Music recommendation method and system based on multi-dimensional data

Abstract

This invention provides a music recommendation method based on multidimensional data. The method includes the following steps: acquiring multidimensional data of the target user, including physiological data, environmental data, user behavior data, and voice data; determining the target user's comprehensive emotional state and current activity type based on the user's multidimensional data; combining the target user's historical music listening behavior and current comprehensive emotional state, and generating a music preference model based on the user's historical music preference data in different scenarios; and generating a personalized recommended music list by fusing and analyzing the comprehensive emotional state, scenario activity type, and music preference model. This invention, by acquiring and comprehensively processing the target user's multidimensional data, can accurately identify the user's current emotional state and activity scenario, achieving a comprehensive perception of the user's emotions and environment.

Description

Technical Field

[0001] This invention relates to the field of music recommendation technology, and in particular to a music recommendation method based on multidimensional data. Background Technology

[0002] With the widespread application of smart wearable devices, large amounts of user physiological data (such as heart rate, blood pressure, and exercise trajectory) can be collected and processed in real time, making the acquisition of personal health and behavioral data more accurate and comprehensive. However, most existing music recommendation systems rely on traditional historical behavioral data (such as listening records and favorite tracks), failing to fully consider the user's current physiological state, environmental factors, and emotional changes. This results in recommendations that lack immediacy and accuracy, and cannot perfectly match the user's music needs in specific situations.

[0003] Currently, many music apps still face several major challenges in personalized recommendations. First, the lack of effective integration of multidimensional data between smart wearable devices and music apps creates data silos, preventing the use of user physiological and environmental data to further optimize music recommendations. Second, most existing systems rely on historical behavioral data for recommendations, lacking accurate identification of the user's current context and activities, potentially leading to music recommendations that don't match the user's current needs. Third, while sentiment analysis technology is used in some music apps, it's mostly limited to text sentiment analysis of user social media platforms or comments, failing to fully utilize real-time physiological data provided by smart wearable devices for emotional state assessment and lacking a dynamic emotional adaptation mechanism.

[0004] Therefore, how to effectively integrate multidimensional data, accurately identify user context and emotional changes, and recommend music based on this information has become a key technical issue in improving personalized music recommendation systems. Summary of the Invention

[0005] To address the problems in existing technologies, the present invention aims to provide a music recommendation method based on multidimensional data. By accurately capturing and integrating users' physiological data, environmental data, behavioral data, and voice data, the method dynamically identifies users' current emotional state and activity scenarios, thereby generating a more personalized and instantly adaptable music recommendation list.

[0006] A first aspect of this invention provides a music recommendation method based on multidimensional data, comprising the following steps:

[0007] Acquire multidimensional data of target users, including physiological data, environmental data, user behavior data, and voice data;

[0008] Based on multidimensional user data, determine the target user's overall emotional state and current activity type.

[0009] By combining the target user's historical music listening behavior and current overall emotional state, and generating a music preference model based on the user's historical music preference data in different scenarios;

[0010] Based on a comprehensive emotional state, scenario activity type, and music preference model, a personalized music recommendation list is generated through integrated analysis.

[0011] In some embodiments of the first aspect, the physiological data in the multidimensional data includes heart rate, blood pressure, sleep quality, and skin conductivity data; the environmental data includes GPS location information, temperature, light intensity, and noise level; the user behavior data includes steps, activity frequency, and device usage time; and the voice data includes voice emotion features and speech rate.

[0012] In some embodiments of the first aspect, the step of determining the overall emotional state includes:

[0013] Emotional state features are extracted from physiological data, and the physiological data are mapped to preliminary emotional states based on a pre-set emotion model.

[0014] Perform speech emotion recognition on speech data, extract speech emotion parameters and generate emotion labels;

[0015] By combining the initial emotional state and emotional labels, a final comprehensive emotional state is generated.

[0016] In some embodiments of the first aspect, determining the scene activity type includes:

[0017] Based on user behavior data, identify the current user's activity status, including sitting, walking, running, and driving;

[0018] Based on environmental data, identify whether the current scene is indoor, outdoor, or commuting environment;

[0019] By combining motion status and scene environment data, the type of scene activity of the target user can be determined.

[0020] In some embodiments of the first aspect, the step of generating a music preference model includes:

[0021] Based on the target users' historical music listening data in different activity scenarios, calculate the distribution of music style preferences in each scenario;

[0022] Based on the user's overall emotional state, extract historical music features that match the emotional state;

[0023] Generate a music preference model that includes the association between scene, emotion, and music features.

[0024] In some embodiments of the first aspect, the step of generating a personalized recommended music list includes:

[0025] Classify the current overall emotional state into emotional intensity levels;

[0026] Based on a scene-dependent music preference model, music that matches the current emotional intensity level and the type of scene activity is selected;

[0027] Based on preset sorting rules and the weight of users' historical preferences, a personalized music recommendation list is generated with priority sorting.

[0028] In some embodiments of the first aspect, the method further includes an emotional transition mechanism, comprising the following steps:

[0029] When the target user’s overall emotional state or the type of activity in the scenario changes, the target emotional state is determined based on the current overall emotional state.

[0030] Adjust the scene-dependent music preference model based on the difference between the target user's current emotional state and the target emotional state.

[0031] In some embodiments of the first aspect, adjusting the music selection of the scene-dependent music preference model further includes the following steps:

[0032] Music with gradually varying emotional intensity is recommended in sequence to match the target's emotional state;

[0033] The volume, rhythm, and melody of the recommended music are dynamically adjusted to match the preset emotional transition curve.

[0034] In some embodiments of the first aspect, the method further includes a real-time feedback and adaptive adjustment step:

[0035] Monitor user behavior feedback, including skipping, pausing, and manually selecting music;

[0036] Adjust the music preference model based on user feedback to optimize user experience and dynamically adjust the emotional transition mechanism.

[0037] A second aspect of this invention provides a music recommendation system based on multidimensional data, used to implement the aforementioned music recommendation method based on multidimensional data.

[0038] A third aspect of the present invention provides a music recommendation program product based on multidimensional data. The program product includes computer instructions that, when executed by a processor, implement the steps of the music recommendation method based on multidimensional data described above.

[0039] A fourth aspect of the present invention provides a music recommendation device based on multidimensional data, comprising:

[0040] processor;

[0041] Memory, which stores the processor's executable instructions;

[0042] The processor is configured to execute the steps of the music recommendation method based on multidimensional data described above by executing executable instructions.

[0043] The music recommendation method based on multidimensional data of the present invention has the following beneficial effects:

[0044] This invention acquires and comprehensively processes multidimensional data from target users to accurately identify their current emotional state and activity scenarios, achieving a comprehensive perception of user emotions and environment. By combining users' historical music listening behavior, a scenario-dependent music preference model is constructed, enabling the generation of personalized recommended music lists based on users' emotional states and activity types in dynamically changing scenarios. Attached Figure Description

[0045] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings.

[0046] Figure 1 This is a schematic diagram illustrating the implementation process of a music recommendation method based on multidimensional data according to an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram of the implementation process for determining the comprehensive emotional state of a target user according to an embodiment of the present invention;

[0048] Figure 3 This is a schematic diagram of the implementation process for determining the type of scene activity according to an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram illustrating the implementation process of a music preference generation model according to an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram illustrating the implementation process of generating a personalized music recommendation list according to an embodiment of the present invention;

[0051] Figure 6 This is a schematic diagram illustrating the implementation process of an emotional transition mechanism according to an embodiment of the present invention;

[0052] Figure 7 This is a schematic diagram illustrating the implementation process of adjusting the music selection of a scene-dependent music preference model according to an embodiment of the present invention;

[0053] Figure 8 This is a schematic diagram of the implementation process of real-time feedback and adaptive adjustment according to an embodiment of the present invention;

[0054] Figure 9 This is a schematic diagram of the structure of a music recommendation device based on multidimensional data according to an embodiment of the present invention. Detailed Implementation

[0055] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0056] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0057] The flowchart shown in the attached diagram is merely an illustrative example and does not necessarily include all steps. For example, some steps may be broken down, while others may be combined or partially combined. Therefore, the actual execution order may change depending on the specific circumstances.

[0058] This invention, based on the technical principles of multidimensional data analysis, aims to accurately identify and assess a user's emotional state and activity scenarios through real-time monitoring and comprehensive analysis of multiple data sources, including physiological, environmental, behavioral, and vocal data. This provides a more dynamic and flexible solution for personalized music recommendations. The invention utilizes real-time information from different data dimensions (such as physiological, environmental, behavioral, and vocal data) and, through comprehensive processing, captures dynamic changes in the user's emotional state and activity scenarios. Combined with the user's historical music listening behavior, it dynamically updates the personalized recommendation model, thereby improving the accuracy and adaptability of the recommendation system. Firstly, there is a close relationship between a user's emotional state and physiological signals. Physiological data such as heart rate variability (HRV), skin conductance, and respiratory rate effectively reflect a user's emotional fluctuations. Analysis of HRV signals reveals the regulatory role of the autonomic nervous system on cardiac function, thus reflecting changes in the user's emotional state. Emotional fluctuations are typically associated with psychological states such as anxiety, stress, and pleasure. Therefore, emotion recognition through physiological data can significantly improve the accuracy of emotion judgment and reflect user emotional changes in real time. This effectively addresses the shortcomings of existing music recommendation systems in capturing users' immediate emotional changes. Secondly, environmental data (such as temperature, humidity, light intensity, and noise level) is used in this invention to identify the user's current activity scenario. Environmental data provides the system with scenario context information, helping the system identify the user's specific activity scenario, such as exercise, work, leisure, or gatherings. Users' music needs and preferences may differ in different scenarios; therefore, analyzing environmental data ensures that music recommendations better match the user's current scenario needs, thereby improving the relevance and satisfaction of music recommendations. The comprehensive processing of physiological, environmental, behavioral, and voice data not only accurately identifies the user's emotional state but also comprehensively captures the user's activity background and environmental changes. Changes in emotional state are usually caused by multiple factors; therefore, relying solely on a single data source is insufficient to accurately determine a user's emotional fluctuations. Through comprehensive analysis of multidimensional data, this invention can compensate for the shortcomings of a single data source, improve the accuracy of emotion recognition, and better reflect the user's personalized needs. Furthermore, this invention also generates a scenario-dependent music preference model by combining the user's historical music listening behavior with their current emotional state and scenario activity type. A user's music preferences are influenced by various factors, including historical music listening behavior, emotional state, and activity scenario. This model can dynamically adjust the recommended music list based on the user's current mood and activity context, thus providing more personalized and contextualized music recommendations. This context-dependent music preference model can flexibly adjust the recommended content according to changes in real-time data, ensuring that music recommendations match the user's immediate needs.This invention, through real-time monitoring and comprehensive analysis of multi-dimensional data, can accurately identify and assess a user's emotional state and activity scenario, thereby generating a personalized music recommendation list, significantly improving the dynamic adaptability and personalization level of the recommendation system. Unlike traditional music recommendation systems that rely on static recommendations based on historical data, this invention can respond to changes in user emotions and scenarios in real time, providing more accurate and personalized recommendation services. This invention effectively solves the problems of inaccurate emotion and scenario recognition and lack of personalized recommendation results in existing music recommendation systems, significantly improving overall user satisfaction and user experience.

[0059] like Figure 1 As shown, this invention provides a music recommendation method based on multidimensional data. It aims to accurately identify and assess a user's emotional state and activity scenarios by real-time monitoring and comprehensive analysis of the user's multidimensional data (including physiological data, environmental data, user behavior data, and voice data), thereby generating a personalized music recommendation list. The following details the various technical features of this invention and their relevance to the technical problem it solves.

[0060] S100: Acquire multidimensional data of the target user, including physiological data, environmental data, user behavior data, and voice data.

[0061] In this embodiment of the invention, multidimensional data refers to data types from different sources, encompassing physiological data, environmental data, user behavior data, and voice data including features such as the content, tone, speed, and volume of the user's speech. Physiological data, such as heart rate, skin conductance, and respiratory rate, can effectively reflect the user's physiological state and emotional fluctuations. Environmental data, including temperature, humidity, and light intensity, can be used to determine the user's activity environment and scenario. User behavior data can cover the user's interactive behaviors, such as clicks, browsing, and playback history, while voice data reflects the user's linguistic emotions and interaction patterns. Through real-time acquisition from multiple data sources, this invention can comprehensively and accurately obtain the user's current emotions and activity background, avoiding the errors or deficiencies caused by a single data source in existing technologies.

[0062] In some optional embodiments, the physiological data in the multidimensional data includes heart rate, blood pressure, sleep quality, and skin conductivity data; environmental data includes GPS location information, temperature, light intensity, and noise level; user behavior data includes steps, activity frequency, and device usage time; and voice data includes voice emotion features and speech rate. The physiological, environmental, user behavior, and voice data in the multidimensional data are selected and processed according to the specific application scenario. This multidimensional data provides the present invention with multi-level, comprehensive user status information, and through multidimensional data fusion, helps the present invention accurately identify the user's emotional state and activity scenario.

[0063] Physiological data refers to various data that reflect a user's physiological state. Specifically, it includes, but is not limited to, heart rate, blood pressure, sleep quality, and skin conductivity data. Heart rate: Fluctuations in heart rate directly reflect a user's emotional state, especially under emotions such as stress, anxiety, or pleasure, where heart rate changes significantly. Therefore, heart rate data can be used to determine a user's emotional changes. Blood pressure: Blood pressure is closely related to emotional changes; for example, blood pressure often rises when a user feels tense or excited. Therefore, blood pressure data helps improve the accuracy of emotion recognition. Sleep quality: Sleep quality reflects a user's health status and their response to environmental or emotional changes. Insufficient sleep or poor sleep quality is usually associated with negative emotions or physical discomfort, which can further optimize the assessment of emotional state. Skin conductivity: Skin conductivity (GSR) is a physiological signal used to measure a user's emotional response (such as anxiety or excitement). Skin conductivity usually increases under stress or anxiety. Real-time monitoring of these physiological data provides the system with immediate feedback on the user's emotions, thereby helping the system accurately determine the user's current emotional state. By integrating physiological data, this embodiment of the invention can provide more accurate personalized music recommendations based on real-time perception of user emotional changes.

[0064] Environmental data includes GPS location information, temperature, light intensity, and noise level. This data helps the invention determine the user's environment and activity scenario. GPS location information: By analyzing the user's GPS location information, the system can identify whether the user is in a specific scenario (such as at home, in the office, or at a sports field) and recommend different types of music based on the needs of different scenarios. Temperature: Temperature data reflects the user's comfort level; in high or low temperatures, the user's mood may fluctuate. Therefore, temperature data can provide auxiliary information for sentiment analysis. Light intensity: Light intensity is closely related to the user's activity and psychological state. For example, in a dark environment, the user may be relaxed or leisurely, while in a bright environment, the user may be working or exercising. By observing changes in light intensity, the system can determine the user's activity type and adjust music recommendations accordingly. Noise level: Noise level affects the user's emotional state and activity type. For example, a high noise level may mean the user is participating in a more active social activity or working in a noisy environment; the system can adjust its recommendation strategy accordingly. The introduction of environmental data effectively improves the system's ability to identify user activity scenarios and helps generate music recommendations that match the user's current scenario.

[0065] User behavior data includes steps, activity frequency, and device usage time. This data helps the system further understand the user's activity intensity and daily habits. Steps: Step count reflects the user's daily activity level. By monitoring steps, the system can determine whether the user is exercising and recommend exercise-related music, such as upbeat or high-energy music. Activity frequency: Activity frequency data helps determine the user's activity patterns. If a certain type of activity occurs frequently, such as frequent gym workouts or outings, the system can dynamically adjust its recommendation strategy based on these habits. Device usage time reflects the user's dependence on the device and can also indicate the user's emotional state. Prolonged device use may indicate that the user is in a high-intensity work or leisure state, and the system can optimize recommendation results based on this information. This behavioral data enables the system to accurately identify user activity patterns and generate personalized music recommendations based on these patterns, thereby improving the user experience.

[0066] Voice data, including emotional features and speech rate, is a crucial way for users to express their emotions. Emotional features: By analyzing characteristics such as pitch, volume, and tone of voice, the system can identify changes in a user's emotions. Sentiment analysis uses machine learning algorithms (such as deep learning or sentiment analysis models) to process a user's speech and determine the emotions behind it. Speech rate: Changes in speech rate can also reflect a user's emotions. For example, a faster speech rate may indicate anxiety or excitement, while a slower speech rate may indicate relaxation or contemplation. Speech rate data can provide auxiliary information on changes in emotional state, thereby improving the accuracy of emotion recognition. The introduction of voice data not only enhances the dimensions of emotion recognition but also helps the system understand the user's immediate emotional state, thus optimizing the real-time performance and accuracy of personalized music recommendations.

[0067] Through real-time monitoring and comprehensive analysis of the aforementioned multi-dimensional data, this embodiment of the invention can accurately capture users' emotional fluctuations and changes in activity scenarios. The fusion analysis of physiological data, environmental data, user behavior data, and voice data can effectively solve the problem that existing music recommendation systems cannot fully perceive users' emotions and activity scenarios, significantly improving the adaptability and personalization level of the recommendation system.

[0068] S200. Based on the user's multidimensional data, determine the target user's overall emotional state and the type of activity in the current scenario.

[0069] This invention, through comprehensive analysis of multi-dimensional data, employs emotion recognition algorithms and scene recognition models to infer a user's emotional state and current activity scenario in real time. Emotional state includes the user's psychological and physiological reactions, such as pleasure, anxiety, and stress. Activity scenario encompasses various situations the user may be in, such as exercise, work, and leisure. By comprehensively evaluating emotional state and scenario type, this invention addresses the shortcomings of existing music recommendation systems that cannot capture real-time changes in user emotions and activity scenarios. Through dynamic emotion and scenario recognition, recommendation strategies can be adjusted in real time, making the recommendations more aligned with the user's current needs.

[0070] S300 combines the target user's historical music listening behavior and current overall emotional state, and generates a music preference model based on the user's historical music preference data in different scenarios.

[0071] In this embodiment of the invention, a scene-dependent music preference model is generated by combining historical behavioral data of users in different scenarios (such as past music listening records) with their current emotional state and activity scenario. This model not only considers changes in the user's emotions but also incorporates differences in user preferences across different scenarios. For example, a user may prefer faster-paced music while exercising, while preferring more relaxing melodies during leisure time. Through a dynamically updated scene-dependent music preference model, this invention can adjust the recommendation strategy based on the user's real-time emotions and scene changes, thereby generating more personalized music recommendations.

[0072] S400: Based on a comprehensive emotional state, scenario activity type, and music preference model, a personalized music recommendation list is generated through fusion analysis.

[0073] In this embodiment of the invention, a personalized music recommendation list is generated by fusing and analyzing the user's emotional state, scene activity type, and scene-related music preference model. This fusing analysis not only considers immediate changes in emotion and scene but also incorporates music preferences from historical behavioral data, forming a multi-dimensional recommendation decision model. Through this comprehensive analysis, the present invention effectively overcomes the limitations of existing technologies that rely solely on historical user behavior data and lack dynamic adaptability, significantly improving the real-time performance and personalization of music recommendations.

[0074] This invention can dynamically identify a user's emotional fluctuations and activity scenarios, and combine these with the user's historical music preferences to generate accurate personalized recommendation lists. Unlike traditional music recommendation systems based on historical data, this invention can respond to users' emotional changes and activity scenarios in real time, ensuring that the recommendation results better match the user's current actual needs. This improves the personalization, adaptability, and real-time performance of the music recommendation system, enhances the user's music experience, and increases user satisfaction.

[0075] In some optional embodiments, step S200 determines the target user's overall emotional state and current activity type based on the user's multidimensional data. For example... Figure 2 As shown, the steps to determine the overall emotional state of the target user include:

[0076] S211. Extract emotional state features from physiological data and map the physiological data to preliminary emotional states based on a preset emotional model.

[0077] In this embodiment of the invention, emotional state features are first extracted from the user's physiological data. Physiological data typically includes, but is not limited to, heart rate, skin conductance, and respiratory rate, as these physiological signals reflect changes in the user's emotions. In the field of emotion analysis, physiological signals (such as heart rate variability (HRV) and skin conductance) are closely related to emotional fluctuations. Emotional state feature extraction from physiological data can be achieved by performing feature engineering on physiological signals, such as frequency domain analysis, time domain analysis, and nonlinear analysis, to extract representative emotion-related features. For example, analyzing the frequency domain components of heart rate variability (HRV) signals can reveal the regulatory role of the autonomic nervous system on cardiac function, thereby inferring the user's emotional fluctuations. Emotional fluctuations are usually directly related to psychological states such as anxiety, stress, and pleasure. Skin conductance can identify emotions such as anxiety or excitement through changes in skin conductance (GSR) signals. Therefore, through the extraction and analysis of physiological data, a relatively objective preliminary assessment of emotional state is provided.

[0078] S212. Perform speech emotion recognition on the speech data, extract speech emotion parameters and generate emotion labels.

[0079] In this embodiment of the invention, speech emotion recognition is performed on speech data to extract speech emotion parameters and generate emotion tags. Speech, as a direct carrier of user emotions, can effectively infer user emotions by analyzing characteristics such as pitch, speech rate, volume, and timbre. Speech emotion recognition technology is typically based on machine learning or deep learning algorithms, such as convolutional neural networks (CNNs) or long short-term memory networks (LSTMs), to classify and recognize speech features through model training, thereby generating speech emotion tags. In this embodiment, speech emotion recognition analyzes the emotional features in speech (such as emotion category and emotion intensity) to convert the user's speech into emotion tags with semantic information. These tags may include, but are not limited to, emotions such as happiness, sadness, anger, anxiety, and relaxation. By supplementing emotional information from multidimensional data, speech emotion recognition can overcome the limitations that may arise from relying solely on physiological data, thus providing a more comprehensive emotion assessment.

[0080] S213. Integrate the initial emotional state and emotional labels to generate the final comprehensive emotional state.

[0081] In this embodiment of the invention, the emotional data obtained in steps S211 and S212 are fused and analyzed to generate a final comprehensive emotional state. Specifically, the system combines the preliminary emotional state extracted based on physiological data with the emotional tags generated based on speech emotion recognition to comprehensively analyze the user's emotional expression. The fusion method can employ techniques such as weighted averaging, logistic regression, and multimodal deep learning. The weighted averaging method assigns different weights to different data sources, weighting and combining physiological data and speech emotion recognition results to obtain a comprehensive score. This score can effectively integrate physiological and emotional information to arrive at a more accurate comprehensive emotional state. The multimodal deep learning model can directly fuse data from different sources (such as physiological signals and speech features) at the feature level, optimizing the fusion strategy and generating the final emotional assessment through automatic learning of the network structure.

[0082] The generation of comprehensive emotional states does not merely consider a single data source (such as isolated physiological or vocal data), but rather maximizes the accuracy and reliability of emotion analysis through the fusion of multi-dimensional data. Therefore, the technical principle of this step, through multi-source information fusion, can more comprehensively and accurately reflect the user's actual emotional state, providing more precise input for subsequent music recommendations. Through comprehensive analysis of physiological and vocal data, this invention can accurately and in real-time assess the user's comprehensive emotional state. Physiological data (such as heart rate and skin conductance response) provides a stable and objective basis for emotion judgment, while vocal data, through real-time capture of the user's emotional expression, compensates for the inadequacy of single physiological data in fully revealing the user's emotions. By fusing this information, this invention overcomes the limitations of traditional emotion recognition methods that rely on a single data source, significantly improving the accuracy and comprehensiveness of emotion assessment.

[0083] In some optional embodiments, step S200 determines the target user's overall emotional state and current activity type based on the user's multidimensional data. For example... Figure 3 As shown, the steps for determining the scene activity type include:

[0084] S221. Based on user behavior data, identify the current user's movement status, including sitting, walking, running, and driving.

[0085] In this embodiment of the invention, the current user's movement state is identified based on user behavior data. User behavior data includes information such as step count, activity frequency, exercise duration, and device usage time. This data typically comes from sensors in smart wearable devices or mobile devices. By analyzing this behavior data, it is possible to identify whether the user is currently in different movement states such as sitting, walking, running, or driving. Sitting: When the user's behavior data indicates that they have remained stationary for an extended period, such as being inactive or sitting, the system identifies them as sitting. Walking: When the user's step count or activity frequency data indicates that they are engaged in low-intensity exercise such as walking, the system identifies them as walking. Running: By identifying the user's step frequency, acceleration, or exercise intensity data, the system determines whether the user is running. Driving: By identifying the user's device's GPS data and movement speed, combined with vehicle characteristics (such as vehicle speed and acceleration), it determines whether the user is driving. This embodiment of the invention can accurately identify the user's current movement state, providing data support for subsequent contextual analysis. The technical principle of this step lies in identifying the user's movement type through behavior data (such as step count, acceleration, etc.), thereby laying the foundation for further determination of the scene activity type.

[0086] S222. Based on environmental data, identify whether the current scene is indoor, outdoor, or commuting environment.

[0087] In this embodiment of the invention, the type of environment in which the user is located is identified by analyzing environmental data. Environmental data includes temperature, humidity, light intensity, noise level, GPS location information, etc. Based on this data, the system can determine whether the user is in an indoor environment, an outdoor environment, or a commuting environment. Indoor environment: When the light intensity in the environmental data is weak and the temperature and humidity are relatively stable, the system can determine that the user is in an indoor environment. Outdoor environment: When the light intensity is high and the temperature fluctuates significantly, the system determines that the user is in an outdoor environment. Commuting environment: Combining GPS location information and motion status data, when the system detects that the user's movement trajectory matches a commuting route and the movement status is walking or driving, the system determines that the user is in a commuting environment. This step determines the type of environment the user is in by analyzing multiple dimensions of the environmental data (such as light intensity, temperature, humidity, etc.), providing necessary environmental background information for determining the type of activity in the scene.

[0088] S223. Based on the comprehensive motion status and scene environment data, determine the scene activity type of the target user.

[0089] In this embodiment of the invention, the system ultimately determines the user's activity type by combining the motion state in S221 and the environmental data in S222. By combining the motion state and environmental data, the system can determine whether the user is in a sports scene, work scene, leisure scene, etc. For example, if the user is walking or running in an outdoor environment, the system may determine that they are in a sports scene; if the user is sitting still in an indoor environment, the system may determine that they are engaged in work or leisure activities; if the user is in a car and their motion state is driving, the system identifies it as a commuting scene. Through this comprehensive analysis, the system can accurately identify the specific activity scene of the user, providing a more accurate basis for subsequent personalized music recommendations.

[0090] By comprehensively analyzing user behavior data and environmental data, this invention can accurately determine the current activity type of a target user in real time, thereby ensuring that the recommendation system can provide more personalized music recommendations based on the user's actual activities. Through comprehensive analysis of multi-dimensional data, the system can identify and determine the user's activity scenario in real time, solving the problem of inaccurate scenario activity type identification in existing technologies, thus improving the adaptability and accuracy of the music recommendation system. Through precise scenario identification, the system can dynamically adjust its recommendation strategy, making music recommendations more aligned with the user's current contextual needs, greatly enhancing the user experience and the system's personalization level.

[0091] In some optional embodiments, step S300 combines the target user's historical music listening behavior and current overall emotional state, and generates a music preference model based on the user's historical music preference data in different scenarios. By analyzing the user's historical behavioral data, emotional state, and scenario activity type, a multi-dimensional model reflecting the user's music preferences is established. Figure 4 As shown, the steps for generating a music preference model include:

[0092] S310. Calculate the distribution of music style preferences in each scenario based on the target user's historical music listening data under different activity types.

[0093] In this embodiment of the invention, based on the target user's historical music listening data under different activity scenarios, the system calculates the music style preference distribution for each scenario. The user's historical music listening data includes information such as the music type, genre, and playback frequency selected by the user at different times, locations, and scenarios. Scenario identification: Based on the scenario activity type (e.g., sports, work, leisure) identified by the above method, the system categorizes the user's music listening behavior in each specific scenario. For example, in a sports scenario, the user may prefer fast-paced, energetic electronic music, while in a leisure scenario, they may prefer soft jazz or classical music. Music style preference distribution: By analyzing the user's music listening behavior in different scenarios, the system can statistically determine the proportion of music styles preferred by the user in each scenario. For example, in a sports scenario, fast-paced styles such as electronic and rock have a high proportion, while in a leisure scenario, light music and classical music have a larger proportion. This preference distribution helps to provide users with music recommendations that better match their scenario needs. This step, through the analysis of historical listening data, forms the music preference distribution for different scenarios and provides a basis for establishing a music preference model.

[0094] S320. Based on the user's overall emotional state, extract historical music features that match the emotional state.

[0095] In this embodiment of the invention, based on the target user's current comprehensive emotional state, historical music features matching this emotional state are further analyzed and extracted. The comprehensive emotional state (such as anxiety, pleasure, fatigue, stress, etc.) is obtained through the emotion recognition process in step S200. Emotion and music matching: Different emotional states are typically matched with specific music styles or emotional characteristics. For example, when a user feels anxious, they may prefer more soothing music, such as gentle piano pieces; while when a user is in a pleasant mood, they may prefer pop or rock music with a strong rhythm and lively atmosphere. Historical music feature extraction: The system extracts music features (such as pitch, rhythm, melody, and lyrical emotion) related to the user's current emotional state based on the user's historical music listening data. For example, under a pleasant mood, a user may tend to listen to music with positive emotional characteristics, such as cheerful and relaxing styles; while under stress, a user may prefer deep and relaxing music styles. Through this step, an association is established between emotional state and music features, laying the foundation for generating music recommendations that match the user's current emotional state.

[0096] S330. Generate a music preference model that includes the association between scene, emotion, and music features.

[0097] In this embodiment of the invention, the music style preference distribution, emotion-matching music features, and scene activity types from steps S310 and S320 are integrated to generate a music preference model that incorporates the association between scene, emotion, and music features. Multi-dimensional model fusion: This model combines the user's music style preferences in different scenarios, music needs under the current emotional state, and historical music features to provide users with more personalized music recommendations. Specifically, this model can dynamically adjust the recommendation strategy in different situations (such as exercise, leisure, work, etc.) while taking into account the user's emotional changes, making the recommendation results more in line with the user's immediate needs. Model updates and dynamic adjustments: As the user's emotional state, activity scenario, and music listening behavior change, the preference model is continuously updated to ensure that music recommendations always maintain high accuracy and personalization. Whenever new historical music data is acquired or a change in the user's emotional state is detected, the model will make corresponding adjustments. Through this music preference model, customized music recommendation lists can be generated for users in different situations, thereby improving the relevance of recommendations and user satisfaction.

[0098] This invention comprehensively analyzes a target user's historical music listening data, current emotional state, and activity scenarios to construct a comprehensive, multi-dimensional music preference model. This model not only considers the user's historical music preferences but also dynamically adapts to the user's emotional fluctuations and different scenario needs, thereby providing personalized and accurate music recommendations. This significantly improves the intelligence and dynamic adaptability of music recommendation systems, overcomes problems such as inaccurate emotion recognition and lack of personalization in existing technologies, and further optimizes the user's music experience.

[0099] In some optional embodiments, step S400 generates a personalized music recommendation list based on a comprehensive emotional state, scenario activity type, and music preference model. Through a comprehensive analysis of emotional intensity levels, activity scenarios, and music preferences, music that best matches the user's current needs is accurately recommended, enhancing the user experience. Figure 5 As shown, the steps to generate a personalized music recommendation list include:

[0100] S410. Classify the current overall emotional state into emotional intensity levels.

[0101] In this embodiment of the invention, the current comprehensive emotional state of the target user is first classified into emotional intensity levels. Through analysis of multi-dimensional data such as the user's physiological data and voice data, the system can obtain a real-time comprehensive emotional state, such as anxiety, pleasure, stress, and fatigue. However, the intensity of these emotional states is not the same, so they need to be further classified into different intensity levels. For example, when the user's emotional state is "anxious," the system classifies the anxiety as mild, moderate, or severe based on the intensity of physiological signals such as heart rate fluctuations and skin conductance. Different intensities of emotional states may require different styles of music for adjustment. For example, for severe anxiety, more relaxing and soothing music may be recommended, while for mild anxiety, faster-paced and motivating music can be chosen. Emotional intensity level classification helps to accurately adjust the recommended music type and style in subsequent steps to ensure that the recommendation results highly match the user's actual emotional needs.

[0102] S420: Based on a scene-dependent music preference model, music that matches the current emotional intensity level and scene activity type is selected.

[0103] In this embodiment of the invention, music that meets the criteria is selected based on a scene-dependent music preference model, the current emotional intensity level, and the type of scene activity. Specifically, suitable music is selected based on the user's current activity scene (such as exercise, work, leisure, etc.) and emotional intensity level (such as mild pleasure, severe anxiety, etc.). The matching of scene activity type and emotional intensity limits the recommended music style based on the user's current activity type (such as exercise, office, family leisure, etc.). For example, in an exercise scene, fast-paced, uplifting electronic music, rock music, etc., may be recommended; in a work scene, relaxing music such as light music, classical music, etc., may be recommended. At the same time, the system considers the current emotional intensity level: if the emotion is relatively anxious, even in an exercise scene, relatively soft music will be recommended to avoid further aggravation of the emotion. The music selection criteria, based on the scene-dependent music preference model and the requirements of emotional intensity, select music that matches the current emotional intensity and scene activity type. This step, through the fusion analysis of multi-dimensional data, ensures that the music recommendation can accurately adapt to the user's real-time needs.

[0104] S430. Generate a personalized recommended music list based on the user's historical preference weights and according to the preset sorting rules.

[0105] In this embodiment of the invention, a personalized music recommendation list is generated based on a preset sorting rule and the user's historical music preference weights. Building upon the matching of emotions and scenarios, the sorting of the recommendation results is further optimized to ensure that the user obtains music that best suits their needs. Historical preference weights assign a weight value to each piece of music based on the user's historical music behavior data (such as songs the user has frequently played, music styles, etc.). Music genres or specific tracks that the user prefers will occupy a higher priority in the recommendation list. The sorting rule sorts the music according to certain rules (e.g., prioritizing music that matches the intensity of emotions and the scenario, or prioritizing music that the user frequently listens to based on historical preference values). This sorting rule ensures that the recommended music not only meets the current emotional and scenario needs but also better satisfies the user's personalized preferences. Finally, a personalized music recommendation list is generated, where music is sorted according to its degree of matching with the current emotion and scenario, the user's historical preferences, and the priority of emotional intensity. This recommendation list will be provided to the user so that they can quickly obtain music content that meets their needs in different scenarios and emotional states.

[0106] This invention enables a comprehensive analysis of a user's emotional intensity, activity context, and music preferences to generate a personalized and accurate music recommendation list. Through multi-dimensional analysis and optimization of emotional intensity, scenario type, and historical preferences, it provides users with more accurate and personalized music recommendations, significantly improving their music experience and satisfaction.

[0107] In some optional embodiments, the present invention further includes an emotional transition mechanism. When the overall emotional state or activity type of the target user changes, the music recommendation model is adjusted according to the current emotional state, thereby ensuring that the recommended music always remains consistent with the user's real-time needs and emotional fluctuations. Figure 6 As shown, the emotional transition mechanism includes the following steps:

[0108] S510. When the target user's overall emotional state or the type of activity in the scenario changes, the target emotional state is determined based on the current overall emotional state.

[0109] In this embodiment of the invention, the system first detects whether the target user's overall emotional state or the type of scene activity has changed. When a change is detected, the system further determines the target emotional state based on the current emotional data and scene information. Specifically, when the target user's emotional state fluctuates, the system judges the current user's emotional state based on real-time physiological signals (such as heart rate, skin conductance response, etc.), voice features (such as voice emotion), behavioral data (such as activity frequency, step count), and environmental data (such as temperature, light intensity, etc.), and determines the target emotional state based on this information. In this embodiment of the invention, changes in overall emotional state are usually closely related to the user's physiological reactions, voice performance, and behavioral characteristics. For example, an increased heart rate or increased skin conductance response may indicate that the user is in a state of anxiety or tension. By analyzing this data, the system can accurately identify emotional changes and update the target emotional state in a timely manner. Changes in the type of scene activity also affect emotional state. For example, changing from a quiet indoor environment to a noisy outdoor environment may change the user's emotional reaction, thereby affecting music recommendations. The system re-evaluates the user's emotional state based on changes in the scene so that these effects are fully considered when recommending music. By monitoring changes in emotion and context in real time, step S510 ensures accurate identification of emotional states, providing reliable basic data for subsequent adjustments to the music preference model and music recommendations.

[0110] S520. Adjust the scene-dependent music preference model based on the difference between the target user's current emotional state and the target emotional state.

[0111] In this embodiment of the invention, the scene-dependent music preference model is adjusted based on the difference between the target user's current emotional state and the target emotional state. Specifically, the difference between the target emotional state and the current emotional state usually reflects the user's emotional fluctuations or changes in needs. This requires the system to dynamically adjust the music recommendation model to adapt to the user's real-time emotional needs. Changes in emotional state may involve multiple aspects, including the type of emotion (e.g., from pleasure to anxiety) and the intensity of emotion (e.g., from mild anxiety to severe anxiety). For example, when a user's emotion changes from pleasure to anxiety, the system will automatically adjust the music recommendation model based on this change, selecting music types more suitable for relieving anxiety, such as light music or relaxing music. The process of adjusting the scene-dependent music preference model includes: re-evaluating the music style preferences in each scene based on changes in the user's emotional state. For example, if the user's emotion changes from anxiety to pleasure, the system will increase the recommendation weight of upbeat and energetic music styles; while when the emotion changes from pleasure to tension, the system will tend to recommend soothing and relaxing music styles. By adjusting the music recommendation model in a timely manner when emotions fluctuate, step S520 ensures that the personalized recommendation system can continuously provide users with music recommendations that meet their immediate needs, improving the user experience.

[0112] This invention, through real-time monitoring of users' emotional states and activity scenarios, enables the system to accurately identify emotional fluctuations or scene transitions in target users and respond immediately. This allows the recommendation system to more accurately reflect users' immediate needs, avoiding the problem of traditional systems failing to capture emotional changes or scene shifts in a timely manner. By adjusting the scene-dependent music preference model based on emotional differences, the system can dynamically optimize the music recommendation list. This adjustment mechanism effectively improves the accuracy of personalized recommendations, allowing users to receive music recommendations that better match their emotional state and activity scenario, increasing the relevance of recommendations and user satisfaction. By sensing and adapting to users' emotional changes in real time, the system can provide users with more personalized and emotionally resonant music recommendations. Regardless of how a user's emotions fluctuate, the recommendation system can provide just the right music, improving the user's listening experience and satisfaction.

[0113] In some optional embodiments, step S520 adjusts the scene-dependent music preference model based on the difference between the target user's current emotional state and the target emotional state, to ensure that the recommended music not only matches the user's current mood but also helps the user transition between emotions through changes in musical elements. For example... Figure 7 As shown, adjusting the music selection of the scene-dependent music preference model also includes the following steps:

[0114] S521. Recommend music with gradually changing emotional intensity in sequence to match the target emotional state.

[0115] In this embodiment of the invention, music with an emotional intensity gradually approaching the target emotional state is recommended based on the difference between the target user's current emotional state and the target emotional state. Through this gradual adjustment of emotional intensity, the recommended music not only matches the user's current emotional state but also guides the user's emotions smoothly to the target emotional state. Emotional intensity is an important dimension of emotional state, typically related to the intensity of emotion and the strength of emotional expression. For example, music with high emotional intensity usually has a strong rhythm, a powerful melody, or a high volume, while music with low emotional intensity may be calmer and gentler. In this process, the target user's emotional intensity is analyzed, and music with corresponding emotional intensity is selected from the audio library. Based on the difference between the current and target emotional states, music with an emotional intensity gradually approaching the target emotional state is recommended. For example, when a user's mood shifts from anxiety to relaxation, music with high emotional intensity is initially recommended to help the user transition, and then the emotional intensity of the recommended music is gradually adjusted to match the user's target emotional state.

[0116] S522. Dynamically adjust the volume, rhythm, and melody of the recommended music to match the preset emotional transition curve.

[0117] In this embodiment of the invention, the emotional transition curve refers to a mechanism that dynamically adjusts the emotional characteristics of music (such as volume, rhythm, and melody) according to changes in the target user's emotional state to achieve a smooth and natural emotional transition. Specifically, the emotional transition curve reflects the pattern of changes in the user's emotional state, and by adjusting the music, it enables the music to adapt to and support these changes when the user experiences emotional shifts, thereby achieving a better emotional regulation effect. Through dynamic adjustment of the music's volume, rhythm, and melody, the system can precisely control the emotional expression of the music during the emotional transition process, thereby optimizing the user's emotional experience and helping the user transition to the target emotional state more naturally. Volume is one of the important parameters of emotional expression. During the emotional transition process, gradual changes in volume can regulate the intensity of the emotion. For example, during the transition from tension to relaxation, the system will gradually reduce the volume to help the user relieve stress and guide them into a relaxed state. Rhythm has a crucial impact on emotions. Intense rhythms are often associated with high-intensity emotional states, such as anxiety or excitement, while gentle rhythms help to soothe tension and guide the user into a relaxed state. In this invention, the system adjusts the tempo of the music according to the target emotional state to adapt to the user's emotional changes. Melodic variation is also a key factor in emotional transition. In this embodiment, the system dynamically adjusts the fluctuations and complexity of the melody according to a preset emotional transition curve. For example, when a user's emotion shifts from a low point to a high point, the system increases the variation in the melody to make it more dynamic; while when the emotion smoothly transitions from a high point to a low point, the melody tends to be simpler and more peaceful. Through dynamic adjustments to volume, rhythm, and melody, the system can accurately grasp the emotional transition process and guide the user to the target emotional state.

[0118] This invention, through recommending music with emotional intensities gradually approaching the target emotional state and dynamically adjusting volume, rhythm, and melody, smoothly guides the user's emotional transition, avoiding discomfort caused by abrupt emotional changes. It helps users adapt to different emotional states more naturally and comfortably, improving their emotional regulation abilities. By adjusting the emotional intensity and musical elements of the music based on the difference between the target and current emotional states, the recommended music better meets the user's immediate needs, thereby enhancing the accuracy and relevance of the recommendations. Compared to traditional music recommendation systems, this invention allows for fine-tuning based on real-time emotional changes, significantly improving the personalization of recommendations. By adjusting various music parameters (such as volume, rhythm, and melody) in real time, it provides users with more nuanced emotional guidance and a personalized music experience. Users experience more pleasant music recommendations that match their emotional state, improving overall user satisfaction.

[0119] In some optional embodiments, the present invention further includes a real-time feedback and adaptive adjustment mechanism to further improve the effectiveness of the personalized music recommendation system. For example... Figure 8 As shown, real-time feedback and adaptive adjustment include the following steps:

[0120] S610 monitors user behavior feedback, including skipping, pausing, and manually selecting music.

[0121] In this embodiment of the invention, user behavior feedback on recommended music is monitored and recorded in real time, mainly including operations such as skipping, pausing, and manually selecting music. By capturing user behavior when using music recommendations, the system identifies the user's acceptance level or dissatisfaction with the currently recommended music, thus providing a basis for subsequent model adjustments. When a user is dissatisfied with the recommended music, they may choose to skip the current song. Skipping usually reflects a user's lack of interest in the music content or that it does not meet their emotional needs. By detecting skipping behavior, it is possible to identify music characteristics that do not meet the user's expectations, such as emotional intensity, melody style, and rhythm, thereby providing data support for optimizing recommendations. A user pausing music may indicate a decrease in interest in the current music or a need to adjust their current emotional state. By monitoring pausing operations, it is possible to infer changes in the user's emotional state; for example, when experiencing drastic emotional fluctuations, a user may choose to pause music to alleviate their emotions. A user manually selecting music means that the recommended music does not fully meet their current needs, and the user has chosen music they prefer. By recording this behavior, it is possible to further analyze the user's preferences and emotional needs, and use this information to adjust subsequent recommendation models.

[0122] S620: Adjust the music preference model based on user feedback to optimize user experience and dynamically adjust the emotional transition mechanism. Adjust the recommendation mechanism in real time based on immediate user feedback, enabling music recommendations to better match users' current needs and ensuring smoother and more accurate emotional transitions.

[0123] In this embodiment of the invention, the association model between user's historical behavior and emotional state and music preferences is adjusted based on user feedback. For example, if a user frequently skips a certain type of music, the system will identify a mismatch between that type of music and the user's emotional needs, and gradually reduce the recommendation probability of that type of music while increasing the recommendation frequency of other music types that match the user's preferences. The emotional transition mechanism involves changes in music under different emotional states, including adjustments to volume, rhythm, and melody. When the system identifies, based on user feedback, that the current emotional transition mechanism is failing to achieve the expected effect, it will automatically adjust the emotional transition strategy of the music. For example, if a user frequently pauses the music, the system may determine that the emotional transition curve is too abrupt or does not meet the user's needs, and then adjust the transition rhythm to make the music transition smoother and more in line with the user's emotional state changes.

[0124] This invention enables music recommendations to adapt to real-time changes by monitoring and dynamically adjusting user behavior feedback. It automatically optimizes recommended content and emotional transition strategies based on users' immediate needs, significantly improving the accuracy and adaptability of personalized recommendations. By capturing user operational feedback and adjusting the recommendation model accordingly, it allows for more precise adjustments based on user preferences and emotional states, thereby enhancing the user experience. Users experience music recommendations that better match their needs and emotions, increasing user satisfaction and engagement. Based on user behavior feedback, the emotional transition mechanism can be finely adjusted. Regardless of whether a user's emotion is rising, falling, or remaining stable, elements such as emotional intensity, rhythm, and volume in music recommendations can be adjusted in real-time to make emotional transitions more natural and smooth, thus improving the user's emotional regulation experience. Through real-time feedback on user behavior and analysis of historical data, it can more accurately capture users' personalized needs. Different users have different emotional responses to music; the real-time feedback mechanism can promptly understand these differences and dynamically adjust the recommendation results, avoiding outdated recommendations and ensuring that music recommendations always meet the user's true needs. Dynamic adjustment of the music emotional transition mechanism ensures the adaptability of music during emotional transitions. For example, when a user's emotions change significantly, the system can adjust the emotional intensity of the music to make the transition smoother and reduce the discomfort caused by emotional fluctuations.

[0125] This invention also provides a music recommendation system based on multidimensional data, used to implement the aforementioned music recommendation method based on multidimensional data. The functionality of the music recommendation system based on multidimensional data of this invention can be implemented using the specific embodiments of the music recommendation method based on multidimensional data described above, which will not be elaborated upon here.

[0126] This invention also provides a music recommendation system based on multidimensional data, a music recommendation program product based on multidimensional data, including computer-executable instructions, which, when executed by a processor, implement the aforementioned music recommendation method based on multidimensional data.

[0127] This application also provides a music recommendation device based on multidimensional data, including a processor; a memory storing executable instructions of the processor; wherein the processor is configured to execute steps of a music recommendation method based on multidimensional data by executing the executable instructions.

[0128] Those skilled in the art will understand that various aspects of this application can be implemented as a system, method, or program product. Therefore, various aspects of this application can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, collectively referred to herein as a "circuit," "module," or "platform."

[0129] The following reference Figure 9 To describe an electronic device 800 according to this embodiment of the present application. Figure 9 The electronic device 800 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0130] like Figure 9 As shown, the electronic device 800 is presented in the form of a general-purpose computing device. The components of the electronic device 800 may include, but are not limited to: at least one processing unit 810, at least one storage unit 820, a bus 830 connecting different system components (including storage unit 820 and processing unit 810), a display unit 840, etc.

[0131] The storage unit stores program code that can be executed by the processing unit 810, causing the processing unit 810 to perform the steps described in the above-described section of this specification regarding the music recommendation method based on multidimensional data, according to various exemplary embodiments of this application. For example, the processing unit 810 can perform actions such as... Figure 1 The steps are shown in the figure.

[0132] The storage unit 820 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 8201 and / or a cache storage unit 8202, and may further include a read-only memory unit (ROM) 8203.

[0133] The storage unit 820 may also include a program / utility 8204 having a set (at least one) program module 8205, such program module 8205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0134] Bus 830 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0135] Electronic device 800 can also communicate with one or more external devices 890 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 800, and / or with any device that enables electronic device 800 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 850. Furthermore, electronic device 800 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 860. Network adapter 860 can communicate with other modules of electronic device 800 via bus 830. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0136] In the music recommendation device based on multidimensional data, when the program in the memory is executed by the processor, it implements the steps of the music recommendation method based on multidimensional data. Therefore, the device can also obtain the technical effects of the music recommendation method based on multidimensional data.

[0137] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of this application and should not be construed as limiting the specific implementation of this application to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of this application, and all such modifications or substitutions should be considered within the scope of protection of this application.

Claims

1. A method of music recommendation based on multi-dimensional data, characterized by, Includes the following steps: Acquire multidimensional data of target users, including physiological data, environmental data, user behavior data, and voice data; Based on multidimensional user data, determine the target user's overall emotional state and current activity type. By combining the target user's historical music listening behavior and current overall emotional state, and generating a music preference model based on the user's historical music preference data in different scenarios; Based on a comprehensive emotional state, scenario activity type, and music preference model, a personalized music recommendation list is generated through integrated analysis.

2. The music recommendation method based on multidimensional data according to claim 1, characterized in that, The physiological data in the multidimensional data includes heart rate, blood pressure, sleep quality, and skin conductivity data; the environmental data includes GPS location information, temperature, light intensity, and noise level; the user behavior data includes steps, activity frequency, and device usage time; and the voice data includes voice emotion features and speech rate.

3. The music recommendation method based on multidimensional data according to claim 2, characterized in that, The steps to determine the overall emotional state of the target user include: Emotional state features are extracted from the physiological data, and the physiological data are mapped to preliminary emotional states based on a preset emotion model. The speech data is subjected to speech emotion recognition, speech emotion parameters are extracted and emotion tags are generated; The initial emotional state and emotional label are combined to generate the final comprehensive emotional state.

4. The music recommendation method based on multidimensional data according to claim 2, characterized in that, The determination of the scenario activity type includes: Based on user behavior data, identify the current user's activity status, including sitting, walking, running, and driving; Based on environmental data, identify whether the current scene is indoor, outdoor, or commuting environment; Based on the combined motion state and scene environment data, the scene activity type of the target user is determined.

5. The music recommendation method based on multidimensional data according to claim 1, characterized in that, The steps for generating the music preference model include: Based on the target users' historical music listening data in different activity scenarios, calculate the distribution of music style preferences in each scenario; Based on the user's overall emotional state, extract historical music features that match the emotional state; Generate a music preference model that includes the association between scene, emotion, and music features.

6. The music recommendation method based on multidimensional data according to claim 1, characterized in that, The steps to generate a personalized music recommendation playlist include: Classify the current overall emotional state into emotional intensity levels; Based on a scene-dependent music preference model, music that matches the current emotional intensity level and the type of scene activity is selected; Based on preset sorting rules and the weight of users' historical preferences, a personalized music recommendation list is generated with priority sorting.

7. The music recommendation method based on multidimensional data according to claim 1, characterized in that, The method also includes an emotional transition mechanism, comprising the following steps: When the target user’s overall emotional state or the type of activity in the scenario changes, the target emotional state is determined based on the current overall emotional state. Adjust the scene-dependent music preference model based on the difference between the target user's current emotional state and the target emotional state.

8. The music recommendation method based on multidimensional data according to claim 7, characterized in that, The music selection process for adjusting the scene-dependent music preference model also includes the following steps: Music with gradually varying emotional intensity is recommended in sequence to match the target's emotional state; The volume, rhythm, and melody of the recommended music are dynamically adjusted to match the preset emotional transition curve.

9. The music recommendation method based on multidimensional data according to claim 7, characterized in that, The method also includes real-time feedback and adaptive adjustment steps: Monitor user behavior feedback, including skipping, pausing, and manually selecting music; Adjust the music preference model based on user feedback to optimize user experience and dynamically adjust the emotional transition mechanism.

10. A music recommendation system based on multidimensional data, characterized in that, This is used to implement the music recommendation method based on multidimensional data as described in any one of claims 1 to 9.

11. A music recommendation program product based on multidimensional data, characterized in that, The program product includes computer instructions that, when executed by a processor, implement the steps of the music recommendation method based on multidimensional data as described in any one of claims 1 to 9.

12. A music recommendation device based on multidimensional data, characterized in that, include: processor; A memory in which executable instructions of the processor are stored; The processor is configured to perform the steps of the music recommendation method based on multidimensional data as described in any one of claims 1 to 9 by executing the executable instructions.

Images