Call information processing method

Abstract

The invention discloses a call information processing method, which is applied to the technical field of data processing. The method comprises the following steps: acquiring audio information of a user and a service staff in a call process, wherein the audio information comprises first audio information of the user; generating voice feature change trend information of the user according to the voice feature information of the first audio information; target emotional feature information is determined according to the voice feature change trend information, and the target emotional feature information reflects the emotional state of the user in the call process; according to the target emotion feature information, an emotion response identifier of the user is generated, and the emotion response identifier is used for generating a service strategy for the user. The method can improve the accuracy of providing services for users.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular, relates to a method for processing call information. Background Art

[0002] A service call is a call made to provide a certain service. A service call usually occurs between a user and a business person on the business side. Business people can communicate with users by answering calls or video calls, etc., to understand users' needs in a timely manner, such as problems encountered by users when using business products or services, and provide users with support, answers, guidance and other service content.

[0003] However, the experience differences among business personnel lead to differences in service quality. Some business personnel may be unable to accurately capture user needs due to their relatively limited experience, which reduces the accuracy of service content provided. Summary of the invention

[0004] The embodiments of the present application provide a method, apparatus, device, storage medium, and program product for processing call information, which can improve the accuracy of providing services to users.

[0005] In a first aspect, an embodiment of the present application provides a method for processing call information, the method comprising:

[0006] Acquire audio information during a call between a user and a service personnel, the audio information including the user's first audio information;

[0007] Generating voice feature change trend information of the user according to the voice feature information of the first audio information;

[0008] Determine target emotion feature information based on the voice feature change trend information, where the target emotion feature information reflects the user's emotional state during the call;

[0009] According to the target emotion feature information, an emotion response identifier of the user is generated, and the emotion response identifier is used to generate a service strategy for the user.

[0010] In a second aspect, an embodiment of the present application provides a device for processing call information, the device comprising:

[0011] A first acquisition module is used to acquire audio information during a call between a user and a service personnel, where the audio information includes first audio information of the user;

[0012] A first generating module, used to generate the user's voice feature change trend information according to the voice feature information of the first audio information;

[0013] A first determination module is used to determine target emotion feature information according to the voice feature change trend information, where the target emotion feature information reflects the emotional state of the user during the call;

[0014] The second generating module is used to generate an emotion response identifier of the user according to the target emotion characteristic information, and the emotion response identifier is used to generate a service strategy for the user.

[0015] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, a method for processing call information as described in any one of the first aspects is implemented.

[0016] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having computer program instructions stored thereon, and when the computer program instructions are executed by a processor, a method for processing call information as described in any one of the first aspects is implemented.

[0017] In a fifth aspect, an embodiment of the present application provides a computer program product, which includes a computer program or instructions, and when the computer program or instructions are executed by a processor, a method for processing call information as described in any one of the first aspects is implemented.

[0018] The method, device, equipment, storage medium and program product for call information processing of the embodiment of the present application can obtain the voice feature change trend information during the user's call by extracting and analyzing the voice feature information in the user's first audio information. The voice feature change trend information characterizes the changes in the user's voice during the call and can reflect the user's emotional fluctuations. Then, the voice feature change trend information is converted into a specific emotional feature, namely the target emotional feature information, to accurately reflect the user's emotional state during the call. Then, based on the target emotional feature information, the user's emotional response identifier is generated, and the emotional response identifier can provide a clear indication to the service personnel, that is, how to adjust the service strategy according to the user's emotional state. In this way, by collecting and analyzing the user's audio information, the user's emotional state during the call is accurately captured, and an emotional response identifier is generated accordingly, providing the service personnel with a targeted service strategy and improving the accuracy of the service provided by the service personnel to the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1A flowchart of a method for processing call information provided by some embodiments of the present application is shown;

[0021] Figure 2 A flowchart showing a specific implementation of step 120 provided in some embodiments of the present application is shown;

[0022] Figure 3 A flowchart showing a specific implementation of step 1201 provided in some embodiments of the present application is shown;

[0023] Figure 4 A flowchart showing a specific implementation of step 140 provided in some embodiments of the present application is shown;

[0024] Figure 5 A flow chart showing a method for generating a service policy in a call information processing method provided in some embodiments of the present application is shown;

[0025] Figure 6 A flowchart showing a specific implementation of the method for processing call information provided in some embodiments of the present application is shown;

[0026] Figure 7 A schematic diagram showing the structure of a device for processing call information provided by some embodiments of the present application;

[0027] Figure 8 A schematic structural diagram of an electronic device provided in some embodiments of the present application is shown. DETAILED DESCRIPTION

[0028] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0029] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0030] It should be noted that the acquisition, storage, use and processing of data in the embodiments of the present application are in compliance with the relevant provisions of national laws and regulations.

[0031] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0032] In order to solve the problems in the above-mentioned related technologies, the embodiments of the present application provide a method, device, equipment, storage medium and program product for processing call information. Figure 1 To Attachment Figure 6 , the call information processing method provided by the embodiment of the present application is described in detail through specific embodiments and their application scenarios.

[0033] Figure 1 FIG. 1 is a flow chart showing a method for processing call information provided by some embodiments of the present application. Figure 1 As shown, the call information processing method may include steps 110 to 140.

[0034] Step 110, obtaining audio information of a conversation between the user and the service personnel, where the audio information includes the user's first audio information.

[0035] Step 120: Generate the user's voice feature change trend information based on the voice feature information of the first audio information.

[0036] Step 130, determining target emotion feature information according to the voice feature change trend information, where the target emotion feature information reflects the emotional state of the user during the call.

[0037] Step 140: Generate an emotion response identifier for the user based on the target emotion feature information. The emotion response identifier is used to generate a service strategy for the user.

[0038] Therefore, by extracting and analyzing the voice feature information in the user's first audio information, the voice feature change trend information during the user's call can be obtained. The voice feature change trend information characterizes the changes in the user's voice during the call and can reflect the user's emotional fluctuations. Next, the voice feature change trend information is converted into specific emotional features, namely, target emotional feature information, to accurately reflect the user's emotional state during the call. Then, based on the target emotional feature information, the user's emotional response identifier is generated, and the emotional response identifier can provide a clear instruction to the service personnel, that is, how to adjust the service strategy according to the user's emotional state. In this way, by collecting and analyzing the user's audio information, the user's emotional state during the call is accurately captured, and an emotional response identifier is generated accordingly, providing the service personnel with targeted service strategies and improving the accuracy of the service provided by the service personnel to the user.

[0039] The above steps are described in detail below, as shown below.

[0040] First, step 110 is involved. The audio information involved in the embodiment of the present application involves the sound data generated during the conversation between the user and the service personnel. For example, the conversation audio between the user and the service personnel can be collected in real time through a recording device or a call system.

[0041] Next, in step 120, the voice feature information involved in the embodiment of the present application refers to data extracted from the audio information that can reflect the voice characteristics of the speaking user, such as basic voice features such as speech speed, volume, tone, rhythm, timbre, and pitch. Voice feature change trend information refers to the trend of voice feature information changing over time, such as a sudden increase in speech speed, a sudden increase in volume, a sudden decrease in tone, etc.

[0042] Exemplarily, extracting the speech feature information of the first audio information may specifically include speech preprocessing and speech feature extraction. Among them, speech preprocessing refers to the application of filtering technology or noise suppression algorithm to remove background noise and interference in the original speech signal to ensure signal clarity. Next, use signal enhancement technology, such as adaptive gain control or noise adaptive filtering, to improve the clarity and recognizability of the speech signal. Then, divide the continuous speech signal into smaller segments or frames to facilitate subsequent feature extraction and analysis. Speech feature extraction may include speech rate feature extraction, volume intensity feature extraction, tone and rhythm feature extraction, and timbre and pitch feature extraction.

[0043] Next, the implementation method of extracting each speech feature of the speech feature will be specifically described.

[0044] For speech rate feature extraction, a support vector machine (SVM) can be used to calculate the speech rate of the segmented speech frames to obtain the number of syllables or words per second.

[0045] Volume intensity feature extraction can calculate the energy value of each speech frame and obtain the overall volume intensity feature by accumulation or averaging to reflect the loudness of the speech signal.

[0046] Tone and rhythm feature extraction can analyze the periodic characteristics of sound signals, such as the change of fundamental frequency, to extract tone features. By analyzing the envelope or energy distribution of speech signals, rhythm features such as pauses and stress can be extracted.

[0047] Timbre and pitch feature extraction can convert the sound signal into the frequency domain and use Fourier transform or related spectrum analysis methods to extract the timbre and pitch features. Specifically, the timbre features can be obtained by analyzing the shape, distribution, and energy of the spectrum. The pitch features can be obtained by detecting the fundamental frequency and its changes.

[0048] Thus, after obtaining the feature vector, multiple feature vectors can be integrated to form a complete feature vector, namely speech feature information. This speech feature information may include multiple aspects of the first audio information, such as speech speed, volume, tone, rhythm, timbre and pitch.

[0049] In some embodiments of the present application, Figure 2 As shown, the above step 120 may specifically include steps 1201 to 1203.

[0050] Step 1201, calculating the statistical feature value of the speech feature value of the speech feature information according to the speech feature information.

[0051] The statistical characteristic value of the speech characteristic value is a characteristic value obtained after statistical analysis of the speech characteristic value, and is used to describe the change trend of the speech characteristic.

[0052] Step 1202: Determine, based on the statistical characteristic value, a target threshold range that matches the statistical characteristic value from a plurality of preset threshold ranges.

[0053] The preset threshold range is a series of numerical ranges set in advance and is used to classify or judge the statistical characteristic values.

[0054] Exemplarily, the preset threshold range may be determined based on a model, which may specifically include the following steps:

[0055] Construct a training set, which includes multiple training samples, each of which includes sample call audio information and reference service rating values. In order to obtain effective call audio information, it is necessary to screen out effective, clear and service-related recordings from a large amount of customer service telephone service recording data, and then remove noise, noise, irrelevant conversations, etc. from the screened recordings to ensure the purity and relevance of the sample call audio information. Next, by manually marking the sample call audio information and using the reference service rating value in the service ticket satisfaction as feedback annotation, the model is provided with accurate emotional labels and satisfaction information, which helps the model better learn the relationship between user emotions and satisfaction. Input the training set, feedback annotation, basic feature items such as speech speed, volume, tone, etc., and change trend types such as sudden increase, sudden decrease, fluctuation, etc. into the trained model. Based on these data, the model will use machine learning or deep learning algorithms to automatically learn and identify the user's emotional change trend during the call, and output a set of change trend thresholds for basic feature items. Then, the change trend threshold output in the previous step is tested for effectiveness. Specifically, this can be achieved by applying the threshold to the new audio information and comparing the degree of consistency between the analysis results and the actual user emotions. If the test results show that the thresholds have high accuracy and reliability, then these thresholds can be confirmed and enabled for actual user sentiment analysis and service optimization.

[0056] Step 1203, based on the preset threshold range and the third association information of the preset voice feature change trend information, the voice feature change trend information associated with the target threshold range is determined as the voice feature change trend information of the user.

[0057] The third association information refers to the correspondence or mapping relationship between the preset threshold range and the preset voice feature change trend information. The preset threshold range and the preset voice feature change trend information can be set and adjusted according to specific needs to meet the voice feature analysis needs of different users and different scenarios.

[0058] Exemplarily, taking the voice feature change trend information including speech rate change, volume change and tone change as an example, the third associated information may refer to the following Table 1.

[0059]

[0060] Table 1

[0061] Take a recording of a user's call as an example, and divide it into multiple time slices, for example, each 10 seconds is a time slice. In a certain time slice such as T2, it is calculated that the user's speaking speed has increased by 35% compared with the previous time slice T1. According to the speech speed change threshold in Table 1, the 35% increase rate belongs to the "medium" level of sudden increase. In another time slice such as T4, it is calculated that the user's volume has decreased by 40% compared with the previous time slice T3. According to the volume change threshold in Table 1, the 40% decrease belongs to the "medium" level of sudden decrease. In a certain continuous time period, such as from T6 to T8, it is calculated that the user's tone has changed significantly in three continuous time slices, such as rising first, then falling, and then rising again. According to the tone change threshold in Table 1, if the number of changes in this tone exceeds the "low" level threshold, that is, more than 1 time, it can be considered that the user's tone fluctuates in this time period. Combining the analysis of the three voice features of voice, volume and tone, a more comprehensive user emotional state assessment can be obtained.

[0062] Therefore, by calculating the statistical feature value, matching the threshold range and searching the third associated information, the automatic analysis of the changing trend of the user's voice feature is realized, thereby improving the efficiency and accuracy of the analysis.

[0063] In some embodiments of the present application, in order to improve the efficiency and accuracy of the calculation of statistical feature values, the first audio information is divided into I consecutive time periods, and the speech feature information in each time period is called speech feature sub-information, that is, the speech feature information includes the speech feature sub-information of the i-th time period in I time periods, i∈[1, 2, 3, ..., I]; the statistical feature value includes the mean change rate and the number of fluctuations. Based on this, if Figure 3 As shown, the above step 1201 may specifically include steps 12011 to 12013.

[0064] Step 12011, obtaining the average value of the speech features in the i-th time period according to the speech feature values ​​of the speech feature sub-information in the i-th time period.

[0065] The speech feature value refers to a specific value extracted from the speech feature sub-information, which is used to quantitatively describe a certain characteristic of speech, such as speech speed, volume, tone, etc. The speech feature average value reflects the general level of speech features in that time period.

[0066] Exemplarily, for each time period i, the speech feature sub-information in the time period is extracted, and the average value of all speech feature values ​​in the time period is calculated to obtain the speech feature average value of the time period.

[0067] Step 12012, calculate the mean change rate of the speech feature value based on the average value of the speech feature in the i-1th time period and the average value of the speech feature in the i-th time period.

[0068] The mean change rate refers to the degree of change of the average value of speech features in adjacent time periods, reflecting the rate at which speech features change over time. For example, for two adjacent time periods i-1 and i, their speech feature average values ​​are calculated respectively. The difference between the two average values ​​is calculated and divided by the time interval of the time period to obtain the mean change rate.

[0069] Step 12013, determine the number of fluctuations of the speech feature value according to I-1 mean change rates of the speech feature value in I time periods.

[0070] The fluctuation times refer to the times that the speech feature value exceeds a certain threshold or changes significantly within a period of time, reflecting the fluctuation of the speech feature. Exemplarily, for the entire time period, i.e., I time period, the mean change rate between all adjacent time periods is calculated. According to a preset threshold or condition, it is determined which mean change rates represent significant changes in speech features. The number of these significant changes is counted to obtain the fluctuation times.

[0071] Therefore, by calculating the statistical feature values ​​of the speech feature values, such as the mean change rate and the number of fluctuations, a set of proprietary feature vectors that can effectively indicate the user's emotions can be obtained. These feature vectors can more accurately reflect the user's emotional state, thereby providing more accurate user emotion analysis.

[0072] In some embodiments of the present application, if the first audio information is from T 0 First, divide the call time into I time periods, each time period is T i , T i ∈{T 1 , T 2 , .., T I The specific steps of processing the first audio information to obtain the voice feature change trend information are shown in steps 1 to 4.

[0073] Step 1: Extraction and calculation of speech feature values. First, extract speech feature values ​​from the speech feature sub-information corresponding to each time period. Then, for each time period, calculate the initial value V of the speech feature value in each time period. start-i , average value V avg-i , maximum value V max-i , minimum value V min-i , End value V end-i .

[0074] Step 2: Calculate the mean change rate and determine the trend of speech feature changes. First, for a certain speech feature, use the mean change rate calculation function C = f(V avg) to solve the mean change rate C of the speech feature. f(V avg ) The mean change rate calculation function supports a variety of general algorithms, such as differential method, calculus method and curve fitting method. After obtaining the mean change rate C, combined with the above Table 1, a set of qualitative results R of change trends and degrees can be obtained. Each element in this set describes the speech feature item, change trend type and change degree in detail, such as the sudden increase / decrease of speech speed (low / medium / high), the sudden increase / decrease of volume (low / medium / high) and the sudden increase / decrease of tone (low / medium / high).

[0075] Step 3: Calculate the fluctuation times N, and calculate the mean change rate of speech features corresponding to two adjacent time periods C = f(V avg ), and these calculation results are recorded in a result set {C i+1 , ..., C I Then, the function f(C j ) to solve the fluctuation number N. Fluctuation number calculation function f(C j ) function supports a variety of general algorithms, such as point-by-point comparison method, cumulative sum method, and binary search method, to accurately obtain the number of peaks, that is, the number of cycles of sudden increase / decrease of the user's voice features within a period of time.

[0076] Step 4, we compare the calculated fluctuation times N with Table 1 to obtain the qualitative result R of the fluctuation degree of a certain speech feature over a period of time. This qualitative result also includes information such as speech feature items, change trend types, and change degrees, such as speech rate fluctuations (low / medium / high), volume fluctuations (low / medium / high), and tone fluctuations (low / medium / high).

[0077] Next, step 130 is involved. The target emotional feature information in the embodiment of the present application refers to data determined based on the voice feature change trend information and can reflect the user's emotional state, such as politeness, tolerance, confusion, urgency, and help-seeking.

[0078] In an embodiment of the present application, in order to improve the efficiency and accuracy of acquiring the target emotional characteristic information, the aforementioned step 130 may specifically include determining the emotional characteristic information associated with the voice characteristic change trend information as the target emotional characteristic information based on the first association information of the preset voice characteristic change trend information and the preset emotional characteristic information. Among them, the second association information of the preset voice characteristic change trend information and the preset emotional characteristic information may be established based on the analysis and learning of a large amount of data, through a machine learning algorithm or an expert system.

[0079] Exemplarily, after determining the voice feature change trend information of the user, the voice feature change trend information is matched with the preset first associated information. According to the matching result, the emotional feature information corresponding to the voice feature change trend, that is, the target emotional feature information, is determined. Thus, by combining the corresponding relationship between the voice feature change trend and the emotional feature, the user's emotional state can be more accurately identified.

[0080] In some embodiments of the present application, before the aforementioned step 130, the method for processing call information may further include performing voice recognition on the audio information to determine call feature information, where the call feature information includes at least one of the following: service type, user interruption information. Based on this, the aforementioned step 130 may further include: determining target emotion feature information based on the voice feature change trend information and the call feature information.

[0081] In some embodiments of the present application, performing speech recognition on audio information and determining the service type may include:

[0082] Performing speech recognition on the audio information to obtain first text information;

[0083] Determining a business scenario keyword corresponding to the first text information from a plurality of preset business scenario keywords;

[0084] When the business scenario keyword meets the first preset condition, the business type associated with the business scenario keyword is determined as the business type of the call based on the association information between the preset business scenario keyword and the preset business type; wherein the first preset condition includes at least one of the following: the number of business scenario keywords within the preset time period is greater than the preset number, and the number of business scenario keywords within the preset time period is greater than the preset number.

[0085] The first text information refers to the text content extracted from the audio information by voice recognition technology. The business scenario keyword refers to a predefined keyword or phrase related to a specific business scenario. The business scenario information is information that describes the business type to which the call belongs, such as consultation, complaint, handling, etc., based on the business scenario keyword. The preset time period refers to a specific time period used to count the number or frequency of business scenario keywords. The first preset condition refers to a condition for determining whether the specific text information is associated with the business scenario keyword.

[0086] Thus, through voice recognition and business scenario keyword matching, the business scenario information of the call can be automatically and accurately determined, improving service efficiency and quality. In addition, by setting the first preset condition, misjudgment caused by accidental factors can be avoided, and the accuracy of business scenario recognition can be improved.

[0087] In some embodiments of the present application, performing voice recognition on audio information and determining user interruption information may include, when the audio information contains an audio segment that meets a second preset condition, determining the user's interruption frequency information based on the number of audio segments, wherein the preset condition includes any one of the following: the duration of the audio segment is greater than the preset duration, and the voiceprint features of the service personnel in the audio segment occur earlier than the voiceprint features of the user.

[0088] Exemplarily, first, the service personnel voiceprint and the user voiceprint can be compared and identified through a preset voice library. Then, in the audio information, find a segment in which two or more voiceprints are superimposed at the same time, and the overlapping time exceeds the preset time t, such as t is not less than 1 second, and determine it as the initial audio segment. Then, for the initial audio segment, if the start time of the service personnel voiceprint is earlier than the start time of the user voiceprint, it is determined that the user interrupted, and the initial audio segment is determined as the target audio segment. Finally, the user's interruption information is determined according to the number of target audio segments, such as according to the number of interruptions, that is, the number of target audio segments, the interruption behavior is divided into four levels: none, low, medium, and high. The specific classification standard can be that the number of interruptions is <1 time for none, 1-2 times for low, 3-5 times for medium, and more than 5 times for high.

[0089] Therefore, the multi-dimensional information fusion combining voice feature change trend information and call feature information can more flexibly adapt to different call scenarios, thereby more accurately identifying user emotions. This helps reduce misjudgments and omissions of user emotions and improves the accuracy of emotion recognition.

[0090] In some embodiments of the present application, the above-mentioned determination of target emotional characteristic information based on voice feature change trend information and call characteristic information may include, based on second association information of preset call information and preset emotional characteristic information, determining emotional characteristic information associated with the call information as target emotional characteristic information, the call information including voice feature change trend information and call characteristic information.

[0091] For example, the target emotional feature information may include positive emotional classes, neutral emotional classes, and negative emotional classes. Each emotional class includes multiple specific emotional states, such as positive emotional classes may include politeness, tolerance, etc.; neutral emotional classes may include confusion, urgency, help-seeking, and complaint, etc.; negative emotional classes may include complaining, anxiety, and confrontation, etc.

[0092] Taking the voice feature change trend information including speech speed change information, volume change information and tone change information, and the call feature information including interruption level and service type as examples, by looking up Table 2, the basic feature vectors in the call process are matched with the emotional state to determine the user's emotional state.

[0093]

[0094] Table 2

[0095] Based on this, the aforementioned step 120 may include, first, dividing the first audio information into a group of adjacent time slices, such as {T i , T i+1 , ..., T i+j}; Next, obtain a set of speech feature vectors such as the change rate C value and the fluctuation number N value of speech speed, volume and tone. Then, according to the above Table 1, the speech feature vector change result set R can be obtained, R = {R i , R i+1 , .., R i+j}, where R i Reflects a time slice T i The key-value pair set of each speech feature change item and the qualitative analysis result in . For example, R i = [(interruption, none), (speech rate sudden increase, medium), (volume sudden increase, low), (tone fluctuation, high), ...], indicating that in time slice T i In the process, no interruption was detected, but changes such as sudden increase in speech speed, sudden increase in volume, and fluctuation in tone were detected.

[0096] Then, step 140 is involved. The emotion response indicator in the embodiment of the present application refers to an indicator generated based on the target emotion feature information and used to guide how to respond to the user's emotions. The service strategy refers to a strategy or action plan formulated based on the emotion response indicator to effectively respond to the user's emotions, such as adjusting the speech speed, changing the tone, providing comfort or suggestions, etc.

[0097] In some embodiments of the present application, the aforementioned step 140 may specifically include, based on fourth association information of preset emotion characteristic information and preset emotion response identifier, determining the target emotion response identifier associated with the emotion characteristic information as the user's emotion response identifier.

[0098] The preset emotional characteristic information is a series of emotional characteristics defined in advance, which are used to describe the emotional state of the user. For example, the preset emotional characteristic information may include politeness, tolerance, confusion, urgency, seeking help, complaint, grumbling, anxiety, confrontation, etc.

[0099] The preset emotion response mark is a response mark preset for specific emotion characteristic information, which is used to guide how to respond to or deal with the user's current emotional state. Exemplarily, the preset emotion response mark may include listening, soothing, guiding, quick handling, etc.

[0100] The fourth association information is a database or a mapping table, which establishes an association between the preset emotion characteristic information and the preset emotion response identifier. Exemplarily, the emotion response identifier matching the emotion characteristic information of the current user can be found in the fourth association information.

[0101] Therefore, by utilizing the fourth association information, the user's emotional state can be identified quickly and accurately, and an emotional response identifier that matches the current user's emotional characteristic information can be found, so as to provide the user with corresponding services based on the emotional response identifier, effectively improving service efficiency.

[0102] For example, the emotional feature information includes the emotional state of the user, and the emotional state of the user can be further divided into specific emotional types. These emotional types can be roughly divided into three categories: positive emotional class, neutral emotional class and negative emotional class. Table 3 describes the specific emotional features, feature descriptions and service strategy references for different emotional types such as positive emotional class, neutral emotional class and negative emotional class.

[0103]

[0104] Table 3

[0105] The positive emotion categories in Table 3 may include politeness and tolerance. Users in these emotional states usually show a relatively calm and friendly communication attitude. The neutral emotion category may include confusion, urgency, seeking help, and complaint. Users in these states may be in a certain psychological state or need guidance or resolution. Negative emotion categories may include complaining, anxiety, and confrontation. Users in these states usually show dissatisfaction, anxiety, or confrontation.

[0106] In some other embodiments of the present application, the target emotion characteristic information includes the first emotion characteristic information of the user in the i-1th time period and the second emotion characteristic information in the i-th time period. Based on this, Figure 4 As shown, the aforementioned step 140 may specifically include step 1401 to step 1402.

[0107] Step 1401: Generate emotion change information according to the first emotion characteristic information and the second emotion characteristic information.

[0108] Among them, the emotion change information refers to the information generated based on the first emotion characteristic information and the second emotion characteristic information, which reflects the changes in the user's emotions in different time periods. On the one hand, the emotion change information can determine the state migration of the emotion characteristics. For example, if the user changes from a state of seeking help to an urgent state, this means that the effectiveness of the telephone communication between the user and the service personnel has significantly decreased, and the user's emotions are unstable, prompting the service personnel to comfort the service personnel and respond to the service personnel's concerns in a timely manner; on the other hand, it can determine whether the degree of the emotion characteristics has intensified. For example, if the user changes from a state of complaining to a state of anxiety, this means that the telephone communication between the service personnel and the user has produced a negative effect, and the conflict between the two parties may break out and intensify.

[0109] In some embodiments of the present application, the above step 1401 may also include:

[0110] In a case where the emotional state of the first emotional characteristic information and the emotional state of the second emotional characteristic information are the same, determining the first emotional change information as the emotional change information, the first emotional change information including information used to reflect that the user's emotions are in a stable state;

[0111] When the emotional state of the first emotional characteristic information and the emotional state of the second emotional characteristic information are different, the emotional change information is determined according to the emotional types and emotional intensities of the first emotional state and the second emotional state.

[0112] Exemplarily, when the emotional states of the first emotional characteristic information and the second emotional characteristic information are the same, it means that the user's emotions remain stable during this period. Therefore, the emotional change information is determined as information reflecting that the user's emotions are in a stable state. For example, if the user is in a "complaining state" at two consecutive time points and the degree of complaint does not change, then the emotional change information is "negative emotion maintenance".

[0113] When the emotional states of the first emotional characteristic information and the second emotional characteristic information are different, it is necessary to determine the emotional change information based on the emotional type and emotional intensity of the two emotional states. This involves judging whether the user's emotions have become more positive or more negative, and the degree of change. For example, if the user changes from a "complaining state" to an "anxious state", and the degree of anxiety is stronger than complaining, then the emotional change information is "deepening negative emotions". On the contrary, if the user changes from a "confrontational state" to an "anxious state", but the degree of anxiety is weaker than confrontation, then the emotional change information is "weakening negative emotions".

[0114] See Table 4 below, which lists the conversion between different emotion feature information and the corresponding emotion change information such as deepening of negative emotions, maintenance of negative emotions, weakening of negative emotions, generation of negative emotions and abnormal fluctuations of emotions.

[0115]

[0116] Table 4

[0117] As an example, a deepening of negative emotions means that the user's negative emotions become more severe or intense. A sustained negative emotion means that the user's negative emotions remain at a relatively stable level without significant changes. A weakening of negative emotions means that the user's negative emotions begin to ease or fade. A generation of negative emotions means that the user begins to show some negative emotions, where there may not be obvious emotions or the emotions are relatively calm before. An abnormal mood fluctuation means that the user's emotions have changed significantly in a short period of time, which may be caused by the alternation of multiple emotional states or a sudden mood shift.

[0118] Therefore, by comparing the user's emotional feature information at different time points, it is possible to accurately identify and process the user's emotional state changes and generate corresponding emotional change information. In this way, the user's emotional needs can be better understood and more personalized services and support can be provided.

[0119] Step 1402: Based on the association information between the preset emotion change information and the preset emotion response identifier, the emotion response identifier associated with the emotion change information is determined as the emotion response identifier of the user.

[0120] Among them, the emotional response indicators associated with the emotional change information can provide service personnel with service strategies or words that can better respond to the user's emotions and needs, thereby providing more personalized services. For example, emotional response indicators can include listening, comforting, guiding, and quick handling.

[0121] For example, if it is found that the user changes from "complaining state" to "anxious state" and it is predicted that this is the judgment condition of "deepening negative emotions", then the service staff may need to adopt a more patient and meticulous service method to avoid further deterioration of the user's emotions, and the corresponding emotional response indicator should be listening. On the contrary, if the user changes from "anxious state" to "complaining state" and it is predicted that this is the judgment condition of "reducing negative emotions", then the service staff may need to adopt a more positive and proactive service method to further alleviate the user's emotions and solve the problem, and the corresponding emotional response indicator should be guidance.

[0122] Thus, by comparing the emotional feature information of the user in different time periods, i.e., the i-1th time period and the ith time period, the change of the user's emotions can be captured. Then, based on the emotional change information, the preset association information, i.e., which emotional response identifiers correspond to different emotional changes, is used to recommend the most appropriate emotional response strategy or suggestion to the user. In this way, the response strategy can be dynamically adjusted according to the actual emotional changes of the user.

[0123] In some embodiments of the present application, in a call scenario between a service personnel and a user, in order to better understand the user's emotional state and provide appropriate service strategy support, the following Table 5 is a detailed explanation of the above four emotional response indicators such as listening, soothing, guiding, and quick response and their applicable emotional change information.

[0124]

[0125] Table 5

[0126] In some embodiments of the present application, the aforementioned call information processing method may further include steps 210 to 220.

[0127] Step 210: Determine the service type provided by the service personnel according to the second audio information.

[0128] Among them, service types may include consulting services, complaint services, processing services, etc.

[0129] As an example, the voice recognition technology can be used to extract keywords from the second audio information of the service personnel. These keywords may directly reflect the type of service, such as "consultation", "handling", "complaint", etc.

[0130] As another example, semantic analysis can be performed on the second audio information of the service personnel. Through natural language processing technology, the specific content and intention expressed by the service personnel in the call can be understood, so as to more accurately determine the service type.

[0131] As another example, the second audio information of the service personnel can be matched with the known service process in combination with the known service process. Different service types usually have different processes and requirements, and the type of service can be further confirmed by matching.

[0132] Step 220: Generate a service strategy for the user based on the service type, target emotion characteristic information and emotion response identifier.

[0133] Exemplarily, the service type, target emotional feature information and emotional response identifier are used as inputs to match with the preset service strategy library of the business end. In the preset service strategy library, there may be multiple service strategies for different service types, different emotional feature information and different emotional response identifiers. In this way, service personnel can obtain real-time decision recommendations, thereby improving service quality in a timely manner and improving user satisfaction. After the user and the service personnel complete the voice communication, the business end collects the satisfaction data after the service and the efficiency and effect data of the service personnel in handling complaints, such as processing time, effectiveness of solutions, etc., through user feedback or satisfaction surveys. Then, the collected satisfaction data and handling efficiency data are fed back to the preset service strategy library. The business end can use these data to perform quality inspection and enhanced training on the preset service strategy library to continuously improve the accuracy of service strategy matching. For example, if it is found that the user satisfaction is high after adopting the appeasement strategy, the business end can adjust the weight in the preset service strategy library so that the appeasement strategy is easier to be matched in similar scenarios.

[0134] Therefore, the service type provides the basic framework of the service, the target emotional feature information reveals the user's emotional needs, and the emotional response indicator provides a specific response method. The service strategy formulated by considering the service type, target emotional feature information and emotional response indicator at the same time is more comprehensive and accurate, and can more accurately meet the needs and expectations of users.

[0135] In some embodiments of the present application, the aforementioned method for generating a preset emotional response identifier may specifically include steps 310 to 340.

[0136] Step 310: construct a generalized feature library.

[0137] For example, the generalized feature library may include vocabulary features covering prohibited terms, illegal behaviors, red line issues, public opinion risks, information security, etc. The keywords in these generalized feature libraries help to efficiently and quickly identify improper tendencies and abnormal trends of both parties in the call.

[0138] Step 320: Generate a semantic tree.

[0139] For example, for audio information with negative emotional changes and abnormal emotional fluctuations, language analysis is performed, including but not limited to word segmentation, part-of-speech tagging, syntactic analysis, semantic analysis, and intent recognition. Through these steps, a detailed semantic tree can be generated, and the tree structure of the semantic tree clearly shows the words, phrases, sentences in the audio information and the grammatical and semantic relationships between them.

[0140] Step 330, determine vocabulary features.

[0141] Exemplarily, based on the generated semantic tree, it is further compared with the generalized feature library to determine the lexical features of the service personnel's terms in the audio information, so as to quickly identify sensitive or inappropriate terms that may exist in the call.

[0142] Step 340: input the model to generate a preset emotional response identifier.

[0143] For example, lexical features, emotional feature information, service types such as consulting services, complaint services, handling services, etc., and key content semantics are input into a pre-trained model. This pre-trained model can comprehensively consider information from multiple dimensions and output a preset emotional response identifier. In specific implementation, the model can use a deep learning framework and a natural language processing library to build and train the model. At the same time, pre-trained language models and sentiment analysis models can be used to improve performance.

[0144] Next, the above-mentioned method of processing call information will be specifically described by taking a user service system as an example.

[0145] In a user service system, in order to improve service quality and user satisfaction, the user service system needs to be able to accurately identify and analyze the user's emotional state and behavioral characteristics, so as to adopt corresponding service strategies. Figure 6 As shown in the figure, when the user starts to communicate with the service staff, the user service system will first capture the user's voice features, such as speaking speed, volume, and tone. As the user and the service staff communicate more deeply, the user's voice features may change, and the voice feature change trend information is obtained. For example, when the user feels dissatisfied or anxious, the speaking speed may increase and the tone may rise. The user service system needs to be able to monitor these changes in real time so as to adjust the response strategy in time. Based on the voice features, the user service system further determines the emotional feature information, such as "politeness" and "tolerance", based on the voice feature change trend information. These emotional feature information reflects the user's emotional state and behavioral tendencies, which helps the system understand the user's needs and expectations more accurately. In addition to the emotional feature information, the user service system also extracts emotional change information, such as "deepening negative emotions", "deepening negative emotions", "weakening negative emotions", etc. These emotional change information further enriches the user's emotional and behavioral portraits. Then, based on the analysis of the above features, the user service system generates emotional response indicators, such as "comfort" and "guidance". These indicators indicate the service strategies that service personnel should adopt to meet user needs and improve satisfaction. Finally, after adopting the service strategy, the user service system will collect user feedback data, such as satisfaction, complaint rate, etc. These data are used to evaluate the effectiveness of the response strategy and serve as the basis for subsequent optimization.

[0146] Based on the method for processing call information provided in the above embodiment, the present application also provides a specific implementation of a device for processing call information. Please refer to the following embodiment.

[0147] See also Figure 7 The apparatus 400 for processing call information provided in the embodiment of the present application includes:

[0148] A first acquisition module 410 is used to acquire audio information during a call between a user and a service personnel, where the audio information includes first audio information of the user;

[0149] A first generating module 420, configured to generate voice feature change trend information of a user according to the voice feature information of the first audio information;

[0150] The first determination module 430 is used to determine target emotion feature information according to the voice feature change trend information, where the target emotion feature information reflects the emotional state of the user during the call;

[0151] The second generating module 440 is used to generate an emotion response identifier of the user according to the target emotion characteristic information, and the emotion response identifier is used to generate a service strategy for the user.

[0152] Thus, by using the first generation module 420 to extract and analyze the voice feature information in the first audio information of the user obtained by the first acquisition module 410, the voice feature change trend information during the user's call can be obtained, and the voice feature change trend information characterizes the change of the user's voice during the call, and can reflect the user's emotional fluctuations. Then, the voice feature change trend information is converted into a specific emotional feature, that is, the target emotional feature information, by using the first determination module 430, so as to accurately reflect the user's emotional state during the call. Then, the second generation module 440 is used to generate the user's emotional response mark according to the target emotional feature information, and the emotional response mark can provide a clear indication for the service personnel, that is, how to adjust the service strategy according to the user's emotional state. In this way, by collecting and analyzing the user's audio information, the user's emotional state during the call is accurately captured, and the emotional response mark is generated accordingly, providing the service personnel with a targeted service strategy, and improving the accuracy of the service provided by the service personnel to the user.

[0153] In some embodiments of the present application, the above-mentioned first determination module 430 is specifically used to determine the emotional feature information associated with the voice feature change trend information as the target emotional feature information based on the first association information of the preset voice feature change trend information and the preset emotional feature information.

[0154] In some embodiments of the present application, the call information processing device 400 may further include:

[0155] The first recognition module is used to perform speech recognition on the audio information to determine the call feature information before determining the target emotion feature information according to the speech feature change trend information, and the call feature information includes at least one of the following: business type and user interruption information. Based on this, the first determination module 430 can be used to determine the target emotion feature information according to the speech feature change trend information and the call feature information.

[0156] In some embodiments of the present application, the above-mentioned first determination module 430 can be used to determine the emotional characteristic information associated with the call information as the target emotional characteristic information based on the second association information of the preset call information and the preset emotional characteristic information, and the call information includes voice feature change trend information and call characteristic information.

[0157] In some embodiments of the present application, the first generating module 420 may include:

[0158] A first calculation submodule, used for calculating the statistical characteristic value of the speech characteristic value of the speech characteristic information according to the speech characteristic information;

[0159] A first determination submodule is used to determine, according to the statistical characteristic value, a target threshold range that matches the statistical characteristic value from a plurality of preset threshold ranges;

[0160] The second determination submodule is used to determine the voice feature change trend information associated with the target threshold range as the user's voice feature change trend information based on the preset threshold range and the third association information of the preset voice feature change trend information.

[0161] In some embodiments of the present application, the first calculation submodule can be used to obtain the average value of the speech feature of the i-th time period according to the speech feature value of the speech feature sub-information of the i-th time period in the case where the speech feature information includes the speech feature sub-information of the i-th time period in I time periods, i∈[1, 2, 3, ..., I]; and the statistical feature value includes the mean change rate and the number of fluctuations;

[0162] Calculate the mean change rate of the speech feature value according to the average value of the speech feature in the i-1th time period and the average value of the speech feature in the i-th time period;

[0163] The number of fluctuations of the speech feature value is determined according to I-1 mean change rates of the speech feature value in I time periods.

[0164] In some embodiments of the present application, the above-mentioned second generation module 440 can be used to determine the target emotion response identifier associated with the emotion characteristic information as the user's emotion response identifier based on the fourth association information of the preset emotion characteristic information and the preset emotion response identifier.

[0165] In some embodiments of the present application, the second generating module 440 may include:

[0166] A first generating submodule, for generating emotion change information according to the first emotion characteristic information and the second emotion characteristic information when the target emotion characteristic information includes the first emotion characteristic information of the user in the i-1th time period and the second emotion characteristic information in the i-th time period;

[0167] The third determination submodule is used to determine the emotion response identifier associated with the emotion change information as the user's emotion response identifier based on the association information between the preset emotion change information and the preset emotion response identifier.

[0168] In some embodiments of the present application, the first generation submodule may be used to:

[0169] In a case where the emotional state of the first emotional characteristic information and the emotional state of the second emotional characteristic information are the same, determining the first emotional change information as the emotional change information, the first emotional change information including information used to reflect that the user's emotions are in a stable state;

[0170] When the emotional state of the first emotional characteristic information and the emotional state of the second emotional characteristic information are different, the emotional change information is determined according to the emotional types and emotional intensities of the first emotional state and the second emotional state.

[0171] In some embodiments of the present application, the above-mentioned call information processing device may further include:

[0172] A second determination module is used to determine the type of service provided by the service personnel according to the second audio information when the audio information includes the second audio information of the service personnel;

[0173] The third generation module is used to generate a service strategy for the user according to the service type, target emotion characteristic information and emotion response identification.

[0174] Each module of the call information processing device 500 provided in the embodiment of the present application can realize Figures 1 to 6 The functions of each step of the method for call information processing are provided, and the corresponding technical effects can be achieved. For the sake of concise description, they will not be repeated here.

[0175] Figure 8 A schematic diagram of the hardware structure of an electronic device provided in some embodiments of the present application is shown.

[0176] The electronic device may include a processor 501 and a memory 502 storing computer program instructions.

[0177] Specifically, the processor 501 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0178] The memory 502 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 502 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 502 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 502 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 502 is a non-volatile solid-state memory.

[0179] In a particular embodiment, the memory 502 may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, generally, the memory 502 includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the data processing method according to the first aspect of the present application.

[0180] The processor 501 implements any one of the call information processing methods in the above embodiments by reading and executing the computer program instructions stored in the memory 502 .

[0181] In one example, the electronic device may further include a communication interface 503 and a bus 510. Figure 8 As shown, the processor 501, the memory 502, and the communication interface 503 are connected via a bus 510 and communicate with each other.

[0182] The communication interface 503 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0183] Bus 510 includes hardware, software or both, and the parts of electronic equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 510 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the application considers any suitable bus or interconnection.

[0184] The electronic device can execute the call information processing method in the embodiment of the present application, thereby realizing the combination Figures 1 to 7 The invention describes a method and apparatus for processing call information.

[0185] In addition, in combination with the method for processing call information in the above-mentioned embodiment, the embodiment of the present application may provide a computer-readable storage medium for implementation. The computer-readable storage medium stores computer program instructions; when the computer program instructions are executed by the processor, any one of the methods for processing call information in the above-mentioned embodiment is implemented. Examples of computer-readable storage media include non-transitory computer-readable storage media, such as portable disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, etc.

[0186] In addition, in combination with the method for processing call information in the above embodiment, the embodiment of the present application may provide a computer program product for implementation. The program product is stored in a storage medium, and may specifically include a computer program or instruction. When the computer program or instruction is executed by a processor, any one of the methods for processing call information in the above embodiment is implemented. The program product is executed by at least one processor to implement the various processes of the above data processing method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0187] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.

[0188] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0189] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.

[0190] Aspects of the present disclosure are described above with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0191] The above is only a specific implementation of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.

Claims

1. A method for processing call information, characterized in that: include: Acquire audio information during a call between a user and a service personnel, wherein the audio information includes first audio information of the user; generating voice feature change trend information of the user according to the voice feature information of the first audio information; Determining target emotion feature information according to the voice feature change trend information, wherein the target emotion feature information reflects the emotional state of the user during the call; An emotion response identifier of the user is generated according to the target emotion characteristic information, and the emotion response identifier is used to generate a service strategy for the user.

2. The method according to claim 1, characterized in that Determining target emotion feature information according to the voice feature change trend information includes: Based on the preset voice feature change trend information and the first association information of the preset emotion feature information, the emotion feature information associated with the voice feature change trend information is determined as the target emotion feature information.

3. The method according to claim 1, characterized in that Before determining the target emotion feature information according to the voice feature change trend information, the method further includes: Performing voice recognition on the audio information to determine call feature information, where the call feature information includes at least one of the following: service type and user interruption information; Determining target emotion feature information according to the voice feature change trend information includes: Target emotion feature information is determined according to the voice feature change trend information and the call feature information.

4. The method according to claim 3, characterized in that The determining target emotion feature information according to the voice feature change trend information and the call feature information includes: Based on second association information of preset call information and preset emotion characteristic information, the emotion characteristic information associated with the call information is determined as the target emotion characteristic information, and the call information includes the voice characteristic change trend information and the call characteristic information.

5. The method according to claim 1, characterized in that The step of generating the voice feature change trend information of the user according to the voice feature information of the first audio information includes: Calculating a statistical feature value of a speech feature value of the speech feature information according to the speech feature information; According to the statistical characteristic value, determining a target threshold range matching the statistical characteristic value from a plurality of preset threshold ranges; According to the preset threshold range and the third association information of the preset voice feature change trend information, the voice feature change trend information associated with the target threshold range is determined as the voice feature change trend information of the user.

6. The method according to claim 5, characterized in that The speech feature information includes the speech feature sub-information of the i-th time period in the I time periods, i∈[1, 2, 3, ..., I]; the statistical feature value includes the mean change rate and the number of fluctuations; The step of calculating the statistical feature value of the speech feature value of the speech feature information according to the speech feature information comprises: Obtaining an average value of the speech features in the i-th time period according to the speech feature values ​​of the speech feature sub-information in the i-th time period; Calculate the mean change rate of the speech feature value according to the average value of the speech feature in the i-1th time period and the average value of the speech feature in the i-th time period; The number of fluctuations of the speech feature value is determined according to I-1 mean change rates of the speech feature value in I time periods.

7. The method according to claim 1, characterized in that Generating the user's emotional response identifier according to the target emotional characteristic information includes: Based on fourth association information of preset emotion characteristic information and preset emotion response identifier, a target emotion response identifier associated with the emotion characteristic information is determined as the emotion response identifier of the user.

8. The method according to claim 1, characterized in that The target emotion characteristic information includes the first emotion characteristic information of the user in the i-1th time period and the second emotion characteristic information in the i-th time period; Generating the user's emotional response identifier according to the target emotional characteristic information includes: generating emotion change information according to the first emotion characteristic information and the second emotion characteristic information; Based on the association information between the preset emotion change information and the preset emotion coping identifier, the emotion coping identifier associated with the emotion change information is determined as the emotion coping identifier of the user.

9. The method according to claim 8, characterized in that The generating the emotion change information according to the first emotion characteristic information and the second emotion characteristic information includes: In a case where the emotional state of the first emotion characteristic information and the emotional state of the second emotion characteristic information are the same, determining the first emotion change information as the emotion change information, wherein the first emotion change information includes information for reflecting that the emotion of the user is in a stable state; In a case where the emotional state of the first emotional characteristic information is different from the emotional state of the second emotional characteristic information, the emotional change information is determined according to the emotional type and emotional intensity of the first emotional state and the second emotional state.

10. The method according to claim 1, characterized in that The audio information includes second audio information of the service personnel; the method further includes: determining the type of service provided by the service personnel according to the second audio information; A service strategy for the user is generated according to the service type, the target emotion characteristic information and the emotion response identifier.

Images