Multi-channel media playing control system and method based on data analysis technology

Abstract

The invention discloses a multi-channel media playing control system and method based on a data analysis technology, and relates to the technical field of media playing control, and the method comprises the steps: analyzing the equipment similarity degree between control equipment and historical control equipment, and obtaining similar historical control equipment; acquiring feature time data of similar historical control equipment of the control equipment, acquiring historical environment data and historical resource change records of the similar historical control equipment, and constructing a resource bottleneck prediction model of the control equipment; constructing a resource prediction model of the control equipment; the resource bottleneck prediction model is utilized to predict the resource use bottleneck of the control equipment in the current period to obtain target resource bottleneck data, the resource prediction model is utilized to predict the resource use condition of the control equipment, and the target resource bottleneck data is combined to optimize and adjust the media information controlled and played by the control equipment. And multi-channel media playing is intelligently controlled.

Description

Technical Field

[0001] The present invention relates to the technical field of media playback control, and in particular to a multi-channel media playback control system and method based on data analysis technology. Background Art

[0002] Multi-channel media refers to the communication media that integrates multiple media forms, such as text, sound, images, etc. It is called multi-channel because it can acquire, process and present information through multiple channels, such as vision and hearing. Compared with the traditional single media form, multi-channel media can provide a richer and more diverse information experience, and multi-channel media can acquire, store, organize and synthesize information in a multi-channel unified manner. Therefore, the demand for multi-channel media playback control exists in various industries, especially in occasions that require collaborative display, information release and audience interaction. Through a multi-channel media playback control system, the playback of audio and video content can be made more efficient, flexible and coordinated, thereby enhancing the audience experience and the effect of information presentation.

[0003] In real life, a control device needs to control video and audio playback on multiple media playback devices. Once the resource usage of the control device is too high, the control device may have problems, thereby affecting the media playback of the media playback device. The conventional method is to set thresholds for key indicators of the media playback device and process the device when resource usage exceeds the bottleneck. However, how to set thresholds for key indicators of the control device is often due to the lack of accurate reference objects, resulting in a lack of actual and sufficient data support for the setting of thresholds for the control device. At the same time, the environment in which the control device is running will also affect the setting of thresholds for key indicators, making it impossible to accurately control the device status of the control device, which will not only lead to a waste of resources, but may even cause serious playback accidents. Summary of the invention

[0004] The object of the present invention is to provide a multi-channel media playback control system and method based on data analysis technology to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a multi-channel media playback control method based on data analysis technology, the method comprising:

[0006] Step S100: construct a media playback control cloud platform, obtain device data of the control device, obtain historical device data of the historical control device, analyze the device similarity between the control device and the historical control device, and obtain similar historical control devices;

[0007] Step S200: Acquire characteristic time data of the control device with similar history, acquire historical environment data and historical resource change records of the control device with similar history based on the characteristic time data, and construct a resource bottleneck prediction model for the control device;

[0008] Step S300: Acquire the historical playback records and historical resource change records of the control device to build a resource prediction model for the control device;

[0009] Step S400: Use the resource bottleneck prediction model to predict the resource usage bottleneck of the control device in the current cycle, obtain target resource bottleneck data, use the resource prediction model to predict the resource usage status of the control device, and combine the target resource bottleneck data to optimize and adjust the media information controlled by the control device for playback, and perform intelligent control of multi-channel media playback.

[0010] Furthermore, step S100 includes:

[0011] Step S101: acquiring device data of a control device, where the control device controls media playback of a playback device, and the device data includes data corresponding to various device indicators of the control device;

[0012] Step S102: Acquire historical device data of the historical control device from the cloud platform, the historical device data including data corresponding to various device indicators of the historical control device;

[0013] Step S103: Based on the data corresponding to each device index of the control device, pre-process each device index, construct a feature vector A of the control device, and calculate the device similarity L between the control device and the historical control device A,B :

[0014]

[0015] Among them, A i Represented as the eigenvalue of the i-th element in the eigenvector A; B i It is represented as the eigenvalue of the i-th element in the eigenvector B of the historical control device;

[0016] Step S104: When the proximity of the devices is L A,B If the value is greater than a preset device similarity threshold, it is determined that the control device and the historical control device are similar, and the historical control device is recorded as a similar historical control device of the control device;

[0017] The above steps are to obtain the similar historical control devices of the control device because in actual situations, the number of failures in the playback control of the control device in normal use is very small, and it is impossible to build an accurate prediction model based on these small data sets. By obtaining similar historical control devices, the accuracy of the constructed model can be greatly improved.

[0018] Further, step S200 includes:

[0019] Step S201: Acquire each of the similar history control devices of the control device, and acquire characteristic time data of the similar history control devices, wherein the characteristic time data is the characteristic historical period of the similar history control devices, wherein the characteristic historical period is the historical period to which the time period during which the playback control failure of the similar history control devices belongs;

[0020] Step S202: Acquire historical environment data and historical resource change records of similar historical control devices within a characteristic historical period, wherein the historical environment data includes data corresponding to various device environment parameters of similar historical control devices within a characteristic historical period;

[0021] Step S203: Obtain preset thresholds D′ of various resource indicators of similar historical control devices from the cloud platform, obtain the average value of resource indicators of similar historical control devices during the time period when the playback control failure occurred from the historical resource change records, and record the average value as the characteristic value D of the resource indicator;

[0022] Step S204: Calculate the index bottleneck value C=D′ / |DD′| of the resource index in the historical resource change record, obtain the maximum value of the index bottleneck value of each resource index in the historical resource change record, and record a resource index corresponding to the maximum value as the characteristic resource index of the similar historical control device in the historical resource change record;

[0023] Step S205: construct a resource bottleneck prediction model for control devices, obtain characteristic values ​​of several characteristic resource indicators of historical resource change records of each similar historical control device within a characteristic historical period, and use them as output data of the resource bottleneck prediction model, and use data corresponding to various device environment parameters of each similar historical control device within the characteristic historical period as input data;

[0024] Step S206: Obtain the output data and input data of the resource bottleneck prediction model as the data set of the resource bottleneck prediction model, divide the data set into a training set and a test set, use the training set to train the resource bottleneck prediction model, and use the test set to calculate the model error M of the resource bottleneck prediction model:

[0025]

[0026] Where n represents the total number of test data in the test set; y i It is represented as the actual value of the characteristic value of the characteristic resource indicator of the i-th test data in the test set; y i ′ represents the resource bottleneck prediction model, which is the predicted value of the characteristic value of the characteristic resource indicator of the i-th test data;

[0027] Step S207: When the model error of the resource bottleneck prediction model is less than the preset model error threshold, it is determined that the resource bottleneck prediction model is constructed; otherwise, the resource bottleneck prediction model is trained and optimized until the model error of the resource bottleneck prediction model is less than the preset model error threshold.

[0028] Furthermore, step S300 includes:

[0029] Step S301: obtaining historical playback records of the control device, and obtaining historical media playback data from the historical playback records, wherein the historical media playback data includes data corresponding to various media playback features of various media information played by the control device in the historical playback records;

[0030] Step S302: Obtain the historical resource change record of the control device within the period of the historical playback record, and obtain the highest value of each resource indicator of the control device from the historical resource change record of the control device;

[0031] Step S303: construct a resource prediction model for the control device, use the highest value of each resource indicator in the historical resource change record of the control device as the output data of the resource prediction model, and use the historical media playback data as the input data of the resource prediction model. The specific construction process of the resource prediction model is as follows:

[0032] Obtain the preset weight matrix W and bias b of the resource prediction model, and obtain the activation function f(z)=1 / (1+e -z );

[0033] Step S304: Obtain the output of each layer in the resource prediction model, where the output a of the gth layer of the resource prediction model is (g) =f(z (g) ), z (g) =W (g) a (g-1) +b (g) , W (g) Represented as the weight matrix of the g-th layer, b (g) Represented as the bias of the g-th layer, a (g-1) Represented as the output of the g-1th layer;

[0034] Get the loss function H of the resource prediction model:

[0035] H=-∑ x x k log(Q′ k ),

[0036] Among them, Q k It is represented as the actual value of the highest value of the resource indicator of the kth input data in the resource prediction model; Q′ k It is expressed as the predicted value of the highest value of the resource indicator of the kth input data in the resource prediction model;

[0037] Update the weight matrix and bias of the g-th layer to obtain the gradient of the loss function in the g-th layer, where the specific formula of the gradient in the g-th layer is:

[0038]

[0039] Among them, m is the total number of input data in the resource prediction model;

[0040] The update algorithm for the weight matrix of the g-th layer is:

[0041]

[0042] The update algorithm for the bias of the gth layer is:

[0043]

[0044] The resource prediction model is trained through multiple iterations, and the weight matrix and bias of the resource prediction model are updated. When the absolute value of the difference between the loss value obtained by the loss function after the t-1th iteration and the loss value after the tth iteration is less than the preset loss change threshold, it is determined that the prediction of the resource prediction model obtained in the tth iteration is in line with expectations.

[0045] Furthermore, step S400 includes:

[0046] Step S401: acquiring the device operating environment data of the control device in the current cycle, the device operating environment data including data corresponding to various device environment parameters, and inputting the device operating environment data into the resource bottleneck prediction model of the control device, obtaining the thresholds of various resource indicators of the control device and aggregating them, and obtaining the target resource bottleneck data of the control device in the current cycle;

[0047] Step S402: acquiring playback media data of the control device in the current cycle, the playback media data including data corresponding to various media playback features of media information played by the control device in a plurality of playback devices, and predicting resource usage of the control device using a resource prediction model to obtain resource prediction data, the resource prediction data being the maximum values ​​of various resource indicators predicted by the control device in the current cycle;

[0048] Step S403: When there is a maximum value of a certain resource indicator among various resource indicators that is greater than a threshold value of a certain resource indicator, it is determined that there is a risk of control and playback failure when the control device controls several playback devices for media playback, and the media information controlled and played by the control device in the current cycle is adjusted until there is no maximum value of a certain resource indicator among various resource indicators that is greater than the threshold value of a certain resource indicator, and the control device is used to intelligently control multi-channel media playback on several playback devices.

[0049] In order to better implement the above method, a multi-channel media playback control system based on data analysis technology is also proposed, the system includes a device proximity analysis module, a resource bottleneck prediction model module, a resource prediction model construction module, and an intelligent playback control module;

[0050] The device similarity analysis module is used to analyze the device similarity between the control device and the historical control device, and obtain the similar historical control device;

[0051] The resource bottleneck prediction model module is used to obtain the historical environmental data and historical resource change records of similar historical control devices, and to build a resource bottleneck prediction model for the control device;

[0052] A resource prediction model building module is used to obtain the historical playback records and historical resource change records of the control device and build a resource prediction model for the control device;

[0053] The intelligent playback control module is used to predict the resource usage bottleneck of the control device in the current cycle, predict the resource usage status of the control device, optimize and adjust the media information played by the control device, and intelligently control multi-channel media playback.

[0054] Further, the device proximity analysis module includes a device proximity degree unit and a device proximity analysis unit;

[0055] A device similarity unit, used to calculate the device similarity between the control device and the historical control device;

[0056] The device similarity analysis unit is used to determine the device similarity between the control device and the historical control device according to the device similarity between the control device and the historical control device, and obtain the similar historical control device of the control device.

[0057] Further, the resource bottleneck prediction model module includes an indicator bottleneck value unit and a resource bottleneck prediction model unit;

[0058] The indicator bottleneck value unit is used to calculate the indicator bottleneck value of the resource indicator in the historical resource change record of the similar historical control device;

[0059] The resource bottleneck prediction model unit is used to construct a resource bottleneck prediction model for the control device according to the indicator bottleneck value and the characteristic resource indicators in the similar historical control device in the historical resource change record.

[0060] Further, the resource prediction model construction module includes a data acquisition unit and a resource prediction model construction unit;

[0061] A data acquisition unit, used to acquire the historical playback records of the control device, and acquire the historical resource change records of the control device within the period of the historical playback records;

[0062] The resource prediction model building unit is used to build a resource prediction model of the control device according to the historical resource change records and historical playback records of the control device.

[0063] Further, the intelligent playback control module includes an intelligent playback control unit;

[0064] The intelligent playback control unit is used to predict the resource usage bottleneck of the control device in the current cycle according to the resource bottleneck prediction model, obtain the target resource bottleneck data, use the resource prediction model to predict the resource usage status of the control device, and optimize and adjust the media information controlled by the control device in combination with the target resource bottleneck data, and use the control device to perform intelligent media playback control on multi-channel media playback on several playback devices.

[0065] Compared with the prior art, the beneficial effect of the present invention is as follows: the present invention realizes intelligent control of multi-channel media playback. For a single control device, it is very likely that a control playback failure has occurred in its historical period, but this does not mean that the resource usage bottleneck of the control device does not exist. Therefore, it is first obtained from a large number of historical control devices in the cloud platform. These historical control devices all have control playback failure records, and the degree of similarity between the control device and these historical control devices is analyzed, so as to obtain similar historical control devices similar to the control device, and analyze the resource usage bottlenecks of these devices in different environments, so that the constructed resource bottleneck prediction model is more in line with the control device, and at the same time, it can also provide early warning of control failures of the control device, and formulate optimization strategies in advance, which greatly reduces the probability of failure of the control device when playing on the playback device. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a method flow chart of a multi-channel media playback control system and method based on data analysis technology of the present invention;

[0067] Figure 2 It is a module schematic diagram of a multi-channel media playback control system and method based on data analysis technology of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a multi-channel media playback control method based on data analysis technology, the method comprising:

[0070] Step S100: construct a media playback control cloud platform, obtain device data of the control device, obtain historical device data of the historical control device, analyze the device similarity between the control device and the historical control device, and obtain similar historical control devices;

[0071] Wherein, step S100 includes:

[0072] Step S101: acquiring device data of a control device, where the control device controls media playback of a playback device, and the device data includes data corresponding to various device indicators of the control device;

[0073] Step S102: Acquire historical device data of the historical control device from the cloud platform, the historical device data including data corresponding to various device indicators of the historical control device;

[0074] For example, various device metrics include the device’s processor frequency, memory size, etc.;

[0075] Step S103: Based on the data corresponding to each device index of the control device, pre-process each device index, construct a feature vector A of the control device, and calculate the device similarity L between the control device and the historical control device A,B :

[0076]

[0077] Among them, A i Represented as the eigenvalue of the i-th element in the eigenvector A; B i It is represented as the eigenvalue of the i-th element in the eigenvector B of the historical control device;

[0078] For example, the historical device data of the control device obtained includes: the processor frequency of the device is 2.5 GHz, the memory of the device is 16, and the power consumption of the device is 100 W;

[0079] The historical device data of the historical control device obtained includes: the processor frequency of the device is 3.0 GHz, the memory of the device is 32, and the power consumption of the device is 120 W;

[0080] Get the characteristic vector A of the control device = [2.5, 16, 100], and get the characteristic vector A of the historical control device = [3, 32, 120];

[0081] Calculate the similarity between the control device and the historical control device L A,B :

[0082]

[0083] Step S104: When the proximity of the devices is L A,B If the value is greater than a preset device similarity threshold, it is determined that the control device and the historical control device are similar, and the historical control device is recorded as a similar historical control device of the control device;

[0084] Step S200: Acquire characteristic time data of the control device with similar history, acquire historical environment data and historical resource change records of the control device with similar history based on the characteristic time data, and construct a resource bottleneck prediction model for the control device;

[0085] Wherein, step S200 includes:

[0086] Step S201: Acquire each of the similar history control devices of the control device, and acquire characteristic time data of the similar history control devices, wherein the characteristic time data is the characteristic historical period of the similar history control devices, wherein the characteristic historical period is the historical period to which the time period during which the playback control failure of the similar history control devices belongs;

[0087] Step S202: Acquire historical environment data and historical resource change records of similar historical control devices within a characteristic historical period, wherein the historical environment data includes data corresponding to various device environment parameters of similar historical control devices within a characteristic historical period;

[0088] For example, various equipment environmental parameters include the equipment's ambient temperature, ambient humidity, etc.;

[0089] Step S203: Obtain preset thresholds D′ of various resource indicators of similar historical control devices from the cloud platform, obtain the average value of resource indicators of similar historical control devices during the time period when the playback control failure occurred from the historical resource change records, and record the average value as the characteristic value D of the resource indicator;

[0090] Step S204: Calculate the index bottleneck value C=D′ / |DD′| of the resource index in the historical resource change record, obtain the maximum value of the index bottleneck value of each resource index in the historical resource change record, and record a resource index corresponding to the maximum value as the characteristic resource index of the similar historical control device in the historical resource change record;

[0091] For example, various resource indicators, including the device’s CPU usage, memory usage, etc.

[0092] Step S205: construct a resource bottleneck prediction model for control devices, obtain characteristic values ​​of several characteristic resource indicators of historical resource change records of each similar historical control device within a characteristic historical period, and use them as output data of the resource bottleneck prediction model, and use data corresponding to various device environment parameters of each similar historical control device within the characteristic historical period as input data;

[0093] Step S206: Obtain the output data and input data of the resource bottleneck prediction model as the data set of the resource bottleneck prediction model, divide the data set into a training set and a test set, use the training set to train the resource bottleneck prediction model, and use the test set to calculate the model error M of the resource bottleneck prediction model:

[0094]

[0095] Where n represents the total number of test data in the test set; y iIt is represented as the actual value of the characteristic value of the characteristic resource indicator of the i-th test data in the test set; y i ′ represents the resource bottleneck prediction model, which is the predicted value of the characteristic value of the characteristic resource indicator of the i-th test data;

[0096] Step S207: When the model error of the resource bottleneck prediction model is less than the preset model error threshold, it is determined that the resource bottleneck prediction model is constructed; otherwise, the resource bottleneck prediction model is trained and optimized until the model error of the resource bottleneck prediction model is less than the preset model error threshold;

[0097] Step S300: Acquire the historical playback records and historical resource change records of the control device to build a resource prediction model for the control device;

[0098] Wherein, step S300 includes:

[0099] Step S301: obtaining historical playback records of the control device, and obtaining historical media playback data from the historical playback records, wherein the historical media playback data includes data corresponding to various media playback features of various media information played by the control device in the historical playback records;

[0100] Step S302: Obtain the historical resource change record of the control device within the period of the historical playback record, and obtain the highest value of each resource indicator of the control device from the historical resource change record of the control device;

[0101] Step S303: construct a resource prediction model for the control device, use the highest value of each resource indicator in the historical resource change record of the control device as the output data of the resource prediction model, and use the historical media playback data as the input data of the resource prediction model. The specific construction process of the resource prediction model is as follows:

[0102] Obtain the preset weight matrix W and bias b of the resource prediction model, and obtain the activation function f(z)=1 / (1+e -z );

[0103] Step S304: Obtain the output of each layer in the resource prediction model, where the output a of the gth layer of the resource prediction model is (g) =f(z (g) ), z (g) =W (g) a (g-1) +b (g) , W (g) Represented as the weight matrix of the g-th layer, b (g) Represented as the bias of the g-th layer, a (g-1) Represented as the output of the g-1th layer;

[0104] Get the loss function H of the resource prediction model:

[0105] H=-∑ k Q k log(Q′ k ),

[0106] Among them, Q k It is represented as the actual value of the highest value of the resource indicator of the kth input data in the resource prediction model; Q′ k It is expressed as the predicted value of the highest value of the resource indicator of the kth input data in the resource prediction model;

[0107] Update the weight matrix and bias of the g-th layer to obtain the gradient of the loss function in the g-th layer, where the specific formula of the gradient in the g-th layer is:

[0108]

[0109] Among them, m is the total number of input data in the resource prediction model;

[0110] The update algorithm for the weight matrix of the g-th layer is:

[0111]

[0112] The update algorithm for the bias of the gth layer is:

[0113]

[0114] The resource prediction model is trained through multiple iterations, and the weight matrix and bias of the resource prediction model are updated. When the absolute value of the difference between the loss value obtained by the loss function at the t-1th iteration and the loss value after the tth iteration is less than the preset loss change threshold, it is determined that the prediction of the resource prediction model obtained at the tth iteration meets expectations;

[0115] Step S400: using the resource bottleneck prediction model, predicting the resource usage bottleneck of the control device in the current cycle, obtaining target resource bottleneck data, using the resource prediction model, predicting the resource usage status of the control device, and optimizing and adjusting the media information played by the control device in combination with the target resource bottleneck data, and performing intelligent control on multi-channel media playback;

[0116] Wherein, step S400 includes:

[0117] Step S401: acquiring the device operating environment data of the control device in the current cycle, the device operating environment data including data corresponding to various device environment parameters, and inputting the device operating environment data into the resource bottleneck prediction model of the control device, obtaining the thresholds of various resource indicators of the control device and aggregating them, and obtaining the target resource bottleneck data of the control device in the current cycle;

[0118] Step S402: acquiring playback media data of the control device in the current cycle, the playback media data including data corresponding to various media playback features of media information played by the control device in a plurality of playback devices, and predicting resource usage of the control device using a resource prediction model to obtain resource prediction data, the resource prediction data being the maximum values ​​of various resource indicators predicted by the control device in the current cycle;

[0119] Step S403: when there is a maximum value of a certain resource indicator among various resource indicators, which is greater than a threshold value of a certain resource indicator, it is determined that there is a risk of control and playback failure when the control device controls several playback devices to play media, and the media information controlled and played by the control device in the current cycle is adjusted until there is no maximum value of a certain resource indicator among various resource indicators, which is greater than the threshold value of a certain resource indicator, and the control device is used to perform intelligent control on multi-channel media playback on several playback devices;

[0120] In order to better implement the above method, a multi-channel media playback control system based on data analysis technology is also proposed, the system includes a device proximity analysis module, a resource bottleneck prediction model module, a resource prediction model construction module, and an intelligent playback control module;

[0121] The device similarity analysis module is used to analyze the device similarity between the control device and the historical control device, and obtain the similar historical control device;

[0122] The resource bottleneck prediction model module is used to obtain the historical environmental data and historical resource change records of similar historical control devices, and to build a resource bottleneck prediction model for the control device;

[0123] A resource prediction model building module is used to obtain the historical playback records and historical resource change records of the control device and build a resource prediction model for the control device;

[0124] Intelligent playback control module, used to predict the resource usage bottleneck of the control device in the current cycle, predict the resource usage status of the control device, optimize and adjust the media information played by the control device, and intelligently control multi-channel media playback;

[0125] Among them, the equipment proximity analysis module includes an equipment proximity degree unit and an equipment proximity analysis unit;

[0126] A device similarity unit, used to calculate the device similarity between the control device and the historical control device;

[0127] A device proximity analysis unit, used to determine the device proximity between the control device and the historical control device according to the device proximity between the control device and the historical control device, and obtain the similar historical control device of the control device;

[0128] Among them, the resource bottleneck prediction model module includes an indicator bottleneck value unit and a resource bottleneck prediction model unit;

[0129] The indicator bottleneck value unit is used to calculate the indicator bottleneck value of the resource indicator in the historical resource change record of the similar historical control device;

[0130] The resource bottleneck prediction model unit is used to construct a resource bottleneck prediction model for the control device according to the bottleneck value of the indicator and the characteristic resource indicators of the control device in the similar history in the historical resource change record;

[0131] Among them, the resource prediction model construction module includes a data acquisition unit and a resource prediction model construction unit;

[0132] A data acquisition unit, used to acquire the historical playback records of the control device, and acquire the historical resource change records of the control device within the period of the historical playback records;

[0133] A resource prediction model building unit, used to build a resource prediction model of the control device according to the historical resource change records and historical playback records of the control device;

[0134] Among them, the intelligent playback control module includes an intelligent playback control unit;

[0135] The intelligent playback control unit is used to predict the resource usage bottleneck of the control device in the current cycle according to the resource bottleneck prediction model, obtain the target resource bottleneck data, use the resource prediction model to predict the resource usage status of the control device, and optimize and adjust the media information controlled by the control device in combination with the target resource bottleneck data, and use the control device to perform intelligent media playback control on multi-channel media playback on several playback devices.

[0136] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A multi-channel media playback control method based on data analysis technology, characterized in that: The method comprises: Step S100: constructing a media playback control cloud platform, obtaining device data of a control device, obtaining historical device data of a historical control device, analyzing the device similarity between the control device and the historical control device, and obtaining similar historical control devices; Step S200: Acquire characteristic time data of the control device with similar history to the control device, acquire historical environment data and historical resource change records of the control device with similar history based on the characteristic time data, and construct a resource bottleneck prediction model for the control device; Step S300: Acquire the historical playback records and historical resource change records of the control device to construct a resource prediction model for the control device; Step S400: Utilize the resource bottleneck prediction model to predict the resource usage bottleneck of the control device in the current cycle to obtain target resource bottleneck data; utilize the resource prediction model to predict the resource usage status of the control device; and optimize and adjust the media information controlled by the control device in combination with the target resource bottleneck data to perform intelligent control on multi-channel media playback.

2. The multi-channel media playback control method based on data analysis technology according to claim 1, characterized in that: The step S100 includes: Step S101: acquiring device data of a control device, wherein the control device controls media playback of a playback device, and the device data includes data corresponding to various device indicators of the control device; Step S102: Acquire historical device data of the historical control device from the cloud platform, wherein the historical device data includes data corresponding to various device indicators of the historical control device; Step S103: Based on the data corresponding to each device indicator of the control device, pre-process the device indicators, construct the feature vector A of the control device, and calculate the device similarity L between the control device and the historical control device. A,B : Among them, A i is represented by the eigenvalue of the i-th element in the eigenvector A; B i is represented as the eigenvalue of the i-th element in the eigenvector B of the historical control device; Step S104: When the proximity L of the devices A,B If the value is greater than a preset device similarity threshold, it is determined that the control device and the historical control device are similar devices, and the historical control device is recorded as a similar historical control device of the control device.

3. The multi-channel media playback control method based on data analysis technology according to claim 2, characterized in that: The step S200 includes: Step S201: Acquire each of the similar history control devices of the control device, and acquire characteristic time data of the similar history control devices, wherein the characteristic time data is the characteristic historical period of the similar history control devices, wherein the characteristic historical period is the historical period to which the time period during which the playback control failure of the similar history control device belongs; Step S202: Acquire historical environment data and historical resource change records of the similar historical control device within a characteristic historical period, wherein the historical environment data includes data corresponding to various device environment parameters of the similar historical control device within the characteristic historical period; Step S203: obtaining preset thresholds D′ of various resource indicators of the similar historical control devices from the cloud platform, obtaining the average value of the resource indicators of the similar historical control devices during the time period when the playback control failure occurred from the historical resource change records, and recording the average value as the characteristic value D of the resource indicator; Step S204: Calculate the index bottleneck value C=D′ / |DD′| of the resource index in the historical resource change record, obtain the maximum value of the index bottleneck values ​​of each resource index in the historical resource change record, and record a resource index corresponding to the maximum value as the characteristic resource index of the similar historical control device in the historical resource change record; Step S205: constructing a resource bottleneck prediction model for the control device, obtaining characteristic values ​​of several characteristic resource indicators of the historical resource change records of each of the similar historical control devices within the characteristic historical period, and using the characteristic values ​​as output data of the resource bottleneck prediction model, and using the data corresponding to the various device environment parameters of each of the similar historical control devices within the characteristic historical period as input data; Step S206: Obtain the output data and input data of the resource bottleneck prediction model as the data set of the resource bottleneck prediction model, divide the data set into a training set and a test set, use the training set to train the resource bottleneck prediction model, and use the test set to calculate the model error M of the resource bottleneck prediction model: Where n represents the total number of test data in the test set; y i y is represented by the actual value of the characteristic value of the characteristic resource indicator of the i-th test data in the test set; i ′ represents the predicted value of the characteristic value of the characteristic resource indicator of the i-th test data by the resource bottleneck prediction model; Step S207: When the model error of the resource bottleneck prediction model is less than the preset model error threshold, it is determined that the resource bottleneck prediction model is constructed; otherwise, the resource bottleneck prediction model is trained and optimized until the model error of the resource bottleneck prediction model is less than the preset model error threshold.

4. The multi-channel media playback control method based on data analysis technology according to claim 3 is characterized in that: The step S300 includes: Step S301: Acquire historical playback records of the control device, and obtain historical media playback data from the historical playback records, wherein the historical media playback data includes data corresponding to various media playback features of various media information played by the control device in the historical playback records; Step S302: obtaining a historical resource change record of the control device within the time period of the historical playback record, and obtaining the highest value of each resource indicator of the control device from the historical resource change record of the control device; Step S303: construct a resource prediction model for the control device, use the highest value of each resource indicator in the historical resource change record of the control device as output data of the resource prediction model, and use the historical media playback data as input data of the resource prediction model. The specific construction process of the resource prediction model is as follows: Obtain the preset weight matrix W and bias b of the resource prediction model, and obtain the activation function f(z)=1 / (1+e -z ); Step S304: Obtain the output of each layer in the resource prediction model, wherein the output a of the gth layer of the resource prediction model is (g) =f(z (g) ), z (g) =W (g) a (g-1) +b (g) , W (g) Represented as the weight matrix of the g-th layer, b (g) Denotes the bias of the g-th layer, a (g-1) Represented as the output of the g-1th layer; Obtain the loss function H of the resource prediction model: H=-∑ k Q k log(Q′ k ), Among them, Q k It is represented by the actual value of the highest value of the resource indicator of the kth input data in the resource prediction model; Q′ k It is represented as the predicted value of the highest value of the resource indicator of the kth input data in the resource prediction model; The weight matrix and bias of the g-th layer are updated to obtain the gradient of the loss function in the g-th layer, wherein the specific formula of the gradient in the g-th layer is: Wherein, m is the total number of each input data in the resource prediction model; The updating algorithm of the weight matrix of the g-th layer is: The update algorithm of the bias of the g-th layer is: The resource prediction model is trained through multiple iterations, and the weight matrix and bias of the resource prediction model are updated. When the absolute value of the difference between the loss value obtained by the loss function after the t-1th iteration and the loss value after the tth iteration is less than the preset loss change threshold, it is determined that the prediction of the resource prediction model obtained in the tth iteration meets expectations.

5. The multi-channel media playback control method based on data analysis technology according to claim 4 is characterized in that: The step S400 includes: Step S401: acquiring the device operating environment data of the control device in the current cycle, the device operating environment data including data corresponding to various device environment parameters, and inputting the device operating environment data into the resource bottleneck prediction model of the control device, obtaining the threshold values ​​of various resource indicators of the control device and aggregating them, and obtaining the target resource bottleneck data of the control device in the current cycle; Step S402: acquiring the playback media data of the control device in the current cycle, the playback media data including data corresponding to various media playback features of the media information played by the control device in a plurality of playback devices, and predicting the resource usage status of the control device using the resource prediction model to obtain resource prediction data, the resource prediction data being the maximum values ​​of various resource indicators predicted by the control device in the current cycle; Step S403: When there is a maximum value of a certain resource indicator among the various resource indicators that is greater than the threshold value of the certain resource indicator, it is determined that there is a risk of control and playback failure when the control device controls several playback devices for media playback, and the media information controlled and played by the control device in the current cycle is adjusted until there is no maximum value of a certain resource indicator among the various resource indicators that is greater than the threshold value of the certain resource indicator, and the control device is used to intelligently control multi-channel media playback on several playback devices.

6. A multi-channel media playback control system based on data analysis technology, used to execute a multi-channel media playback control method based on data analysis technology as described in any one of claims 1 to 5, characterized in that: The system includes a device proximity analysis module, a resource bottleneck prediction model module, a resource prediction model construction module, and an intelligent playback control module; The device similarity analysis module is used to obtain similar historical control devices based on the device similarity between the control device and the historical control devices; The resource bottleneck prediction model module is used to obtain the historical environment data and historical resource change records of the similar historical control device and to construct a resource bottleneck prediction model for the control device; The resource prediction model building module is used to acquire the historical playback records and historical resource change records of the control device to build a resource prediction model of the control device; The intelligent playback control module is used to predict the resource usage bottleneck of the control device in the current cycle, predict the resource usage status of the control device, optimize and adjust the media information controlled by the control device for playback, and intelligently control multi-channel media playback.

7. The multi-channel media playback control system based on data analysis technology according to claim 6, characterized in that: The device proximity analysis module includes a device proximity degree unit and a device proximity analysis unit; The device similarity unit is used to calculate the device similarity between the control device and the historical control device; The device proximity analysis unit is used to determine the device proximity between the control device and the historical control device according to the device proximity between the control device and the historical control device, and obtain the similar historical control device of the control device.

8. The multi-channel media playback control system based on data analysis technology according to claim 6, characterized in that: The resource bottleneck prediction model module includes an indicator bottleneck value unit and a resource bottleneck prediction model unit; The indicator bottleneck value unit is used to calculate the indicator bottleneck value of the resource indicator in the historical resource change record of the similar historical control device; The resource bottleneck prediction model unit is used to construct a resource bottleneck prediction model for the control device according to the indicator bottleneck value and the characteristic resource indicators in the similar historical control device in the historical resource change record.

9. The multi-channel media playback control system based on data analysis technology according to claim 6, characterized in that: The resource prediction model construction module includes a data acquisition unit and a resource prediction model construction unit; The data acquisition unit is used to acquire the historical playback record of the control device, and acquire the historical resource change record of the control device within the time period of the historical playback record; The resource prediction model building unit is used to build a resource prediction model of the control device according to the historical resource change records and historical playback records of the control device.

10. The multi-channel media playback control system based on data analysis technology according to claim 6, characterized in that: The intelligent playback control module includes an intelligent playback control unit; The intelligent playback control unit is used to predict the resource usage bottleneck of the control device in the current cycle according to the resource bottleneck prediction model, obtain target resource bottleneck data, use the resource prediction model to predict the resource usage status of the control device, and optimize and adjust the media information controlled by the control device in combination with the target resource bottleneck data, and use the control device to perform intelligent media playback control on multi-channel media playback on multiple playback devices.

Images