Intelligent Bluetooth sound box sub-control adjusting system and method based on Internet

Abstract

The invention discloses an intelligent Bluetooth sound box sub-control adjustment system and method based on the Internet, and relates to the technical field of intelligent Bluetooth sound box sub-control adjustment of the Internet, and the system comprises a Bluetooth sound box intelligent collection module, a preprocessing module, a volume adjustment module, a music preference evaluation module, a model fusion module and an execution module. The Bluetooth sound box intelligent acquisition module is divided into an environment sound intensity unit and a historical music playing unit, the volume adjustment module constructs an intelligent volume adjustment model by using a random forest algorithm, and the music preference evaluation module defines a music preference type and obtains user music preference. The Bluetooth sound box intelligent acquisition technology, the random forest algorithm and the CART algorithm are closely combined with the modern information technology, real-time and comprehensive monitoring of the volume of the intelligent Bluetooth sound box and the music preference of the user is achieved, and the intelligent degree in the intelligent Bluetooth sound box sub-control adjustment process of the Internet is remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent Bluetooth speaker sub-control adjustment of the Internet, and specifically relates to an intelligent Bluetooth speaker sub-control adjustment system and method based on the Internet. Background Art

[0002] With the rapid development of technology, people have put forward higher requirements for the functions and experiences of audio devices. Traditional Bluetooth speakers are limited by local connections and simple control modes and are unable to cope when facing complex scenarios. The single control method of traditional speakers cannot meet the adjustment of the functions of multi-scenario intelligent speakers. At the same time, the booming development of Internet technology provides the possibility to solve these problems. Through the Internet, intelligent Bluetooth speakers break through traditional technical limitations. It is extremely necessary to develop an intelligent Bluetooth speaker sub-control adjustment system based on the Internet. It can rely on the powerful data transmission and processing capabilities of the Internet to break through the limitations of traditional speakers, provide users with a more intelligent, flexible and personalized audio control experience, and meet the growing audio control needs in various scenarios;

[0003] Although the existing technology has made great progress in the direction of intelligent Bluetooth speaker sub-control adjustment of the Internet, there are still some problems to be optimized. It is difficult for the existing technology to adjust the volume of the intelligent Bluetooth speaker according to the sound intensity of the environment where the intelligent Bluetooth speaker is located, and it is not easy to analyze the music preferences of users through the historical music file playback information of users and recommend relevant audio content in a personalized manner. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent Bluetooth speaker sub-control adjustment system and method based on the Internet to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the technical solutions adopted by the present invention are as follows: In the first aspect, an intelligent Bluetooth speaker sub-control adjustment system based on the Internet includes a Bluetooth speaker intelligent acquisition module, a preprocessing module, a volume adjustment module, a music preference evaluation module, a model fusion module and an execution module, wherein each module is communicatively connected;

[0006] The Bluetooth speaker intelligent acquisition module is divided into an environmental sound intensity unit and a historical music playback unit. Among them, the environmental sound intensity unit is used to collect the environmental sound intensity of the intelligent Bluetooth speaker, and the historical music playback unit is used to collect the names and playback times of the historical music played by the intelligent Bluetooth speaker, providing data support for the implementation of subsequent modules;

[0007] The preprocessing module preprocesses the environmental sound intensity of the intelligent Bluetooth speaker and the names and playback times of the historical music played, and divides the volume range of the intelligent Bluetooth speaker, providing data reserves for constructing an intelligent volume adjustment model;

[0008] The volume adjustment module uses the random forest algorithm to construct an intelligent volume adjustment model, solving the problem in the prior art that it is difficult to adjust the volume of the intelligent Bluetooth speaker according to the sound intensity of the environment where the intelligent Bluetooth speaker is located;

[0009] The music preference evaluation module defines music preference types, obtains user music preferences, and uses the CART algorithm to construct an intelligent music preference evaluation model, solving the problem in the prior art that it is not easy to analyze the music preferences of users through the playback information of user historical music files and recommend relevant audio in a personalized manner;

[0010] The model fusion module fuses the intelligent volume adjustment model and the intelligent music preference evaluation model through meta-learning to construct an intelligent Bluetooth speaker sub-control adjustment model, providing technical support for enhancing the adaptability of the system;

[0011] The execution module adjusts the volume of the intelligent Bluetooth speaker according to the volume range of the intelligent Bluetooth speaker; recommends music to the user according to the user's music preferences.

[0012] A further improvement of the technical solution of the present invention lies in that: the process of the ambient sound intensity unit collecting the ambient sound intensity of the intelligent Bluetooth speaker includes:

[0013] Place the sound level meter 1 meter away from the intelligent Bluetooth speaker, adjust the height of the sound level meter so that the sound level meter and the speaker of the intelligent Bluetooth speaker are at the same horizontal height, turn on the intelligent Bluetooth speaker, and wait for the playback state of the intelligent Bluetooth speaker to stabilize, then press the measurement button of the sound level meter to collect the ambient sound intensity of the intelligent Bluetooth speaker.

[0014] A further improvement of the technical solution of the present invention lies in that: the process of the historical music playback unit collecting the names and playback times of the historical music played by the intelligent Bluetooth speaker includes:

[0015] Apply for access rights to the historical playback record database storing the intelligent Bluetooth speaker, analyze the specific data in the historical playback record of the intelligent Bluetooth speaker, determine the storage format of the names and playback times of the historical music played by the intelligent Bluetooth speaker, and use a Python script to write a query statement according to the storage format of the names and playback times of the historical music played by the intelligent Bluetooth speaker to extract the names and playback times of the historical music played in the historical playback record database of the intelligent Bluetooth speaker.

[0016] A further improvement of the technical solution of the present invention lies in that: the process of the preprocessing module preprocessing the ambient sound intensity of the intelligent Bluetooth speaker and the names and playback times of the historical music played and dividing the volume range of the intelligent Bluetooth speaker includes:

[0017] Clean the data of the environmental sound intensity of the smart Bluetooth speaker, the name of the historical played music and its playing times, and remove abnormal data and duplicate data;

[0018] Based on the non-linear perception metric standard of the human ear for sound intensity, the process of calculating the sound pressure level of the Bluetooth speaker environment is as follows:

[0019]

[0020] Among them, L is the sound pressure level of the Bluetooth speaker environment, in decibels, I is the environmental sound intensity of the smart Bluetooth speaker, and I 0 is the environmental reference sound intensity of the smart Bluetooth speaker, usually taking I 0 = 10 -12 W / m 2 ;

[0021] Set two sound pressure level thresholds, represented by Z1 and Z2, and Z1 < Z2. The sound pressure level interval from 0 to Z1 is used as the low sound intensity interval of the Bluetooth speaker environment, the sound pressure level interval from Z1 to Z2 is used as the medium sound intensity interval of the Bluetooth speaker environment, and the sound pressure level interval greater than Z2 is used as the high sound intensity interval of the Bluetooth speaker environment;

[0022] When the environment of the smart Bluetooth speaker is in the low sound intensity interval, set the volume interval of the smart Bluetooth speaker to 0% - 20%; when the environment of the smart Bluetooth speaker is in the medium sound intensity interval, set the volume interval of the smart Bluetooth speaker to 20% - 50%; when the environment of the smart Bluetooth speaker is in the high sound intensity interval, set the volume interval of the smart Bluetooth speaker to 50% - 100%.

[0023] A further improvement of the technical solution of the present invention is that the volume adjustment module uses the random forest algorithm to construct the intelligent volume adjustment model, and the process includes:

[0024] Encode the volume intervals of the smart Bluetooth speaker of 0% - 20%, 20% - 50% and 50% - 100% as 0, 1 and 2 respectively. Through the random forest algorithm, construct a random forest model, set the random forest model parameters, and use the sound pressure level of the Bluetooth speaker environment and its corresponding volume interval encoding of the smart Bluetooth speaker as the data set, and divide it into a training set and a test set according to the ratio of 7:3;

[0025] Input the training set data into the random forest model, train each decision tree in the random forest model, adjust the splitting conditions of the decision tree, use the sound pressure level of the Bluetooth speaker environment as the input, and the volume interval encoding of the smart Bluetooth speaker as the output, learn the non-linear relationship between the sound pressure level of the Bluetooth speaker environment and its corresponding volume interval encoding of the smart Bluetooth speaker, and obtain the trained random forest model;

[0026] Input the test set data into the trained random forest model to output the encoded volume range of the smart Bluetooth speaker. By comparing the encoded volume range of the smart Bluetooth speaker output by the random forest model with the actual volume range of the smart Bluetooth speaker, if the random forest model is overfitted, optimize the random forest model by reducing the decision tree depth and increasing the minimum number of samples for node splitting; if the random forest model is underfitted, optimize the random forest model by increasing the number of decision trees;

[0027] Deploy the optimized random forest model to an Internet-based intelligent Bluetooth speaker sub-control adjustment system to obtain an intelligent volume adjustment model;

[0028] Input the sound pressure level of the smart Bluetooth speaker environment into the intelligent volume adjustment model. When the intelligent volume adjustment model outputs 0, it means that the sound pressure level of the smart Bluetooth speaker environment is in the low sound intensity range, and the corresponding power range of the smart Bluetooth speaker is 0% - 20%; when the intelligent volume adjustment model outputs 1, it means that the sound pressure level of the smart Bluetooth speaker environment is in the medium sound intensity range, and the corresponding power range of the smart Bluetooth speaker is 20% - 50%; when the intelligent volume adjustment model outputs 2, it means that the sound pressure level of the smart Bluetooth speaker environment is in the high sound intensity range, and the corresponding power range of the smart Bluetooth speaker is 50% - 100%.

[0029] A further improvement of the technical solution of the present invention lies in: the music preference evaluation module defines music preference types, and the process of obtaining the user's music preference includes:

[0030] The music preferences are divided into two categories. According to the music style, they are divided into pop music, classical music, and rock music; according to the emotional needs, they are divided into cheerful, soothing, and sad. Assign music preference labels to the historical played music of the smart Bluetooth speaker. The music preference label includes a music style preference label and an emotional need preference label;

[0031] Traverse the playback times of all the historical played music of the smart Bluetooth speaker, retrieve the historical played music with the highest playback times, record the music style preference label and the emotional need preference label of the historical played music respectively, and obtain the user's music preference.

[0032] A further improvement of the technical solution of the present invention lies in: the music preference evaluation module uses the CART algorithm, and the process of constructing an intelligent music preference evaluation model includes:

[0033] Encode pop music, classical music, rock music, upbeat, soothing, and sad as A, B, C, D, E, and F respectively, obtain the user's music preference encoding, use the playback times of the music played by the smart Bluetooth speaker in history and its corresponding user music preference encoding as the dataset, and divide it into a training set and a test set according to the ratio of 8:2. Select the playback times of the music played by the smart Bluetooth speaker in history as the feature variable, and select the user music preference encoding as the target variable;

[0034] Extract sample data from the dataset. According to the sample data, divide the sample types, evenly distribute each type of sample to each node, traverse all features and split points of each node, calculate the Gini impurity after splitting, and select the feature and split point with the largest decrease in Gini impurity for splitting. The formula for this Gini impurity is: where Gini(p) is the Gini impurity after splitting, and p i is the proportion of the i-th type of sample in the node. Set the Gini impurity threshold. Starting from the root node, for each node, according to the feature and split point with the largest decrease in Gini impurity, split the node to generate a left child node and a right child node. Through recursion, repeat the splitting process until the Gini impurity after splitting is lower than the Gini impurity threshold, then stop splitting to obtain the CART decision tree model;

[0035] Use the training set data to train the CART decision tree model, adjust the structure and node splitting conditions of the CART decision tree, learn the non-linear relationship between the playback times of the music played by the smart Bluetooth speaker in history and the user music preference encoding, and obtain the trained CART decision tree model;

[0036] Input the test set data into the CART decision tree model, compare the output result of the CART decision tree model with the encoding of the user music preference corresponding to the actual playback times of the music played by the smart Bluetooth speaker in history, evaluate the performance of the CART decision tree model, and optimize the performance of the CART decision tree model through pre-pruning and post-pruning;

[0037] Deploy the optimized CART decision tree model to an Internet-based smart Bluetooth speaker sub-control adjustment system to obtain a smart music preference evaluation model;

[0038] Input the playback times of the music played by the smart Bluetooth speaker in history into the smart music preference evaluation model to obtain the corresponding user music preference encoding. The output user music preference encoding includes A, B, C, D, E, and F, and the corresponding user music preferences are pop music, classical music, rock music, upbeat, soothing, and sad respectively.

[0039] A further improvement of the technical solution of the present invention lies in: the model fusion module, through meta-learning, fuses the intelligent volume adjustment model and the intelligent music preference evaluation model. The process of constructing the intelligent Bluetooth speaker sub-control adjustment model includes:

[0040] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent volume adjustment model, and the intelligent volume adjustment model outputs the corresponding intelligent Bluetooth speaker power range encoding; input the playback times of the intelligent Bluetooth speaker's historical played music into the intelligent music preference evaluation model, and the intelligent music preference evaluation model outputs the corresponding user music preference encoding;

[0041] Combine the actual intelligent Bluetooth speaker power range encoding, the actual user music preference encoding, the output result of the intelligent volume adjustment model, and the output result of the intelligent music preference evaluation model to obtain metadata. Select the MLP meta-model architecture and divide the metadata into a training set, a validation set, and a test set according to the ratio of 7:2:1;

[0042] Use the training set metadata to train the meta-model, adjust the meta-model parameters through the backpropagation algorithm, learn the non-linear relationship between the sound pressure level of the Bluetooth speaker environment and its corresponding intelligent Bluetooth speaker volume range encoding, and the non-linear relationship between the playback times of the intelligent Bluetooth speaker's historical played music and the user music preference encoding. Combine the validation set metadata to adjust the hyperparameters of the meta-model to obtain a trained meta-model;

[0043] Input the test set metadata into the trained meta-model, compare the output result of the meta-model with the actual intelligent Bluetooth speaker power range encoding and the actual user music preference encoding, evaluate the performance of the meta-model, adjust the meta-model parameters, optimize the performance of the meta-model, and obtain an optimized meta-model;

[0044] Deploy the optimized meta-model to an Internet-based intelligent Bluetooth speaker sub-control adjustment system to obtain an intelligent Bluetooth speaker sub-control adjustment model;

[0045] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent Bluetooth speaker sub-control adjustment model, and the intelligent Bluetooth speaker sub-control adjustment model outputs the same output result as the intelligent volume adjustment model; input the playback times of the intelligent Bluetooth speaker's historical played music into the intelligent Bluetooth speaker sub-control adjustment model, and the intelligent Bluetooth speaker sub-control adjustment model outputs the same output result as the intelligent music preference evaluation model.

[0046] A further improvement of the technical solution of the present invention lies in: the execution module, according to the intelligent Bluetooth speaker volume range, adjusts the volume of the intelligent Bluetooth speaker; according to the user's music preference, the process of recommending music for the user includes:

[0047] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent Bluetooth speaker sub-control adjustment model to obtain the intelligent Bluetooth speaker volume range code, and based on the intelligent Bluetooth speaker volume range corresponding to the intelligent Bluetooth speaker volume range code, obtain the intelligent Bluetooth speaker volume range;

[0048] Input the playback times of the music played by the intelligent Bluetooth speaker historically into the intelligent Bluetooth speaker sub-control adjustment model to obtain the user music preference code, and based on the user music preference corresponding to the user music preference code, obtain the user music preference;

[0049] When the sound pressure level of the intelligent Bluetooth speaker environment is in the low sound intensity range, adjust the intelligent Bluetooth speaker power range to 0% - 20%; when the sound pressure level of the intelligent Bluetooth speaker environment is in the medium sound intensity range, adjust the intelligent Bluetooth speaker power range to 20% - 50%; when the sound pressure level of the intelligent Bluetooth speaker environment is in the high sound intensity range, adjust the intelligent Bluetooth speaker power range to 50% - 100%;

[0050] In terms of music style, when the user music preference is pop music, the intelligent Bluetooth speaker increases the recommendation times for pop music; when the user music preference is classical music, the intelligent Bluetooth speaker increases the recommendation times for classical music; when the user music preference is rock music, the intelligent Bluetooth speaker increases the recommendation times for rock music;

[0051] In terms of emotional needs, when the user music preference is lively, the intelligent Bluetooth speaker increases the recommendation times for lively music; when the user music preference is soothing, the intelligent Bluetooth speaker increases the recommendation times for soothing music; when the user music preference is sad, the intelligent Bluetooth speaker increases the recommendation times for sad music.

[0052] Second, an intelligent Bluetooth speaker sub-control adjustment method based on the Internet, used to implement the above-mentioned intelligent Bluetooth speaker sub-control adjustment system based on the Internet, consists of the following steps:

[0053] Step 1: Use a sound level meter to collect the ambient sound intensity of the intelligent Bluetooth speaker; through the historical playback record database of the intelligent Bluetooth speaker, obtain the name and playback times of the music played by the intelligent Bluetooth speaker historically;

[0054] Step 2: Clean the data of the ambient sound intensity of the intelligent Bluetooth speaker and the name and playback times of the music played historically, calculate the sound pressure level of the Bluetooth speaker environment based on the non-linear perception metric standard of the human ear for sound intensity, set two sound pressure level thresholds, and divide the intelligent Bluetooth speaker volume range according to the two sound pressure level thresholds;

[0055] Step 3: Use the random forest algorithm to construct an intelligent volume adjustment model;

[0056] Step 4: Define music preference types, assign music preference tags to the names and play counts of the music played by the smart Bluetooth speaker in history, obtain the user's music preferences based on the play counts of the music played in history, and use the CART algorithm to construct a smart music preference evaluation model;

[0057] Step 5: Through meta-learning, fuse the smart volume adjustment model and the smart music preference evaluation model to construct a smart Bluetooth speaker sub-control adjustment model;

[0058] Step 6: Adjust the volume of the smart Bluetooth speaker according to the volume range of the smart Bluetooth speaker; recommend music to the user according to the user's music preferences.

[0059] The beneficial effects of the present invention are as follows: A sub-control adjustment system and method for a smart Bluetooth speaker based on the Internet. Compared with the traditional sub-control adjustment system and method for a smart Bluetooth speaker based on the Internet, technologies such as the smart acquisition technology of the Bluetooth speaker, the random forest algorithm, and the CART algorithm in the method of the present invention are closely combined with modern information technologies. Through the smart acquisition technology of the Bluetooth speaker, the environmental sound intensity of the smart Bluetooth speaker and the names and play counts of the music played in history are accurately captured. The random forest algorithm is used to construct a smart volume adjustment model, and the CART algorithm is used to construct a smart music preference evaluation model, achieving real-time and comprehensive monitoring of the volume of the smart Bluetooth speaker and the user's music preferences, solving the problems that it is difficult to adjust the volume of the smart Bluetooth speaker according to the sound intensity of the environment where the smart Bluetooth speaker is located in the prior art, and it is not easy to analyze the user's music preferences through the user's historical music file play information and recommend relevant audio in a personalized manner, ensuring that the method in the present invention can refine the dynamic monitoring standard for the sub-control adjustment system and method of the smart Bluetooth speaker within a more accurate range, making the monitored data more accurate indicators under the same conditions. The research and application of this method significantly enhance the degree of intelligence in the sub-control adjustment process of the smart Bluetooth speaker based on the Internet. Description of the Drawings

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0061] Figure 1 It is a block diagram of a sub-control adjustment system for a smart Bluetooth speaker based on the Internet according to the present invention;

[0062] Figure 2 It is a flowchart of a sub-control adjustment method for a smart Bluetooth speaker based on the Internet according to the present invention. Detailed Embodiments

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0064] Embodiment 1, as Figure 1 shown, the present invention provides an intelligent Bluetooth speaker sub-control adjustment system based on the Internet, including a Bluetooth speaker intelligent acquisition module, a preprocessing module, a volume adjustment module, a music preference evaluation module, a model fusion module, and an execution module. Among them, each module is communicatively connected;

[0065] The Bluetooth speaker intelligent acquisition module is divided into an ambient sound intensity unit and a historical music playback unit. Among them, the ambient sound intensity unit is used to collect the ambient sound intensity of the intelligent Bluetooth speaker, and the historical music playback unit is used to collect the names and playback times of the historical music played by the intelligent Bluetooth speaker, providing data support for the implementation of subsequent modules;

[0066] The preprocessing module preprocesses the ambient sound intensity of the intelligent Bluetooth speaker and the names and playback times of the historical music played, and divides the volume range of the intelligent Bluetooth speaker, providing data reserves for building an intelligent volume adjustment model;

[0067] The volume adjustment module uses the random forest algorithm to build an intelligent volume adjustment model, solving the problem in the prior art that it is difficult to adjust the volume of the intelligent Bluetooth speaker according to the sound intensity of the environment where the intelligent Bluetooth speaker is located;

[0068] The music preference evaluation module defines music preference types, obtains user music preferences, and uses the CART algorithm to build an intelligent music preference evaluation model, solving the problem in the prior art that it is not easy to analyze the music preferences of users and recommend relevant audio personalized through the user's historical music file playback information;

[0069] The model fusion module fuses the intelligent volume adjustment model and the intelligent music preference evaluation model through meta-learning to build an intelligent Bluetooth speaker sub-control adjustment model, providing technical support for enhancing the adaptability of the system;

[0070] The execution module adjusts the volume of the intelligent Bluetooth speaker according to the volume range of the intelligent Bluetooth speaker; recommends music to the user according to the user's music preferences.

[0071] Preferably, the process of the ambient sound intensity unit collecting the ambient sound intensity of the intelligent Bluetooth speaker includes:

[0072] Place the sound level meter 1 meter away from the smart Bluetooth speaker. Adjust the height of the sound level meter so that it is at the same horizontal level as the speaker of the smart Bluetooth speaker. Turn on the smart Bluetooth speaker. After the playback state of the smart Bluetooth speaker stabilizes, press the measurement button of the sound level meter to collect the ambient sound intensity of the smart Bluetooth speaker.

[0073] Preferably, for the historical music playback unit, the process of collecting the names of the historical music played by the smart Bluetooth speaker and their playback times includes:

[0074] Apply for access rights to the database storing the historical playback records of the smart Bluetooth speaker, analyze the specific data in the historical playback records of the smart Bluetooth speaker, determine the storage format of the names and playback times of the historical music played by the smart Bluetooth speaker, and use a Python script to write a query statement according to the storage format of the names and playback times of the historical music played by the smart Bluetooth speaker to extract the names of the historical music played by the smart Bluetooth speaker and their playback times from the historical playback record database of the smart Bluetooth speaker.

[0075] Preferably, for the preprocessing module, the process of preprocessing the ambient sound intensity of the smart Bluetooth speaker and the names of the historical music played by the smart Bluetooth speaker and their playback times and dividing the volume range of the smart Bluetooth speaker includes:

[0076] Perform data cleaning on the ambient sound intensity of the smart Bluetooth speaker and the names of the historical music played by the smart Bluetooth speaker and their playback times to remove abnormal data and duplicate data;

[0077] Based on the non-linear perception metric standard of the human ear for sound intensity, the process of calculating the sound pressure level of the Bluetooth speaker environment is as follows:

[0078]

[0079] Where L is the sound pressure level of the Bluetooth speaker environment, in decibels, I is the ambient sound intensity of the smart Bluetooth speaker, and I 0 is the ambient reference sound intensity of the smart Bluetooth speaker, usually taking I 0 = 10 -12 W / m 2 ;

[0080] Set two sound pressure level thresholds, denoted as Z1 and Z2, and Z1 < Z2. The sound pressure level range from 0 to Z1 is taken as the low sound intensity range of the Bluetooth speaker environment, the sound pressure level range from Z1 to Z2 is taken as the medium sound intensity range of the Bluetooth speaker environment, and the sound pressure level range greater than Z2 is taken as the high sound intensity range of the Bluetooth speaker environment;

[0081] When the environment of the intelligent Bluetooth speaker is in the low sound intensity range, the volume range of the intelligent Bluetooth speaker is set to 0% - 20%; when the environment of the intelligent Bluetooth speaker is in the medium sound intensity range, the volume range of the intelligent Bluetooth speaker is set to 20% - 50%; when the environment of the intelligent Bluetooth speaker is in the high sound intensity range, the volume range of the intelligent Bluetooth speaker is set to 50% - 100%.

[0082] Preferably, for the volume adjustment module, the process of constructing the intelligent volume adjustment model using the random forest algorithm includes:

[0083] Encode the volume ranges of the intelligent Bluetooth speaker at 0% - 20%, 20% - 50%, and 50% - 100% as 0, 1, and 2 respectively. Through the random forest algorithm, construct a random forest model, set the parameters of the random forest model, and use the sound pressure level of the intelligent Bluetooth speaker environment and its corresponding volume range encoding of the intelligent Bluetooth speaker as the data set, and divide it into a training set and a test set according to a ratio of 7:3;

[0084] Input the training set data into the random forest model, train each decision tree in the random forest model, adjust the splitting conditions of the decision tree, use the sound pressure level of the Bluetooth speaker environment as the input, and the volume range encoding of the intelligent Bluetooth speaker as the output, learn the non-linear relationship between the sound pressure level of the Bluetooth speaker environment and its corresponding volume range encoding of the intelligent Bluetooth speaker, and obtain the trained random forest model;

[0085] Input the test set data into the trained random forest model, output the volume range encoding of the intelligent Bluetooth speaker. By comparing the volume range of the intelligent Bluetooth speaker corresponding to the encoding output by the random forest model with the actual volume range of the intelligent Bluetooth speaker, if the random forest model is overfitted, optimize the random forest model by reducing the decision tree depth and increasing the minimum number of samples for node splitting; if the random forest model is underfitted, optimize the random forest model by increasing the number of decision trees;

[0086] Deploy the optimized random forest model to an Internet-based intelligent Bluetooth speaker sub-control adjustment system to obtain the intelligent volume adjustment model;

[0087] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent volume adjustment model. When the intelligent volume adjustment model outputs 0, it means that the sound pressure level of the intelligent Bluetooth speaker environment is in the low sound intensity range, and the corresponding intelligent Bluetooth speaker power range is 0% - 20%; when the intelligent volume adjustment model outputs 1, it means that the sound pressure level of the intelligent Bluetooth speaker environment is in the medium sound intensity range, and the corresponding intelligent Bluetooth speaker power range is 20% - 50%; when the intelligent volume adjustment model outputs 2, it means that the sound pressure level of the intelligent Bluetooth speaker environment is in the high sound intensity range, and the corresponding intelligent Bluetooth speaker power range is 50% - 100%.

[0088] Preferably, the music preference evaluation module defines music preference types. The process of obtaining the user's music preferences includes:

[0089] Among them, music preferences are divided into two categories. According to music styles, they are divided into pop music, classical music, and rock music; according to emotional needs, they are divided into lively, soothing, and sad. Music preference labels are assigned to the historical played music of the smart Bluetooth speaker. The music preference label includes a music style preference label and an emotional need preference label;

[0090] Traverse the play counts of all historical played music of the smart Bluetooth speaker, retrieve the historical played music with the highest play count, record the music style preference label and the emotional need preference label of this historical played music respectively, and obtain the user's music preferences.

[0091] Preferably, the music preference evaluation module uses the CART algorithm. The process of constructing a smart music preference evaluation model includes:

[0092] Encode pop music, classical music, rock music, lively, soothing, and sad as A, B, C, D, E, and F respectively to obtain the user's music preference encoding. Use the play counts of the historical played music of the smart Bluetooth speaker and their corresponding user music preference encodings as a data set, and divide it into a training set and a test set according to a ratio of 8:2. Select the play counts of the historical played music of the smart Bluetooth speaker as the feature variable and select the user music preference encoding as the target variable;

[0093] Extract sample data from the data set, divide the sample types according to the sample data, evenly distribute each type of sample to each node, traverse all features and split points of each node, calculate the Gini impurity after splitting, and select the feature and split point with the largest decrease in Gini impurity for splitting. The calculation formula for this Gini impurity is: Among them, Gini(p) is the Gini impurity after splitting, and p i is the proportion of the i-th type of sample in the node. Set the Gini impurity threshold. Starting from the root node, for each node, according to the feature and split point with the largest decrease in Gini impurity, split the node to generate a left child node and a right child node. Through a recursive method, repeat the splitting process until the Gini impurity after splitting is lower than the Gini impurity threshold, then stop splitting to obtain the CART decision tree model;

[0094] Use the training set data to train the CART decision tree model, adjust the structure and node splitting conditions of the CART decision tree, learn the non-linear relationship between the play counts of the historical played music of the smart Bluetooth speaker and the user music preference encoding, and obtain the trained CART decision tree model;

[0095] Input the test set data into the CART decision tree model, compare the output result of the CART decision tree model with the encoding of the actual number of plays of the music played by the intelligent Bluetooth speaker corresponding to the user's music preference, evaluate the performance of the CART decision tree model, and optimize the performance of the CART decision tree model through pre-pruning and post-pruning;

[0096] Deploy the optimized CART decision tree model to an Internet-based intelligent Bluetooth speaker sub-control adjustment system to obtain an intelligent music preference evaluation model;

[0097] Input the number of plays of the music played by the intelligent Bluetooth speaker into the intelligent music preference evaluation model to obtain the corresponding user music preference encoding. The output user music preference encoding includes A, B, C, D, E, and F, corresponding to the user music preferences of pop music, classical music, rock music, lively, soothing, and sad, respectively.

[0098] Preferably, the model fusion module, through meta-learning, fuses the intelligent volume adjustment model and the intelligent music preference evaluation model. The process of constructing the intelligent Bluetooth speaker sub-control adjustment model includes:

[0099] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent volume adjustment model, and the intelligent volume adjustment model outputs the corresponding intelligent Bluetooth speaker power range encoding; input the number of plays of the music played by the intelligent Bluetooth speaker into the intelligent music preference evaluation model, and the intelligent music preference evaluation model outputs the corresponding user music preference encoding;

[0100] Combine the actual intelligent Bluetooth speaker power range encoding, the actual user music preference encoding, the output result of the intelligent volume adjustment model, and the output result of the intelligent music preference evaluation model to obtain metadata, select the MLP meta-model architecture, and divide the metadata into a training set, a validation set, and a test set according to the ratio of 7:2:1;

[0101] Use the training set metadata to train the meta-model, adjust the meta-model parameters through the backpropagation algorithm, learn the non-linear relationship between the sound pressure level of the Bluetooth speaker environment and its corresponding intelligent Bluetooth speaker volume range encoding and the non-linear relationship between the number of plays of the music played by the intelligent Bluetooth speaker and the user music preference encoding, and combine the validation set metadata to adjust the hyperparameters of the meta-model to obtain the trained meta-model;

[0102] Input the test set metadata into the trained meta-model, compare the output result of the meta-model with the actual intelligent Bluetooth speaker power range encoding and the actual user music preference encoding, evaluate the performance of the meta-model, adjust the meta-model parameters, optimize the performance of the meta-model, and obtain the optimized meta-model;

[0103] Deploy the optimized meta-model to an Internet-based intelligent Bluetooth speaker sub-control adjustment system to obtain an intelligent Bluetooth speaker sub-control adjustment model;

[0104] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent Bluetooth speaker sub-control adjustment model, and the intelligent Bluetooth speaker sub-control adjustment model outputs the same output result as the intelligent volume adjustment model; input the playback times of the music played by the intelligent Bluetooth speaker in the past into the intelligent Bluetooth speaker sub-control adjustment model, and the intelligent Bluetooth speaker sub-control adjustment model outputs the same output result as the intelligent music preference evaluation model.

[0105] Preferably, the execution module adjusts the volume of the intelligent Bluetooth speaker according to the intelligent Bluetooth speaker volume range; the process of recommending music to the user according to the user's music preference includes:

[0106] Input the sound pressure level of the intelligent Bluetooth speaker environment into the intelligent Bluetooth speaker sub-control adjustment model, obtain the intelligent Bluetooth speaker volume range code, and obtain the intelligent Bluetooth speaker volume range according to the intelligent Bluetooth speaker volume range corresponding to the intelligent Bluetooth speaker volume range code;

[0107] Input the playback times of the music played by the intelligent Bluetooth speaker in the past into the intelligent Bluetooth speaker sub-control adjustment model, obtain the user music preference code, and obtain the user music preference according to the user music preference corresponding to the user music preference code;

[0108] When the sound pressure level of the intelligent Bluetooth speaker environment is in the low sound intensity range, adjust the intelligent Bluetooth speaker power range to 0% - 20%; when the sound pressure level of the intelligent Bluetooth speaker environment is in the medium sound intensity range, adjust the intelligent Bluetooth speaker power range to 20% - 50%; when the sound pressure level of the intelligent Bluetooth speaker environment is in the high sound intensity range, adjust the intelligent Bluetooth speaker power range to 50% - 100%;

[0109] In terms of music style, when the user's music preference is pop music, the intelligent Bluetooth speaker increases the recommendation times of pop music; when the user's music preference is classical music, the intelligent Bluetooth speaker increases the recommendation times of classical music; when the user's music preference is rock music, the intelligent Bluetooth speaker increases the recommendation times of rock music;

[0110] In terms of emotional needs, when the user's music preference is lively, the intelligent Bluetooth speaker increases the recommendation times of lively music; when the user's music preference is soothing, the intelligent Bluetooth speaker increases the recommendation times of soothing music; when the user's music preference is sad, the intelligent Bluetooth speaker increases the recommendation times of sad music.

[0111] Embodiment 2, as Figure 2 shown, on the basis of Embodiment 1, the present invention provides a technical solution: an intelligent Bluetooth speaker sub-control adjustment method based on the Internet, used to implement the above-mentioned intelligent Bluetooth speaker sub-control adjustment system based on the Internet, and consists of the following steps:

[0112] Step 1: Use a sound level meter to collect the ambient sound intensity of the smart Bluetooth speaker; obtain the names and play counts of the music played by the smart Bluetooth speaker historically through the historical play record database of the smart Bluetooth speaker.

[0113] Step 2: Clean the data of the ambient sound intensity of the smart Bluetooth speaker and the names and play counts of the music played historically. Based on the non-linear perception metric standard of sound intensity by the human ear, calculate the sound pressure level of the Bluetooth speaker environment, set two sound pressure level thresholds, and divide the volume range of the smart Bluetooth speaker according to the two sound pressure level thresholds.

[0114] Step 3: Use the random forest algorithm to construct an intelligent volume adjustment model.

[0115] Step 4: Define music preference types, assign music preference labels to the names and play counts of the music played by the smart Bluetooth speaker historically, obtain the user's music preferences according to the play counts of the music played historically, and use the CART algorithm to construct an intelligent music preference evaluation model.

[0116] Step 5: Through meta-learning, fuse the intelligent volume adjustment model and the intelligent music preference evaluation model to construct an intelligent Bluetooth speaker sub-control adjustment model.

[0117] Step 6: Adjust the volume of the smart Bluetooth speaker according to the volume range of the smart Bluetooth speaker; recommend music to the user according to the user's music preferences.

[0118] As described above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

Claims

1. An Internet-based smart Bluetooth speaker sub-control and adjustment system, including a Bluetooth speaker intelligent acquisition module, a preprocessing module, a volume adjustment module, a music preference evaluation module, a model fusion module and an execution module, wherein: Each module is communicatively connected, and is characterized in that: The intelligent acquisition module of the Bluetooth speaker is divided into an ambient sound intensity unit and a historical music playback unit. Among them, the ambient sound intensity unit is used to collect the ambient sound intensity of the intelligent Bluetooth speaker, and the historical music playback unit is used to collect the name and playback times of the historical music played by the intelligent Bluetooth speaker; The preprocessing module preprocesses the ambient sound intensity of the intelligent Bluetooth speaker and the name and playback times of the historical music played, and divides the volume range of the intelligent Bluetooth speaker; The volume adjustment module uses the random forest algorithm to construct an intelligent volume adjustment model; The music preference evaluation module defines music preference types, obtains user music preferences, and uses the CART algorithm to construct an intelligent music preference evaluation model; The model fusion module fuses the intelligent volume adjustment model and the intelligent music preference evaluation model through meta-learning to construct an intelligent Bluetooth speaker sub-control adjustment model; The execution module adjusts the volume of the intelligent Bluetooth speaker according to the volume range of the intelligent Bluetooth speaker; and recommends music to the user according to the user's music preferences.

2. According to claim 1, an Internet-based smart Bluetooth speaker sub-control adjustment system is characterized in that: The process of the ambient sound intensity unit collecting the ambient sound intensity of the intelligent Bluetooth speaker includes: Place the sound level meter 1 meter away from the intelligent Bluetooth speaker, adjust the height of the sound level meter so that the sound level meter and the speaker of the intelligent Bluetooth speaker are at the same horizontal height. Turn on the intelligent Bluetooth speaker, and after waiting for the playback state of the intelligent Bluetooth speaker to stabilize, press the measurement button of the sound level meter to collect the ambient sound intensity of the intelligent Bluetooth speaker.

3. According to claim 2, an Internet-based smart Bluetooth speaker sub-control and adjustment system is characterized in that: The process of the historical music playback unit collecting the name and playback times of the historical music played by the intelligent Bluetooth speaker includes: Apply for access rights to the database storing the historical playback records of the intelligent Bluetooth speaker, analyze the specific data in the historical playback records of the intelligent Bluetooth speaker, determine the storage format of the name and playback times of the historical music played by the intelligent Bluetooth speaker, and use a Python script to write a query statement according to the storage format of the name and playback times of the historical music played by the intelligent Bluetooth speaker to extract the name and playback times of the historical music played from the historical playback record database of the intelligent Bluetooth speaker.

4. The Internet-based smart Bluetooth speaker sub-control and adjustment system according to claim 3 is characterized in that: The process of the preprocessing module preprocessing the ambient sound intensity of the intelligent Bluetooth speaker and the name and playback times of the historical music played, and dividing the volume range of the intelligent Bluetooth speaker includes: Perform data cleaning on the ambient sound intensity of the intelligent Bluetooth speaker and the name and playback times of the historical music played to remove abnormal data and duplicate data; Based on the non-linear perception metric standard of the human ear for sound intensity, the process of calculating the sound pressure level of the Bluetooth speaker environment is as follows: Where L is the sound pressure level of the Bluetooth speaker environment, in decibels, I is the ambient sound intensity of the smart Bluetooth speaker, and I0 is the ambient reference sound intensity of the smart Bluetooth speaker, usually I0=10 -12 W / m 2 ; Set two sound pressure level thresholds, denoted as Z1 and Z2, and Z1 < Z2. The sound pressure level range of 0 to Z1 is used as the low sound intensity range of the intelligent Bluetooth speaker environment, the sound pressure level range of Z1 to Z2 is used as the medium sound intensity range of the intelligent Bluetooth speaker environment, and the sound pressure level range greater than Z2 is used as the high sound intensity range of the intelligent Bluetooth speaker environment; When the environment of the smart Bluetooth speaker is in a low sound intensity range, the volume range of the smart Bluetooth speaker is set to 0% to 20%; when the environment of the smart Bluetooth speaker is in a medium sound intensity range, the volume range of the smart Bluetooth speaker is set to 20% to 50%; when the environment of the smart Bluetooth speaker is in a high sound intensity range, the volume range of the smart Bluetooth speaker is set to 50% to 100%.

5. The Internet-based smart Bluetooth speaker sub-control and adjustment system according to claim 4 is characterized in that: The volume adjustment module uses the random forest algorithm to construct an intelligent volume adjustment model, which includes: The volume intervals of the smart Bluetooth speakers of 0% to 20%, 20% to 50%, and 50% to 100% are encoded as 0, 1, and 2, respectively. A random forest model is constructed using the random forest algorithm. The parameters of the random forest model are set. The sound pressure level of the smart Bluetooth speaker environment and its corresponding volume interval encoding of the smart Bluetooth speaker are used as a data set and divided into a training set and a test set in a ratio of 7:

3. Input the training set data into the random forest model, train each decision tree in the random forest model, adjust the splitting condition of the decision tree, take the sound pressure level of the Bluetooth speaker environment as input, take the volume interval code of the smart Bluetooth speaker as output, learn the nonlinear relationship between the sound pressure level of the Bluetooth speaker environment and the volume interval code of its corresponding smart Bluetooth speaker, and obtain a trained random forest model; Input the test set data into the trained random forest model, output the volume range code of the smart Bluetooth speaker, and compare the volume range of the smart Bluetooth speaker corresponding to the code output by the random forest model with the actual volume range of the smart Bluetooth speaker. If the random forest model is overfitting, optimize the random forest model by reducing the depth of the decision tree and increasing the minimum number of samples for node splitting; if the random forest model is underfitting, optimize the random forest model by increasing the number of decision trees; The optimized random forest model is deployed into an Internet-based smart Bluetooth speaker sub-control adjustment system to obtain an intelligent volume adjustment model; The sound pressure level of the smart Bluetooth speaker environment is input into the intelligent volume adjustment model. When the intelligent volume adjustment model outputs 0, it means that the sound pressure level of the smart Bluetooth speaker environment is in the low sound intensity range, and the corresponding smart Bluetooth speaker power range is 0% to 20%; when the intelligent volume adjustment model outputs 1, it means that the sound pressure level of the smart Bluetooth speaker environment is in the medium sound intensity range, and the corresponding smart Bluetooth speaker power range is 20% to 50%; when the intelligent volume adjustment model outputs 2, it means that the sound pressure level of the smart Bluetooth speaker environment is in the high sound intensity range, and the corresponding smart Bluetooth speaker power range is 50% to 100%.

6. The Internet-based smart Bluetooth speaker sub-control and adjustment system according to claim 5 is characterized in that: The music preference evaluation module defines the music preference type and the process of obtaining the user's music preference includes: The music preferences are divided into two categories: pop music, classical music and rock music according to music style; cheerful, soothing and sad according to emotional needs, and music preference tags are assigned to the historically played music of the smart Bluetooth speaker, and the music preference tags include music style preference tags and emotional needs preference tags; Traverse the number of times all smart Bluetooth speakers have played music in history, retrieve the music with the highest number of times played in history, record the music style preference label and emotional demand preference label of the music in history, and obtain the user's music preference.

7. The Internet-based smart Bluetooth speaker sub-control and adjustment system according to claim 6 is characterized in that: The music preference evaluation module uses the CART algorithm to construct an intelligent music preference evaluation model, including: Pop music, classical music, rock music, cheerful, soothing and sad music are coded as A, B, C, D, E and F respectively, and the user music preference code is obtained. The number of times the smart Bluetooth speaker plays music in the past and its corresponding user music preference code are used as the data set, which is divided into a training set and a test set in a ratio of 8:

2. The number of times the smart Bluetooth speaker plays music in the past is selected as the feature variable, and the user music preference code is selected as the target variable. Extract sample data from the data set, divide the sample types according to the sample data, divide each type of sample equally into each node, traverse all the features and splitting points of each node, calculate the Gini impurity after splitting, and select the features and splitting points with the largest decrease in Gini impurity for splitting. The calculation formula of the Gini impurity is: Among them, Gini(p) is the Gini impurity after splitting, p i Set the Gini impurity threshold for the proportion of the i-th type of samples in the node, start from the root node, and for each node, split the node according to the feature and split point with the largest Gini impurity drop to generate a left child node and a right child node. Repeat the splitting process recursively until the Gini impurity after the split is lower than the Gini impurity threshold, stop splitting, and obtain the CART decision tree model; The training set data is used to train the CART decision tree model, the structure and node splitting conditions of the CART decision tree are adjusted, the nonlinear relationship between the number of times the smart Bluetooth speaker has played music in history and the user's music preference encoding is learned, and a trained CART decision tree model is obtained; The test set data is input into the CART decision tree model, and the output of the CART decision tree model is compared with the encoding of the user's music preference corresponding to the number of times the music is played in the history of the actual smart Bluetooth speaker, and the performance of the CART decision tree model is evaluated. The performance of the CART decision tree model is optimized through pre-pruning and post-pruning; The optimized CART decision tree model is deployed into an Internet-based smart Bluetooth speaker sub-control adjustment system to obtain an intelligent music preference evaluation model; The number of times the smart Bluetooth speaker has played music in the past is input into the smart music preference evaluation model to obtain the corresponding user music preference code. The output user music preference code includes A, B, C, D, E and F, and the corresponding user music preferences are pop music, classical music, rock music, cheerful, soothing and sad, respectively.

8. The Internet-based smart Bluetooth speaker sub-control and adjustment system according to claim 7 is characterized in that: The model fusion module integrates the intelligent volume adjustment model and the intelligent music preference evaluation model through meta-learning to construct the intelligent Bluetooth speaker sub-control adjustment model. The process includes: The sound pressure level of the smart Bluetooth speaker environment is input into the smart volume adjustment model, and the smart volume adjustment model outputs the corresponding smart Bluetooth speaker power interval code; the number of times the smart Bluetooth speaker has played music in history is input into the smart music preference evaluation model, and the smart music preference evaluation model outputs the corresponding user music preference code; Combine the actual smart Bluetooth speaker power range encoding, the actual user music preference encoding, the output results of the smart volume adjustment model, and the output results of the smart music preference evaluation model to obtain metadata, select the MLP metamodel architecture, and divide the metadata into training set, validation set, and test set in a ratio of 7:2:1; Use the training set metadata to train the meta-model, adjust the meta-model parameters through the back-propagation algorithm, learn the nonlinear relationship between the sound pressure level of the Bluetooth speaker environment and the volume range encoding of its corresponding smart Bluetooth speaker, and the nonlinear relationship between the number of times the smart Bluetooth speaker has played music in history and the encoding of the user's music preference, and adjust the hyper-parameters of the meta-model in combination with the validation set metadata to obtain the trained meta-model; Input the test set metadata into the trained meta-model, compare the output of the meta-model with the actual smart Bluetooth speaker power range encoding and the actual user music preference encoding, evaluate the performance of the meta-model, adjust the meta-model parameters, optimize the meta-model performance, and obtain the optimized meta-model; Deploy the optimized meta-model to an Internet-based smart Bluetooth speaker sub-control adjustment system to obtain a smart Bluetooth speaker sub-control adjustment model; The sound pressure level of the smart Bluetooth speaker environment is input into the smart Bluetooth speaker sub-control adjustment model, and the smart Bluetooth speaker sub-control adjustment model outputs the same output result as the smart volume adjustment model; the number of times the smart Bluetooth speaker has played music in history is input into the smart Bluetooth speaker sub-control adjustment model, and the smart Bluetooth speaker sub-control adjustment model outputs the same output result as the smart music preference evaluation model.

9. The Internet-based smart Bluetooth speaker sub-control and adjustment system according to claim 8, characterized in that: The execution module adjusts the volume of the smart Bluetooth speaker according to the volume range of the smart Bluetooth speaker; The process of recommending music to users based on their music preferences includes: Input the sound pressure level of the smart Bluetooth speaker environment into the smart Bluetooth speaker sub-control adjustment model to obtain the smart Bluetooth speaker volume interval code, and obtain the smart Bluetooth speaker volume interval according to the smart Bluetooth speaker volume interval corresponding to the smart Bluetooth speaker volume interval code; Input the number of times the smart Bluetooth speaker has played music in the past into the smart Bluetooth speaker sub-control adjustment model to obtain the user music preference code, and obtain the user music preference according to the user music preference corresponding to the user music preference code; When the sound pressure level of the smart Bluetooth speaker environment is in the low sound intensity range, adjust the smart Bluetooth speaker power range to 0% to 20%; when the sound pressure level of the smart Bluetooth speaker environment is in the medium sound intensity range, adjust the smart Bluetooth speaker power range to 20% to 50%; when the sound pressure level of the smart Bluetooth speaker environment is in the high sound intensity range, adjust the smart Bluetooth speaker power range to 50% to 100%; In terms of music style, when the user's music preference is pop music, the smart Bluetooth speaker increases the number of recommendations for pop music; when the user's music preference is classical music, the smart Bluetooth speaker increases the number of recommendations for classical music; when the user's music preference is rock music, the smart Bluetooth speaker increases the number of recommendations for rock music; In terms of emotional needs, when the user's music preference is cheerful, the smart Bluetooth speaker increases the number of recommendations for cheerful music; when the user's music preference is soothing, the smart Bluetooth speaker increases the number of recommendations for soothing music; when the user's music preference is sad, the smart Bluetooth speaker increases the number of recommendations for sad music.

10. An Internet-based smart Bluetooth speaker sub-control adjustment method, implemented based on an Internet-based smart Bluetooth speaker sub-control adjustment system as described in any one of claims 1 to 9, characterized in that: It consists of the following steps: Step 1: Use a sound level meter to collect the ambient sound intensity of the smart Bluetooth speaker; obtain the name and number of times the smart Bluetooth speaker has played historical music through the smart Bluetooth speaker's historical playback record database; Step 2: Perform data cleaning on the ambient sound intensity of the smart Bluetooth speaker and the name and number of times the music has been played in the past. Based on the nonlinear perception metric of the human ear to the sound intensity, calculate the sound pressure level of the Bluetooth speaker environment, set two sound pressure level thresholds, and divide the volume range of the smart Bluetooth speaker according to the two sound pressure level thresholds. Step 3: Use the random forest algorithm to build an intelligent volume adjustment model; Step 4: Define the music preference type, assign music preference labels to the names of the music played in the history of the smart Bluetooth speaker and its play times, obtain the user's music preference based on the play times of the music played in the history, and use the CART algorithm to build an intelligent music preference evaluation model; Step 5: Through meta-learning, the intelligent volume adjustment model and the intelligent music preference evaluation model are integrated to build a smart Bluetooth speaker sub-control adjustment model; Step 6: Adjust the volume of the smart Bluetooth speaker according to the volume range of the smart Bluetooth speaker; recommend music to the user according to the user's music preferences.

Images