optimization result, adjusts the data synchronization in real time, and generates the TV program data synchronization adjustment result.

[0011] As a further solution of the present invention, the step of analyzing the acquisition mode of the user switching between multiple devices is specifically:

[0012] Collect the user's device usage data and mark each device usage data as an element of array D, that is, D = [s1, s2, ..., s n ], using the formula:

[0013]

[0014] Calculate the count C of device status S in the data set S ;

[0015] Among them, s i represents the device status of the user at the i-th moment, C S is the count of the device state S, I is the indicative function, if the condition s i =S holds true, then I(s i =S) = 1, otherwise 0, n is the total number of data points in the device usage data;

[0016] According to the count C of the device state S in the data set S , using the formula:

[0017]

[0018] Calculate the prior probability P(S) of the device state S;

[0019] According to the prior probability P(S) of the device state S, the formula is adopted:

[0020]

[0021] Calculate the conditional probability P(S|D) of observing data set D under device state S;

[0022] Among them, S' is the overall device status.

[0023] As a further solution of the present invention, the step of obtaining the device switching prediction result is specifically:

[0024] Integrate the viewing progress v of the TV program and the content type b into an array V = [(v1, b1), (v2, b2), ..., (v m ,b m )], according to the conditional probability P(S|D), the formula is adopted:

[0025] R j =P(S|D)×v j ×f(b j )

[0026] Calculate the probability R of a user switching devices j ;

[0027] Among them, f(b j ) is the content type b j The impact function on the probability of switching devices, v j is the viewing progress of the jth data point, b j is the content type of the jth data point;

[0028] According to the probability R of the user switching devices j , combined with the user history switching data set H, that is, H = [h1,h2,…,h m ], using the formula:

[0029]

[0030] Calculate the accuracy index A of the prediction of the user's device switching time;

[0031] Among them, h j is the device switching situation at the jth time point, and m is the total number of historical switching data points of the user;

[0032] According to the accuracy index A of the prediction of the user switching device time, the formula is adopted:

[0033] T=max({R j |A>α1})

[0034] Calculate the switching time T and obtain the device switching prediction result;

[0035] Among them, α1 is the threshold for evaluating accuracy, and max is used to select the maximum value in the set.

[0036] As a further solution of the present invention, the step of obtaining the key environmental factors that affect the user equipment selection is specifically:

[0037] Collect environmental data, including user location information U, that is, U = [u1,u2,...,u x ], the device power state Q, that is, Q = [q1, q2, ..., q x ] and the device network status W, that is, W = [w1,w2,...,w x ], according to the switching time T, the formula is adopted:

[0038] G k =t k ×(v u ·u T +v q ·q T +v w ·w T )

[0039] Calculate the environmental impact value G at each time point k ;

[0040] Among them, G k is the environmental impact value at the kth time point, t k is the predicted probability of device switching corresponding to the switching time T, v u 、v q 、v w They are the weight factors of user location information, power status and network status. According to historical data analysis, u T ,q T 、w T are the values ​​of user location information, power status, and network status at time point T, respectively. k is used to iterate the environment data array;

[0041] According to the environmental impact value G at each time point k , using the formula:

[0042]

[0043] and

[0044]

[0045] and

[0046]

[0047] Calculate the total influence G of each position u , the total influence of electricity G q and the total influence of the network G w , x is the total number of environmental data points;

[0048] According to the total influence G of the position u , the total influence of electricity G q and the total network influence G w , using the formula:

[0049]

[0050] Calculate the impact score Z of each environmental factor;

[0051] Among them, x represents the u , G q , G w Each value selected in .

[0052] As a further solution of the present invention, the step of obtaining the environmental factor analysis result is specifically:

[0053] According to the impact score Z of each environmental factor, the formula is used:

[0054] P new (S) = P (S) + γ × Z

[0055] Adjust the Bayesian network prediction model to reflect the influence of environmental factors and obtain the prior probability P of the adjusted device state S new (S);

[0056] Among them, γ is the adjustment coefficient used to quantify the impact of environmental factors;

[0057] According to the prior probability P of the adjusted device state S new (S), using the formula:

[0058]

[0059] Calculate the conditional probability P of the new device state S new (S|D) to obtain the environmental factor analysis results.

[0060] As a further solution of the present invention, the step of obtaining the estimated network speed and the current buffer size is specifically as follows:

[0061] According to the conditional probability P of the new device state S new (S|D), using the formula:

[0062]

[0063] Calculate the expected value F of the network speed;

[0064] Among them, w k is the speed or quality indicator of the kth network state, k is used to iterate the environment data array, and x is the total number of environment data points;

[0065] According to the expected value F of the network speed, the formula is adopted:

[0066] O=F×δ1

[0067] Calculate the current buffer size O;

[0068] Where δ1 is the buffer size adjustment factor.

[0069] As a further solution of the present invention, the step of obtaining the buffer parameter optimization result is specifically:

[0070] According to the current buffer size O, the formula is adopted:

[0071] Q'=O×(1+λ1·F)

[0072] Calculate the adjusted video buffer size Q';

[0073] Among them, λ1 is the weight coefficient used to adjust the buffer size;

[0074] According to the adjusted video cache size Q', the formula is adopted:

[0075] R'=Q'×η1

[0076] Calculate the optimal buffer parameter R' and obtain the buffer parameter optimization result;

[0077] Among them, η1 is the performance evaluation coefficient.

[0078] As a further solution of the present invention, the step of obtaining the result of synchronization adjustment of the television program data is specifically as follows:

[0079] According to the optimal buffer parameter R', the formula is adopted:

[0080]

[0081] Calculate the user's total viewing progress P on multiple devices track ;

[0082] Among them, E a is the current viewing progress of the ath device, a represents the device number, and g is the total number of devices used by the user;

[0083] According to the total viewing progress P of the user on multiple devicestrack , using the formula:

[0084] S sync =α2×P track

[0085] Calculate the synchronization adjustment S sync , obtain the TV program data synchronization adjustment result;

[0086] Among them, α2 is the synchronous efficiency coefficient.

[0087] Compared with the prior art, the advantages and positive effects of the present invention are:

[0088] In the present invention, by analyzing the user's switching mode between multiple devices and combining it with the viewing progress and content type of TV programs, not only the user experience is optimized, but also the seamless switching capability of program viewing is enhanced. In addition, the real-time collection and analysis of environmental factors such as location information, device power and network status enable the system to adjust the prediction model according to these key environmental factors, thereby achieving more accurate video cache adjustment in the network speed and current buffer size evaluation. It effectively reduces the buffering and loading time, improves the possibility of smooth viewing, and can still maintain a good viewing experience, especially when the network conditions are poor. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 is a system flow chart of the present invention;

[0090] Figure 2 A flowchart of the steps for analyzing the switching mode of a user between multiple devices in the present invention;

[0091] Figure 3 A flow chart of the steps for obtaining the prediction results of the device switching of the present invention;

[0092] Figure 4 A flow chart of the acquisition steps for identifying key environmental factors that affect user device selection for the present invention;

[0093] Figure 5 A flowchart of the steps for obtaining the environmental factor analysis results of the present invention;

[0094] Figure 6 A flow chart of the acquisition steps for evaluating the network speed and the current buffer size of the present invention;

[0095] Figure 7 A flow chart of the steps for obtaining the buffer parameter optimization results of the present invention;

[0096] Figure 8 The present invention is a flowchart of the steps for obtaining the result of synchronous adjustment of television program data. DETAILED DESCRIPTION

[0097] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0098] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0099] See also Figure 1 , a cross-platform TV program tracking system includes:

[0100] The device prediction module collects user device usage data, builds a Bayesian network prediction model, analyzes the user's switching pattern between multiple devices, combines the viewing progress and content type of TV programs, predicts the time when users change devices through statistical analysis, and generates device switching prediction results;

[0101] The environmental data analysis module collects environmental factor data associated with the user's device, including user location information, device power status, and device network status. Based on the device switching prediction results, it identifies key environmental factors that affect the user's device selection, adjusts the Bayesian network prediction model, and generates environmental factor analysis results.

[0102] The buffer management module evaluates the network speed and current buffer size through the Markov decision process according to the results of environmental factor analysis, dynamically adjusts the video cache in combination with the current network conditions, and generates buffer parameter optimization results;

[0103] The real-time adjustment module monitors the viewing progress of users on multiple devices based on the buffer parameter optimization results, adjusts data synchronization in real time, and generates TV program data synchronization adjustment results;

[0104] The device switching prediction results include the recommended switching device, the estimated switching time point and the predicted viewing interruption time. The environmental factor analysis results include the user's current geographic location, ambient light conditions and network connection speed. The buffer parameter optimization results include the buffer size and data packet transmission speed adjusted according to network delay. The TV program data synchronization adjustment results include the updated user viewing progress, the timestamp of the synchronized data and the cross-device viewing status record.

[0105] See also Figure 2 ,The specific steps for analyzing the switching mode of users between multiple devices are:

[0106] Collect the user's device usage data and mark each device usage data as an element of array D, that is, D = [s1, s2, ..., s n ], using the formula:

[0107]

[0108] Calculate the count C of device status S in the data set S ;

[0109] Among them, s i is the i-th element in the array, indicating the device status of the user at the i-th moment, C S is the count of the device state S, I is the indicative function, if the condition s i =S holds true, then I(s i =S) = 1, otherwise 0, n is the total number of data points in the device usage data;

[0110] According to the count C of the device status S in the data set S , using the formula:

[0111]

[0112] Calculate the prior probability P(S) of the device state S;

[0113] According to the prior probability P(S) of the device state S, the formula is used:

[0114]

[0115] Calculate the conditional probability P(S|D) of observing data set D under device state S;

[0116] Among them, S′ is the total device state, which is used to calculate the denominator of probability normalization;

[0117] The calculation process is as follows:

[0118] The user's device usage logs are extracted from the system database. These logs record the status of the devices used by the user at different time points.

[0119] For the formula:

[0120]

[0121] Set the parameter D = [M′, T′, M′, P′, T′, T′], where M′ represents the mobile phone, T′ represents the TV, and P′ represents the tablet;

[0122] C M′ =I(M′=M′)+I(T′=M′)+I(M′=M′)+I(P′=M′)+I(T′=M′)+I(T′=M′

[0123] =1+0+1+0+0+0=2

[0124] C T′ =3

[0125] C P′ =1

[0126] Set parameter n=6;

[0127] For the formula:

[0128]

[0129] For the formula:

[0130]

[0131] Calculate P(M′|D):

[0132]

[0133] Calculate P(T′|D):

[0134]

[0135] Calculate P(P′|D):

[0136]

[0137] Results P(M′|D)≈0.5: A high probability user will use a mobile phone. P(T′|D)≈0.1875: A low probability user will use a TV. P(P′|D)≈0.313: A medium probability user will use a tablet.

[0138] See also Figure 3 , the specific steps for obtaining the device switching prediction results are:

[0139] Integrate the viewing progress v of the TV program and the content type b into an array V = [(v1, b1), (v2, b2), ..., (v m ,b m )], according to the conditional probability P(S|D), the formula is:

[0140] R j =P(S|D)×v j ×f(b j )

[0141] Calculate the probability R of a user switching devices j ;

[0142] Among them, f(b j ) is the content type b j The impact function on the probability of switching devices, v j is the viewing progress of the jth data point, b j is the content type of the jth data point;

[0143] According to the probability R of users switching devices j , combined with the user history switching data set H, that is, H = [h1,h2,…,h m ], using the formula:

[0144]

[0145] Calculate the accuracy index A of the prediction of the user's device switching time;

[0146] Among them, h j is the device switching situation at the jth time point, m is the total number of historical switching data points of the user, which is used to traverse historical data;

[0147] According to the accuracy index A of the prediction of the user's device switching time, the formula is:

[0148] T=max({R j |A>α1})

[0149] Calculate the switching time T and obtain the device switching prediction result;

[0150] Among them, α1 is the threshold for evaluating accuracy, and max is used to select the maximum value in the set;

[0151] The calculation process is as follows:

[0152] v j is a value from 0 to 1 indicating the percentage of views that have not been started but have been completed. j It can be "Drama", "Comedy", "Documentary", etc.

[0153] Calculate when the user is using a mobile phone and set the parameter f(b j ) is f("plot") = 1.2 (the probability of switching devices when watching plot content is 1.2 times the baseline probability), P(M|D) ≈ 0.522, v j =0.75 (75% viewed);

[0154] For the formula:

[0155] R j =P(S|D)×vj ×f(b j )

[0156] but:

[0157] R j =P(M|D)×v j ×f(b j )

[0158] R j =0.522×0.75×1.2≈0.471

[0159] For the formula:

[0160]

[0161] Set parameter α1 = 0.8;

[0162] For the formula:

[0163] T=max({R j |A>α1})

[0164] T=max({R j})(because A=1.0>α1)

[0165] T=max({0.471,0.471,0.471})=0.471

[0166] The result T=0.471 indicates that the user is most likely to switch devices when the viewing progress reaches about 47%.

[0167] See also Figure 4 ,The steps for identifying the key environmental factors that affect the user’s device selection are as follows:

[0168] Collect environmental data, including user location information U, that is, U = [u1,u2,...,u x ], the device power state Q, that is, Q = [q1, q2, ..., q x ] and the device network status W, that is, W = [w1,w2,...,w x ], according to the switching time T, the formula is:

[0169] G k =t k ×(v u ·u T +v q ·q T +v w ·w T )

[0170] Calculate the environmental impact value G at each time point k;

[0171] Among them, G k is the environmental impact value at the kth time point, which is used to evaluate how environmental factors at a given time point affect the device switching probability, t k is the predicted probability of device switching corresponding to the switching time T, v u 、v q 、v w They are the weight factors of user location information, power status, and network status, which are obtained based on historical data analysis and reflect the influence of different environmental factors on the probability of device switching. T ,q T 、w T are the values ​​of user location information, power status, and network status at time point T, respectively. k is used to iterate the environment data array;

[0172] According to the environmental impact value G at each time point k , using the formula:

[0173]

[0174] and

[0175]

[0176] and

[0177]

[0178] Calculate the total influence G of each position u , the total influence of electricity G q and the total network influence G w , x is the total number of environmental data points;

[0179] According to the total influence G of the position u , the total influence of electricity G q and the total network influence G w , using the formula:

[0180]

[0181] Calculate the impact score Z of each environmental factor;

[0182] Among them, x represents the u , G q , G w Each value selected in;

[0183] The calculation process is as follows:

[0184] For the formula:

[0185] Gk =t k ×(v u ·u T +v q ·q T +v w ·w T )

[0186] Set the parameters T = 0.471, v u =0.3, v q =0.5, v w =0.2, u T =1,q T =0.2, w T =0.8, t T =0.471, where k = T;

[0187] G T =0.471×(0.3·1+0.5·0.2+0.2·0.8)=0.471×(0.3+0.1+0.16)=0.471×0.56

[0188] =0.264

[0189] For the formula:

[0190]

[0191] and

[0192]

[0193] and

[0194]

[0195] Set the parameter x = 5, and each u k ,q k 、w k Data is collected (for example: k =[1,0.9,0.8,1,0.95], etc.);

[0196] G u =0.3×(1+0.9+0.8+1+0.95)=0.3×4.65=1.395

[0197] G q =0.5×(0.2+0.3+0.4+0.1+0.2)=0.5×1.2=0.6

[0198] G w =0.2×(0.8+0.85+0.75+0.8+0.9)=0.2×4.1=0.82

[0199] The most influential factor: max(1.395,0.6,0.82)=1.395 (position information)

[0200] For the formula:

[0201]

[0202] Explanation of the results The calculated Z = 1.726 represents the environmental impact score after comprehensively considering the location, power and network status. The baseline value Z = 1.0 is set as the limit of the degree of influence of environmental factors on the probability of device switching. The calculated comprehensive impact score Z = 1.726, which exceeds the baseline value of 1.0, indicates that at the considered time point T, the combined impact of location information, power status and network status on user device switching is significant. This means that these environmental factors are strong enough to prompt users to switch devices at this time, and should be paid close attention to when optimizing the system or predicting user behavior.

[0203] See also Figure 5 , the specific steps for obtaining the environmental factor analysis results are:

[0204] According to the impact score Z of each environmental factor, the formula is used:

[0205] P new (S) = P (S) + γ × Z

[0206] Adjust the Bayesian network prediction model to reflect the influence of environmental factors and obtain the prior probability P of the adjusted device state S new (S);

[0207] Among them, γ is the adjustment coefficient used to quantify the impact of environmental factors;

[0208] According to the prior probability P of the adjusted device state S new (S), using the formula:

[0209]

[0210] Calculate the conditional probability P of the new device state S new (S|D), get the environmental factor analysis results;

[0211] The calculation process is as follows:

[0212] Original conditional probability: P(M′|D)≈0.5, P(T′|D)≈0.1875, P(P′|D)≈0.313

[0213] Set the parameters γ = 0.05, Z = 1.726;

[0214] For the formula:

[0215] P new (S) = P (S) + γ × Z

[0216] Mobile phone M':

[0217] P new (M′)=P(M′)+γ×ZP new (M′)=0.5+0.05×1.726=0.5863

[0218] TV T':

[0219] P new (T′)=P(T′)+γ×ZP new (T′)=0.1875+0.05×1.726=0.2741

[0220] Plate P′:

[0221] P new (P′)=P(P′)+γ×ZP new (P′)=0.313+0.05×1.726=0.3996

[0222] For the formula:

[0223]

[0224] Conditional probability calculation:

[0225] Mobile phone M':

[0226]

[0227] TV T':

[0228]

[0229] Plate P′:

[0230]

[0231] Results Explanation new (M′|D)≈0.3927: After adjustment, the probability of users using mobile phones is slightly reduced. new (T′|D)≈0.4274: After adjustment, the probability of users using TV increases significantly. new (P′|D)≈0.1786: After adjustment, the probability of users using tablets decreases slightly.

[0232] See also Figure 6 , the steps to evaluate the network speed and the current buffer size are as follows:

[0233] According to the conditional probability P of the new device state S new (S|D), using the formula:

[0234]

[0235] Calculate the expected value F of the network speed;

[0236] Among them, w k is the speed or quality indicator of the kth network state, k is used to iterate the environment data array, and x is the total number of environment data points;

[0237] According to the expected value F of the network speed, the formula is used:

[0238] O=F×δ1

[0239] Calculate the current buffer size O;

[0240] Among them, δ1 is the buffer size adjustment coefficient, which depends on the network speed;

[0241] The calculation process is as follows:

[0242] Calculation is performed when the user is using a mobile phone, for the formula:

[0243]

[0244] but:

[0245]

[0246] Set the parameters W = [100,50,150] Mbps, x = 3, P new {M′|D)≈0.3927;

[0247]

[0248] F=39.27+19.635+58.905=117.81Mbps

[0249] For the formula:

[0250] O = F × δ1;

[0251] Set parameter δ1 to 0.1;

[0252] O = 117.81 × 0.1 = 11.781 MB

[0253] The recalculated buffer size O=11.781MB indicates that the buffer should be set to about 11.8MB based on the current network speed and the buffering strategy set by the system. This helps to ensure the optimization of user experience and the maximization of buffering efficiency when the network speed is 117.81Mbps.

[0254] See also Figure 7 , the specific steps for obtaining the buffer parameter optimization results are:

[0255] According to the current buffer size O, use the formula:

[0256] Q'=O×(1+λ1·F)

[0257] Calculate the adjusted video buffer size Q';

[0258] Among them, λ1 is the weight coefficient used to adjust the buffer size to adapt to changes in network speed;

[0259] According to the adjusted video cache size Q', the formula is used:

[0260] R'=Q'×η1

[0261] Calculate the optimal buffer parameter R' and obtain the buffer parameter optimization result;

[0262] Where η1 is the performance evaluation coefficient, which is used to evaluate the buffering performance according to the adjusted cache size;

[0263] The calculation process is as follows:

[0264] For the formula:

[0265] Q′=O×(1+λ1·F)

[0266] Set the parameters O to 11.781MB, F to 76.26Mbps, and λ1 to 0.01;

[0267] Q′=11.781×(1+0.01×76.26)

[0268] Q′=11.781×(1+0.7626)

[0269] Q′=11.781×1.7626≈20.77MB

[0270] For the formula:

[0271] R′=Q′×η1

[0272] Set parameter η1 to 0.1;

[0273] R′=20.77×0.1R′=2.077MB

[0274] The buffer parameter R′=2.077 MB means that at the current network speed, after adjusting the buffer size, the system should maintain a buffer of approximately 2.077 MB to optimize the smoothness of the video stream and reduce the buffer delay.

[0275] See also Figure 8 The specific steps for obtaining the result of synchronous adjustment of TV program data are as follows:

[0276] According to the optimal buffer parameter R', the formula is used:

[0277]

[0278] Calculate the user's total viewing progress P on multiple devices track ;

[0279] Among them, E a is the current viewing progress of the ath device, a represents the device number, and g is the total number of devices used by the user;

[0280] Based on the user's total viewing progress on multiple devices track , using the formula:

[0281] S sync =α2×P track

[0282] Calculate the synchronization adjustment S sync , obtain the TV program data synchronization adjustment result;

[0283] Among them, α2 is the synchronization efficiency coefficient, which adjusts the synchronization process to match the playback status between multiple devices;

[0284] The calculation process is as follows:

[0285] For the formula:

[0286]

[0287] Set the parameters R′ to 2.077MB, g to 3 (for example, the user uses a mobile phone, TV, and tablet at the same time), E1=0.9, E2=0.5, E3=0.8;

[0288] P track =2.077×(0.9+0.5+0.8)=2.077×2.2=4.5694

[0289] For the formula:

[0290] S sync =α2×P track

[0291] Set parameter α2 to 0.95;

[0292] S sync =0.95×4.5694=4.34093

[0293] S sync=4.34093 means that under the current network and device status, the system should adjust the data to maintain a synchronization efficiency of approximately 4.34, setting the baseline value to 3.0, S sync If the value is higher than the benchmark, it means that the system's data synchronization efficiency exceeds the normal performance level under the current network and device status. This means that the system's adjustment measures are very effective and can ensure faster and more accurate synchronization of playback progress on multiple devices, thereby optimizing the user's viewing experience.

[0294] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A cross-platform TV program tracking system, characterized in that: The system comprises: The device prediction module collects user device usage data, builds a Bayesian network prediction model, analyzes the user's switching pattern between multiple devices, combines the viewing progress and content type of TV programs, predicts the time when users change devices through statistical analysis, and generates device switching prediction results; The environmental data analysis module collects environmental factor data associated with the user device, including user location information, device power status, and device network status, identifies key environmental factors that affect the user's device selection based on the device switching prediction results, adjusts the Bayesian network prediction model, and generates environmental factor analysis results; The buffer management module evaluates the network speed and the current buffer size through the Markov decision process according to the environmental factor analysis results, dynamically adjusts the video cache in combination with the current network status, and generates a buffer parameter optimization result; The real-time adjustment module monitors the viewing progress of the user on multiple devices based on the buffer parameter optimization result, adjusts the data synchronization in real time, and generates the TV program data synchronization adjustment result.

2. The cross-platform TV program tracking system according to claim 1, characterized in that: The acquisition steps of analyzing the switching mode of the user between multiple devices are specifically as follows: Collect the user's device usage data and mark each device usage data as an element of array D, that is, D = [s1, s2, ..., s n ], using the formula: Calculate the count C of device status S in the data set S ; Among them, s i represents the device status of the user at the i-th moment, C S is the count of the device state S, I is the indicative function, if the condition s i =S holds true, then I(s i =S) = 1, otherwise 0, n is the total number of data points in the device usage data; According to the count C of the device state S in the data set S , using the formula: Calculate the prior probability P(S) of the device state S; According to the prior probability P9S of the device state S, the formula is adopted: Calculate the conditional probability P(S|D) of observing data set D under device state S; Among them, S' is the overall device status.

3. The cross-platform TV program tracking system according to claim 2, characterized in that: The steps for obtaining the device switching prediction result are specifically as follows: Integrate the viewing progress v of the TV program and the content type b into an array V = [(v1, b1), (v2, b2), ..., (v m ,b m )], according to the conditional probability P(S|D0), the formula is: R j =P(S|D)×v j ×f(b j ) Calculate the probability R of a user switching devices j ; Among them, f(b j ) is the content type b j The impact function on the probability of switching devices, v j is the viewing progress of the jth data point, b j is the content type of the jth data point; According to the probability R of the user switching devices j , combined with the user history switching data set H, that is, H = [h1,h2,…,h m ], using the formula: Calculate the accuracy index A of the prediction of the user's device switching time; Among them, h j is the device switching situation at the jth time point, and m is the total number of historical switching data points of the user; According to the accuracy index A of the prediction of the user switching device time, the formula is adopted: T=max({R j |A>α1}) Calculate the switching time T and obtain the device switching prediction result; Among them, α1 is the threshold for evaluating accuracy, and max is used to select the maximum value in the set.

4. The cross-platform TV program tracking system according to claim 3, characterized in that: The steps of obtaining the key environmental factors that affect the user's device selection are specifically as follows: Collect environmental data, including user location information U, that is, U = [u1,u2,...,u x ], the device power state Q, that is, Q = [q1, q2, ..., q x ] and the device network status W, that is, W = [w1,w2,...,w x ], according to the switching time T, the formula is adopted: G k =t k ×(v u ·u T +v q ·q T +v w ·w T ) Calculate the environmental impact value G at each time point k ; Among them, G k is the environmental impact value at the kth time point, t k is the predicted probability of device switching corresponding to the switching time T, v u 、v q 、v w They are the weight factors of user location information, power status and network status. According to historical data analysis, u T ,q T 、w T are the values ​​of user location information, power status, and network status at time point T, respectively. k is used to iterate the environment data array; According to the environmental impact value G at each time point k , using the formula: and and Calculate the total influence G of each position u , the total influence of electricity G q and the total network influence G w , x is the total number of environmental data points; According to the total influence G of the position u , the total influence of electricity G q and the total network influence G w , using the formula: Calculate the impact score Z of each environmental factor; Among them, x represents the u , G q , G w Each value selected in .

5. The cross-platform TV program tracking system according to claim 4, characterized in that: The steps for obtaining the environmental factor analysis results are specifically as follows: According to the impact score Z of each environmental factor, the formula is used: P new (S)=P(S)+γ×Z Adjust the Bayesian network prediction model to reflect the influence of environmental factors and obtain the prior probability P of the adjusted device state S new (S); Among them, γ is the adjustment coefficient used to quantify the impact of environmental factors; According to the prior probability P of the adjusted device state S new (S), using the formula: Calculate the conditional probability P of the new device state S new (S|D) to obtain the environmental factor analysis results.

6. The cross-platform TV program tracking system according to claim 5, characterized in that: The steps of evaluating the network speed and obtaining the current buffer size are specifically as follows: According to the conditional probability P of the new device state S new (S|D), using the formula: Calculate the expected value F of the network speed; Among them, w k is the speed or quality indicator of the kth network state, k is used to iterate the environment data array, and x is the total number of environment data points; According to the expected value F of the network speed, the formula is adopted: O=F×δ1 Calculate the current buffer size O; Where δ1 is the buffer size adjustment factor.

7. The cross-platform TV program tracking system according to claim 6, characterized in that: The steps for obtaining the buffer parameter optimization result are specifically as follows: According to the current buffer size O, the formula is adopted: Q'=O×(1+λ1·F) Calculate the adjusted video buffer size Q'; Among them, λ1 is the weight coefficient used to adjust the buffer size; According to the adjusted video cache size Q', the formula is adopted: R'=Q'×η1 Calculate the optimal buffer parameter R' and obtain the buffer parameter optimization result; Among them, η1 is the performance evaluation coefficient.

8. The cross-platform TV program tracking system according to claim 7, characterized in that: The steps for obtaining the result of the synchronization adjustment of the television program data are specifically as follows: According to the optimal buffer parameter R', the formula is adopted: Calculate the user's total viewing progress P on multiple devices track ; Among them, E a is the current viewing progress of the ath device, where a represents the device number and g is the total number of devices used by the user; According to the total viewing progress P of the user on multiple devices track , using the formula: S sync =α2×P track Calculate the synchronization adjustment S sync , obtain the TV program data synchronization adjustment result; Among them, α2 is the synchronous efficiency coefficient.

Images