Report generation methods, apparatus, electronic devices and storage media
By performing emotion recognition and event extraction on multi-round interactive data from children's smart terminals, an emotion change report is generated, which solves the problem that existing technologies cannot reflect psychological state and enables guardians to intuitively understand children's feelings.
Patent Information
- Application Number
- CN202611111972.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-24
- Publication Date
- 2026-11-17
AI Technical Summary
Existing interactive analysis reports for children's smart terminals fail to reflect users' psychological state, and the content is highly technical, making it difficult for guardians to intuitively understand children's true feelings, resulting in a poor reading experience.
By acquiring multi-round user interaction data, emotion recognition and event extraction are performed to generate reports based on emotion type, emotion expression events, and interaction timestamps, including emotional change analysis text and chart displays, thus lowering the reading threshold.
It intuitively and accurately presents users' emotional fluctuations and psychological states, reducing the difficulty for guardians to understand and optimizing the reading experience of the report.
Smart Images

Figure CN123120177A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information processing technology, and in particular to a report generation method, apparatus, electronic device, and storage medium. Background Technology
[0002] Currently, smart terminals typically possess interactive analysis capabilities, meaning they can generate corresponding interaction analysis reports based on human-computer interaction data over a period of time, allowing users to understand their terminal usage. This function is particularly prevalent in the field of children's smart terminals, where the generated analysis reports can be pushed to the guardian's terminal, enabling the guardian to monitor the child's terminal usage.
[0003] However, in the field of children's smart devices, current analytical reports tend to focus on children's learning outcomes, failing to reflect their true psychological state during daily use. Furthermore, the reports are highly technical, demanding a high level of comprehension from guardians, making it difficult for them to intuitively understand children's actual feelings during daily use, resulting in a poor reading experience. Summary of the Invention
[0004] This invention provides a report generation method, apparatus, electronic device, and storage medium to address the shortcomings of interactive analysis reports in related technologies that fail to reflect the user's psychological state.
[0005] This invention provides a report generation method, comprising: Acquire user interaction data from multiple rounds within the first cycle; Emotion recognition and event extraction are performed on the interaction data of each round to obtain the emotion type and emotion expression events of each round of interaction data; Based on the emotion type, emotion expression event, and interaction timestamp of each round of interaction data, a first report is generated that shows the user's emotional changes during the first period.
[0006] According to a report generation method provided by the present invention, the step of generating a first report showing the user's emotional changes within a first period based on the emotion type, emotion expression events, and interaction timestamps of the interaction data in each round includes: Based on the interaction timestamp of each round of interaction data, the second cycle in which each round of interaction data is located is determined, and the first cycle includes multiple second cycles; Based on the sentiment types of multi-round interaction data within each second cycle, the representative sentiment type of each second cycle is determined; Based on the order of the multiple second periods within the first period, the representative emotion type of each second period, and the candidate performance events within each second period, a first report is generated that shows the user's emotional changes within the first period; the candidate performance events are emotional performance events of the interaction data with the emotion type of the representative emotion type.
[0007] According to a report generation method provided by the present invention, the step of generating a first report showing the user's emotional changes in the first period based on the arrangement order of the plurality of second periods in the first period, the representative emotion type of each second period, and the candidate performance events in each second period includes: Based on the arrangement order of the multiple second cycles in the first cycle, the representative emotion types of each second cycle are sorted to obtain the emotion type sequence of the first cycle. Based on the representative emotion types of the multiple second periods, or based on the emotion types of all interaction data within the first period, the percentage of emotion types in the first period is calculated. Based on the order of the multiple second periods in the first period, the representative emotion type of each second period, the candidate performance events in each second period, and the emotion intensity corresponding to the candidate performance events, an emotion change analysis text for the first period is generated; the emotion change analysis text is used to describe the user's emotion changes, emotion change events, and emotion optimization suggestions for the user in the first period. The first report is generated based on at least one of the emotion type sequence, the proportion of the emotion type, and the emotion change analysis text.
[0008] According to a report generation method provided by the present invention, for each second cycle, it further includes: Based on the sentiment types of multiple rounds of interaction data within the second period, the percentage of sentiment types in the second period is calculated. Based on the emotion types and emotion expression events of the multi-round interaction data in the second cycle, representative events for each emotion type are selected from the emotion expression events corresponding to each emotion type. Based on the representative emotion type, the percentage of each emotion type, and at least one representative event for each emotion type in the second period, a second report is generated that reflects the user's emotions in the second period.
[0009] According to a report generation method provided by the present invention, the step of performing emotion recognition and event extraction on each round of interaction data includes: For each round of interaction data, determine the functional category of the agent corresponding to the interaction data; When the function category is communication, emotion recognition and event extraction are performed on the interaction data.
[0010] A report generation method according to the present invention further includes: Based on the multi-round interaction data, the interaction duration of the user for each intelligent agent is calculated. Based on the user's interaction duration for each agent, agent interaction statistics are generated and added to the first report.
[0011] A report generation method according to the present invention further includes: The communication information between the user and the monitoring terminal during the first period is statistically analyzed, and the communication information is added to the first report.
[0012] A report generation method according to the present invention further includes: The first report is presented in chart form.
[0013] The present invention also provides a report generation apparatus, comprising: The acquisition unit is used to acquire user interaction data from multiple rounds within the first cycle. The analysis unit is used to perform emotion recognition and event extraction on the interaction data of each round, and to obtain the emotion type and emotion expression events of each round of interaction data. The generation unit is used to generate a first report showing the user's emotional changes in the first period based on the emotion type, emotion expression event and interaction timestamp of the interaction data in each round.
[0014] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the report generation method as described above.
[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the report generation method as described above.
[0016] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the report generation method as described above.
[0017] The report generation method, apparatus, electronic device, and storage medium provided by this invention perform emotion recognition and event extraction on multi-round interaction data within a first period. This yields the emotion type and emotional expression events of the multi-round interaction data, which, combined with interaction timestamps, generate a first report reflecting the user's emotional changes. This report can intuitively and accurately present the full picture and specific triggers of the user's emotional fluctuations within the first period, thereby indirectly reflecting the user's mental health status. Furthermore, by introducing emotional expression events when generating the first report, the analysis of the user's emotional changes becomes interpretable, effectively lowering the reading and comprehension threshold for guardians and optimizing the report reading experience. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in this invention or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 This is a flowchart illustrating the report generation method provided by the present invention.
[0020] Figure 2 This is a schematic diagram of the display interface of the second report provided by the present invention.
[0021] Figure 3 This is a schematic diagram of the display interface for intelligent agent interaction statistics provided by the present invention.
[0022] Figure 4 This is a schematic diagram of the communication information display interface provided by the present invention.
[0023] Figure 5 This is a schematic diagram of the report selection interface provided by the present invention.
[0024] Figure 6 This is a schematic diagram of the report generation device provided by the present invention.
[0025] Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation
[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0027] All actions involving the acquisition of signal information or data in this invention are carried out in compliance with the relevant data protection laws and policies of the country where the device is located, and with the authorization granted by the owner of the device.
[0028] A smart terminal refers to a portable computing device that has an independent operating system, can access a network, and supports users in installing and running third-party applications. Examples include smartphones, tablets, dictionary pens, and smartwatches.
[0029] Most current smart terminals possess interactive analysis capabilities, meaning they can generate interaction analysis reports based on human-computer interaction data over a period of time, allowing users to understand their terminal usage. Interaction analysis functionality is even more widely used in the field of children's smart terminals, where the generated reports are typically called growth reports. Furthermore, to facilitate guardians' understanding of children's terminal usage, these growth reports are usually sent to third-party devices such as the guardian's mobile phone for display.
[0030] However, most current growth reports tend to analyze and statistically summarize what knowledge children have learned using smart devices, emphasizing children's learning outcomes, but lack tracking and analysis of the dynamic changes in children's actual feelings while using them. At the same time, the content analysis of current growth reports is often too professional, and the various professional learning data or psychological assessments create a high reading threshold, making it difficult for guardians to intuitively and easily understand children's actual feelings in daily use, resulting in a poor reading experience for growth reports.
[0031] To address this, the present invention provides a report generation method, which can be executed by a smart terminal or cloud server with data processing capabilities. Specifically, the smart terminal can be a smartphone, tablet computer, dictionary pen, smartwatch, learning machine, object recognition device, translator, etc.
[0032] Figure 1 This is a flowchart illustrating the report generation method provided by the present invention, as shown below. Figure 1 As shown, the method includes: Step 110: Obtain user interaction data from multiple rounds within the first cycle.
[0033] Here, the first period refers to the time interval used for statistics and report generation. The duration of the first period can be a system default preset or a user-defined setting. For example, the first period can be one day, one week, one month, etc.
[0034] Interaction data refers to the historical data of human-computer interaction between a user and a smart terminal during the user's use of the smart terminal. Each round of human-computer interaction generates one round of interaction data. Interaction data can include information such as the user's voice commands, text input, and touch click records.
[0035] The user's multi-round interaction data within the first cycle refers to the interaction data recorded by the smart terminal within the time frame defined by the first cycle, representing the multiple human-computer interaction operations performed by the user. For example, when generating the first cycle report based on preset rules or under the operation of a guardian, the interaction data generated from all rounds of human-computer interaction between the user and the smart terminal within the first cycle can be obtained.
[0036] Step 120: Perform emotion recognition and event extraction on the interaction data of each round to obtain the emotion type and emotion expression event of each round of interaction data.
[0037] Specifically, after obtaining the multi-round interaction data in the first cycle, it is necessary to perform sentiment analysis and event extraction for each round of interaction data.
[0038] Furthermore, for a round of interaction data, emotion recognition algorithms can be used to analyze and identify the user emotions reflected in that round of interaction data. Specifically, if the interaction data includes user speech, the type of emotion reflected in the user speech can be identified by analyzing acoustic features such as speech rate, pitch, and volume. Alternatively, the user speech can be transcribed, and the transcribed text can be used to identify the emotion type based on natural language processing algorithms. Or, if the interaction data includes user input text, the emotion type can also be identified based on natural language processing algorithms, such as algorithms that extract the emotion type of each word segment in the text. If the interaction data includes user facial videos or images, image classification algorithms can also be used to identify the emotion type of the user facial videos or images. This embodiment of the invention does not specifically limit the specific methods used in this regard.
[0039] Therefore, the emotion type of this round of interaction data can be obtained. The emotion type here can include three types: positive, neutral, and negative, or it can be a more specific emotion type, such as happy, excited, disappointed, impatient, sad, etc.
[0040] Furthermore, for a round of interaction data, events reflected in that round of interaction data can be extracted based on event extraction algorithms. Specifically, if the interaction data includes user speech, speech-to-text transcription can be performed, and event extraction can be performed on the transcribed text using natural language processing (NLP) algorithms. Similarly, if the interaction data includes user input text, event extraction can also be performed on the user input text using NLP algorithms. These NLP algorithms can include named entity recognition, dependency parsing, and other algorithms. Through NLP algorithms, actions, content, or topics that trigger or accompany the aforementioned emotional types can be extracted from a round of interaction data, especially from user speech, user input text, and the agent's responses, as emotional expression events in the interaction data.
[0041] It is understandable that, since both emotion type and emotion expression event refer to a single round of interaction data, meaning that the data source for emotion type and emotion expression event is the same round of interaction data, there is a natural connection between emotion type and emotion expression event, and emotion expression event can be understood as the trigger for emotion type.
[0042] For example, in one round of interaction data, a child user cheerfully says to the smart terminal, "I heard a really funny joke today, let me tell you about it." Through emotion recognition, the emotion type of this round of interaction data can be determined to be positive; through event extraction, the emotional expression event of this round of interaction data can be extracted as listening to a joke.
[0043] Step 130: Based on the emotion type, emotion expression event, and interaction timestamp of each round of interaction data, generate a first report showing the user's emotional changes within the first period.
[0044] Specifically, each round of interaction data carries a corresponding interaction timestamp when it is generated. For each round of interaction data, the interaction timestamp records the specific time point when the interaction occurred. By combining the interaction timestamps, emotion types, and emotional expression events of each round of interaction data in the first cycle, the originally discrete interaction data in the first cycle can be linked together in chronological order to reconstruct the dynamic changes in the user's psychological emotions in the first cycle, thus obtaining the first report.
[0045] Here, the first report is an interactive analysis report that shows the dynamic changes in the user's emotional type during the first period, and can also be understood as the user's growth report during the first period.
[0046] Compared to traditional growth reports that focus on analyzing user learning content or simply listing usage time, this invention's first report uses the interaction timestamps and emotion types of each round of interaction data within the first cycle. Therefore, the first report can objectively and accurately show the emotional fluctuations the user experienced during the first cycle, such as when they were in a positive state and when they were in a negative state. Furthermore, since the emotion type and emotional expression events of each round of interaction data are naturally correlated, the first report can not only show the emotional fluctuations the user experienced during the first cycle but also link them to specific emotional expression events, thereby showing the reasons for the emotional fluctuations. For example, a positive emotion may arise from listening to a joke, while a negative emotion may result from being criticized by a teacher. This allows guardians to intuitively understand the child user's emotional changes and their causes when reading the first report.
[0047] In the method provided in this embodiment of the invention, emotion recognition and event extraction are performed on the multi-round interaction data within the first period, thereby obtaining the emotion type and emotion expression events of the multi-round interaction data. Combined with the interaction timestamps, a first report reflecting the user's emotional changes is generated. This can intuitively and accurately present the full picture and specific triggers of the user's emotional fluctuations within the first period, thereby indirectly reflecting the user's mental health status. Furthermore, by introducing emotion expression events when generating the first report, the analysis of the user's emotional changes becomes interpretable, effectively reducing the reading and comprehension threshold for guardians and optimizing the report reading experience.
[0048] Based on the above embodiments, step 130, which involves generating a first report reflecting the user's emotional changes within the first period based on the emotion type, emotional expression events, and interaction timestamps of each round of interaction data, includes: Based on the interaction timestamp of each round of interaction data, the second cycle in which each round of interaction data is located is determined, and the first cycle includes multiple second cycles; Based on the sentiment types of multi-round interaction data within each second cycle, the representative sentiment type of each second cycle is determined; Based on the order of the multiple second periods within the first period, the representative emotion type of each second period, and the candidate performance events within each second period, a first report is generated that shows the user's emotional changes within the first period; the candidate performance events are emotional performance events of the interaction data with the emotion type of the representative emotion type.
[0049] Specifically, a first period with a large time span can be divided into multiple second periods with smaller time spans. For example, when the first period is one week, the second period can be one day, so the first period can consist of 7 second periods; when the first period is one month, the second period can be one day or one week, and this embodiment of the invention does not limit this.
[0050] After obtaining the multi-round interaction data within the first cycle, the interaction timestamps inherent in each round of interaction data can be used to determine which second cycle within the first cycle the interaction occurred in, thus obtaining the second cycle in which each round of interaction data belongs. For example, if the first cycle is one week and the second cycle is one day, all interaction data within one week can be categorized into seven second cycles, such as Monday, Tuesday, Wednesday, etc., thereby completing the interaction data clustering based on interaction timestamps.
[0051] After obtaining the interaction data for each second cycle, the emotion types can be summarized for each second cycle, thus obtaining the representative emotion type for each second cycle. Here, the representative emotion type reflects the representative emotion type under multiple rounds of interaction within the second cycle, and can also be understood as the user's main emotion type within the second cycle.
[0052] Furthermore, the emotion type that appears most frequently in the second period can be used as the representative emotion type. For example, if there are 10 interactions in the second period, and 7 of them are positive, then the representative emotion type is positive. Alternatively, the emotion type with the highest emotional intensity in the second period can be used as the representative emotion type. This embodiment of the invention does not specifically limit this.
[0053] Furthermore, for each representative emotion type in the second period, the reasons for its generation can be traced, i.e., candidate performance events for each second period can be obtained. Here, candidate performance events specifically refer to the emotion performance events corresponding to interactive data whose emotion type matches the representative emotion type of the second period. For example, if the representative emotion type of a certain second period is positive, then interactive data with the same positive emotion type can be selected from the interactive data of that second period, and their corresponding emotion performance events can be used as candidate performance events.
[0054] Based on this, a first report can be generated by combining the order of multiple second cycles within the first cycle, the representative emotion type of each second cycle, and the corresponding candidate performance events. For example, the order of multiple second cycles within the first cycle, the representative emotion type of each second cycle, and the corresponding candidate performance events can all be input into a large language model, which will then combine this information to generate the first report.
[0055] In this embodiment of the invention, by dividing a long first period into multiple second periods and extracting representative emotion types and associated candidate performance events for each second period, a first report is generated in chronological order. This effectively compresses massive, fragmented interaction data within the first period into representative emotion types and candidate performance events across multiple second periods, filtering out redundant information that lacks representativeness. Furthermore, by combining the order of the multiple second periods, emotion change analysis is achieved using the second period as the unit of analysis. The resulting first report is more organized and its readability is improved.
[0056] Based on any of the above embodiments, step 130, which involves generating a first report demonstrating the user's emotional changes within the first period based on the order of the plurality of second periods within the first period, the representative emotion type of each second period, and the candidate performance events within each second period, includes: Based on the arrangement order of the multiple second cycles in the first cycle, the representative emotion types of each second cycle are sorted to obtain the emotion type sequence of the first cycle. Based on the representative emotion types of the multiple second periods, or based on the emotion types of all interaction data within the first period, the percentage of emotion types in the first period is calculated. Based on the order of the multiple second periods in the first period, the representative emotion type of each second period, the candidate performance events in each second period, and the emotion intensity corresponding to the candidate performance events, an emotion change analysis text for the first period is generated; the emotion change analysis text is used to describe the user's emotion changes, emotion change events, and emotion optimization suggestions for the user in the first period. The first report is generated based on at least one of the emotion type sequence, the proportion of the emotion type, and the emotion change analysis text.
[0057] Specifically, after obtaining the representative emotion type for each second cycle within the first cycle, these second cycles can be arranged according to their chronological order within the first cycle, thereby generating an emotion type sequence reflecting the fluctuation of emotions over time. In some embodiments, the emotion type sequence can be displayed as a timeline or line graph, showing the representative emotion type of each second cycle over time within the first cycle. For example, each node on the timeline corresponds to an emoji icon representing the emotion type of a second cycle, thus clearly presenting the user's mood fluctuation trajectory within the first cycle.
[0058] Furthermore, after obtaining the representative emotion type for each second period within the first period, the frequency or duration of each representative emotion type can be summarized and statistically analyzed to calculate the percentage of each representative emotion type in the entire first period, denoted as the emotion type percentage. Alternatively, the frequency or duration of emotion types in all interaction data within the first period can be summarized and statistically analyzed to calculate the percentage of each emotion type in the entire first period, denoted as the emotion type percentage.
[0059] Understandably, the resulting percentage of emotion types can be interpreted as the percentage of each emotion state within the total number of interactions or the total duration of interactions during the first period. For example, the percentage of emotion types can be displayed in the form of a pie chart.
[0060] Furthermore, based on the order of all second periods within the first period, the representative emotion type of each second period, the candidate performance events within each second period, and the emotion intensity corresponding to the candidate performance events, the emotion change analysis text for the first period can be generated together. Here, emotion intensity reflects the intensity of the user's emotion when the candidate performance event occurs. Emotion intensity can be obtained through emotion recognition; that is, the emotion intensity corresponding to the candidate performance event can be the intensity of the emotion type obtained by emotion recognition of the interaction data from which the candidate performance event originates. For example, a large language model or a preset generation template can be used, combined with the aforementioned order, representative emotion type, candidate performance time, and emotion intensity, to generate the emotion change text.
[0061] The application of emotion intensity can assign influence weights to multiple candidate events representing the same emotion type when generating emotion change analysis text. This facilitates the extraction of the candidate event with the greatest impact on triggering the representative emotion type, which can then be used as an example of emotion change events in the emotion change analysis text. It is understandable that the higher the emotion intensity of the candidate event, the greater its impact on triggering the representative emotion type.
[0062] The resulting emotional change analysis text can describe the user's emotional changes during the first period, such as "The child's overall mood was positive this week, but there was a brief period of low mood on Wednesday." It can also describe the events that led to the emotional changes, such as "On Wednesday, the child mentioned that his mother did not reward him for his good grades, but on Friday, he became positive because he was praised by the teacher for reciting the text well." In addition, the emotional change analysis text can also provide suggestions for optimizing the user's emotions, such as "It is recommended that parents communicate with their children in a timely manner to understand their psychological state."
[0063] Based on this, at least one of the following can be formatted and combined: emotion type sequence, emotion type proportion, and emotion change analysis text, thus obtaining the first report.
[0064] In this embodiment of the invention, by constructing an emotion type sequence, statistically analyzing the proportion of emotion types, and generating at least one of emotion change analysis text, a foundation is provided for generating an easy-to-understand and richly illustrated first report, which helps to further improve the readability of the first report and thus optimize the reading experience of the first report.
[0065] Based on any of the above embodiments, for each second period, the report generation method further includes: Based on the sentiment types of multiple rounds of interaction data within the second period, the percentage of sentiment types in the second period is calculated. Based on the emotion types and emotion expression events of the multi-round interaction data in the second cycle, representative events for each emotion type are selected from the emotion expression events corresponding to each emotion type. Based on the representative emotion type, the percentage of each emotion type, and at least one representative event for each emotion type in the second period, a second report is generated that reflects the user's emotions in the second period.
[0066] Specifically, for each second cycle under the first cycle, the frequency or duration of each emotion type can be statistically analyzed based on the emotion types of all interaction data within the second cycle, thereby obtaining the proportion of each emotion type in the second cycle. This proportion refers to the percentage of each emotion state within the total number of interactions or the total interaction duration in the second cycle.
[0067] Furthermore, for each emotion type that emerges during the second period, representative emotion expression events can be extracted from multiple corresponding emotion expression events for each emotion type, thus obtaining representative events for each emotion type. For example, for an emotion type during the second period, the emotion expression events of all interaction data corresponding to that emotion type can be obtained. Based on this, the emotion expression event with the highest emotional intensity can be selected as the representative event for that emotion type.
[0068] Based on this, the representative emotion types, their percentages, and at least one representative event for each emotion type can be formatted and combined to obtain a second report for the second period. Here, the second report is a more granular interactive analysis report with a smaller time span compared to the first report. For example, if the first period is one week and the second period is one day, the first report could be a weekly report, while the second report could be a daily report.
[0069] For example, Figure 2This is a schematic diagram showing the interface of the second report provided by the present invention. Figure 2 In this context, a second cycle is defined as one day, hence the second report is designated as a daily report. Figure 2 The daily report for the 25th identified positive emotion icons representing the day's mood and displayed the percentage of each emotion type in a pie chart. Specifically, on the 25th, 65% of users experienced positive emotions, 25% experienced neutral emotions, and 10% experienced negative emotions. Furthermore, the report also showed representative events for each of the positive, neutral, and negative emotions. For example, a representative event for positive emotions was "I was praised by my Chinese teacher for reciting the text well today!"; a representative event for neutral emotions was "I went to class with Xiaohong today"; and a representative event for negative emotions was "I got 100 on my math test today, but my mom didn't say anything. That's so unfair!"
[0070] In this embodiment of the invention, by analyzing the interaction data within the second cycle, a second report is generated on a second cycle basis, enabling guardians to understand the emotional distribution and emotional triggers of child users in a granular manner within each second cycle.
[0071] Based on any of the above embodiments, step 120, which involves performing emotion recognition and event extraction on the interaction data for each round, includes: For each round of interaction data, determine the functional category of the agent corresponding to the interaction data; When the function category is communication, emotion recognition and event extraction are performed on the interaction data.
[0072] Specifically, an intelligent agent is an artificial intelligence interactive object with an independent character and personality, pre-configured in a smart terminal. Users can interact with various intelligent agents built into the smart terminal while using it. For example, a smart terminal can have multiple built-in intelligent agents, such as Alpha Egg, Sister Parrot, Dr. Science, and Hug Bear.
[0073] Each intelligent agent is assigned a corresponding application scenario or capability label when configured, which is understood as the functional category of the intelligent agent in this embodiment of the invention. For example, Alpha Egg is positioned as a voice assistant, with functions such as voice interaction and device control, and supports casual communication with users; therefore, Alpha Egg's functional category includes the communication category. Parrot Sister is positioned to play games such as riddles and idiom chain games; therefore, Parrot Sister's functional category is the game category. Science Doctor is positioned to provide users with popular science knowledge; therefore, Science Doctor's functional category is the knowledge category. Hug Bear is positioned to be a child's emotional support channel, providing emotional comfort to children; therefore, Hug Bear's functional category also includes the communication category.
[0074] When users interact with non-communication-type agents, they typically focus on the specific functions provided by the agent itself and do not express their own emotions and experiences during the interaction. The interaction data generated by interacting with such agents usually cannot accurately reflect the user's emotions.
[0075] Therefore, after acquiring multiple rounds of user interaction data within the first cycle, the agent corresponding to each round of interaction data can be identified, thereby determining the functional category of the agent corresponding to each round of interaction data. In this case, to ensure the reliability of emotion recognition and emotion event extraction, emotion recognition and event extraction can be performed only on the interaction data corresponding to agents whose functional category is communication, and not on the interaction data corresponding to agents whose functional category is not communication. This filters out the impact of interaction data that does not contain user emotional components on the analysis and tracking of user emotion fluctuations.
[0076] In this embodiment of the invention, emotion recognition and event extraction are performed on the interaction data only when the function category of the intelligent agent is communication. This filters out interaction data that is not related to emotion, saving the computing resources and computing power required to generate the report. Furthermore, applying interaction data with emotional communication attributes can effectively improve the accuracy of emotion recognition and subsequent report generation, and avoid the report generation being interfered with by redundant data.
[0077] Based on any of the above embodiments, the report generation method further includes: Based on the multi-round interaction data, the interaction duration of the user for each intelligent agent is calculated. Based on the user's interaction duration for each agent, agent interaction statistics are generated and added to the first report.
[0078] Specifically, in cases where a smart terminal has multiple built-in agents, one agent participates in each round of human-computer interaction. Therefore, after acquiring multiple rounds of interaction data within the first cycle, the agent corresponding to each round of interaction data can be identified. Based on the agent corresponding to each round of interaction data and the duration of each round of interaction data itself, the interaction time between the user and each agent within the first cycle can be calculated. This interaction time is the cumulative duration of interaction between the user and the agents within the first cycle.
[0079] After obtaining the interaction duration of each agent, the interaction durations of all agents can be aggregated to obtain agent interaction statistics. These statistics can include the absolute interaction duration of each agent, such as the cumulative number of minutes or hours, as well as information such as the percentage of each agent's interaction duration in the total interaction duration, the most frequently used agent, and other relevant information.
[0080] Subsequently, the aforementioned agent interaction statistics can be added to the first report as part of it. For example, the agent interaction statistics can be included in the first report as data representing the agent's interaction time dimension.
[0081] For example, Figure 3 This is a schematic diagram of the display interface for intelligent agent interaction statistics provided by the present invention. Figure 3 The agent interaction statistics shown are part of the information presented in the first report. Figure 3 In this embodiment, the agent interaction statistics are displayed in the form of a bar chart, showing the interaction duration of each agent, and the position of each agent in the bar chart is identified by its avatar. Furthermore, the agent interaction statistics may also include a summary of some agents' dialogues, such as the number of knowledge points asked to the science popularizer or the number of voice games played with the parrot sister; however, this embodiment does not impose specific limitations on this.
[0082] In this embodiment of the invention, by statistically analyzing the interaction duration of multiple intelligent agents built into the smart device and adding it to the first report, the interaction status of the intelligent agents within the first period can be objectively and accurately reflected in the first report, thereby reflecting the user's interests and preferences and helping guardians understand the development direction of the child user's interests.
[0083] Based on any of the above embodiments, the report generation method further includes: The communication information between the user and the monitoring terminal during the first period is statistically analyzed, and the communication information is added to the first report.
[0084] Here, the monitoring terminal refers to a third-party terminal device associated with the user's currently used smart terminal or account, typically a smartphone, tablet, or other smart terminal used by the guardian. Communication information between the user and the monitoring terminal during the first period refers to various two-way or one-way communication data generated between the user and the monitoring terminal through the smart terminal during the first period. Specific types of communication information may include voice messages, video call logs, and voice call records.
[0085] Based on this, the statistically obtained communication information can be added to the first report as part of it.
[0086] For example, Figure 4 This is a schematic diagram of the communication information display interface provided by the present invention. Figure 4 The communication information shown is part of the first report. Figure 4 The communication information lists the number of video calls and voice calls made and received, as well as the cumulative call duration, and also lists the identifier of the monitoring terminal that made the most calls, such as "Dad".
[0087] In this embodiment of the invention, by statistically analyzing the communication information between the user and the monitoring terminal within the first period and adding it to the first report, the frequency of parent-child interaction can be presented intuitively in the first report.
[0088] Based on any of the above embodiments, the report generation method further includes: The first report is presented in chart form.
[0089] Specifically, after receiving the first report, the various information in the first report can be displayed in chart form. Here, chart form refers to using graphical and visual data representation methods to replace or supplement plain text descriptions. Chart forms can specifically include pie charts, line charts, bar charts, tables, and visualization panels with emoticons, etc.
[0090] For example, in Figure 2 In the text, various emojis are used to represent positive, neutral, and negative emotions, and pie charts are used to show the percentage of each emotion type in the second cycle; in Figure 3 In the middle, a bar chart is used to compare the cumulative time of user interaction with different intelligent agents.
[0091] In this embodiment of the invention, the first report is presented in the form of charts, which lowers the reading threshold of the first report and enables guardians to quickly grasp the true psychological state and usage preferences of child users, greatly optimizing the guardians' reading experience and application feel.
[0092] Based on any of the above embodiments, this embodiment of the invention provides a report generation method, including: First, acquire multi-round interaction data within the first or second cycle.
[0093] Here, the first cycle can be a week, a month, etc., and the second cycle can be a day, a week, etc., with the time span of the first cycle being longer than that of the second cycle. In this embodiment of the invention, an example is given with the first cycle being a week and the second cycle being a day.
[0094] Secondly, based on the multi-round interaction data within the first cycle, generate the first report for the first cycle; based on the multi-round interaction data within the second cycle, generate the second report for the second cycle.
[0095] Here, the first report can be a weekly report; the second report can be a daily report.
[0096] Furthermore, the weekly report includes statistics on agent interactions, communication information, and emotional change data from a mental health perspective. Specifically, the agent interaction statistics include the cumulative interaction time between the user and each agent; the communication information includes statistical analysis results of communication data between the user and the monitoring terminal; and the emotional change data includes a weekly sequence of emotion types, the percentage of each emotion type over the week, representative events for each emotion type within the week, and a weekly emotional change analysis text. For example, the emotional change analysis text could be: "Great job! Children are the little ambassadors of life's enthusiasm." The child's mood fluctuated somewhat this week. The first half of the week was happier because the language arts teacher praised the child for reciting the text well in class; the second half of the week was a little down, mentioning that the mother did not reward the child for getting a perfect score of 100. Parents are advised to notice this in time, communicate with the child, understand the child's psychological state, and intervene as appropriate.
[0097] In some embodiments, the display interface for sentiment change data in the weekly report is similar to... Figure 2 The second report shown has a similar interface; the sentiment type sequence in the weekly report can be... Figure 2 The report displays the emotions in a sequence, arranged from first to last, according to the emotion type each day; the percentage of each emotion type in the weekly report can be... Figure 2 The pie chart in the report is used to illustrate the emotions; representative events for each emotion type in the weekly report can be presented in the form of... Figure 2 The emotional change analysis text in the weekly report can be displayed in a text box on the interface afterward.
[0098] In addition, the daily report can include users' sentiment analysis data, specifically including the user's representative emotion type for the day, the percentage of each emotion type for the day, and representative events for each emotion type.
[0099] Next, it can receive the user's display selection instruction and display the report corresponding to the above display selection instruction.
[0100] For example, Figure 5 This is a schematic diagram of the report selection interface provided by the present invention. Figure 5 The system displays daily and weekly reports with their own icons. Users can tap the icon to send a selection command to the smart device. Upon receiving the selection command, the smart device can then display the daily or monthly report as an icon, thus enabling selective display of growth reports for children.
[0101] In this embodiment of the invention, a growth report reflecting the user's emotional changes can be generated, thereby intuitively and accurately presenting the full picture and specific causes of the user's emotional fluctuations over a period of time, thus indirectly reflecting the user's mental health status. Furthermore, by introducing emotional expression events when generating the first report, the analysis of the user's emotional changes becomes interpretable, effectively reducing the reading and comprehension threshold for guardians and optimizing the report reading experience.
[0102] The report generation apparatus provided by the present invention is described below. The report generation apparatus described below and the report generation method described above can be referred to in correspondence.
[0103] Figure 6 This is a schematic diagram of the report generation device provided by the present invention, as shown below. Figure 6 As shown, the device includes: Acquisition unit 610 is used to acquire user interaction data from multiple rounds within the first cycle; Analysis unit 620 is used to perform emotion recognition and event extraction on each round of interaction data to obtain the emotion type and emotion expression events of each round of interaction data. The generation unit 630 is used to generate a first report showing the user's emotional changes in the first period based on the emotion type, emotion expression event and interaction timestamp of the interaction data in each round.
[0104] The device provided in this invention performs emotion recognition and event extraction on multi-round interaction data within a first cycle, thereby obtaining the emotion type and emotional expression events of the multi-round interaction data. Combined with interaction timestamps, it generates a first report reflecting the user's emotional changes. This report can intuitively and accurately present the full picture and specific triggers of the user's emotional fluctuations within the first cycle, thus indirectly reflecting the user's mental health status. Furthermore, by introducing emotional expression events when generating the first report, the analysis of the user's emotional changes becomes interpretable, effectively reducing the reading and comprehension threshold for guardians and optimizing the report reading experience.
[0105] Based on the above embodiments, the generation unit is specifically used for: Based on the interaction timestamp of each round of interaction data, the second cycle in which each round of interaction data is located is determined, and the first cycle includes multiple second cycles; Based on the sentiment types of multi-round interaction data within each second cycle, the representative sentiment type of each second cycle is determined; Based on the order of the multiple second periods within the first period, the representative emotion type of each second period, and the candidate performance events within each second period, a first report is generated that shows the user's emotional changes within the first period; the candidate performance events are emotional performance events of the interaction data with the emotion type of the representative emotion type.
[0106] Based on any of the above embodiments, the generating unit is specifically used for: Based on the arrangement order of the multiple second cycles in the first cycle, the representative emotion types of each second cycle are sorted to obtain the emotion type sequence of the first cycle. Based on the representative emotion types of the multiple second periods, or based on the emotion types of all interaction data within the first period, the percentage of emotion types in the first period is calculated. Based on the order of the multiple second periods in the first period, the representative emotion type of each second period, the candidate performance events in each second period, and the emotion intensity corresponding to the candidate performance events, an emotion change analysis text for the first period is generated; the emotion change analysis text is used to describe the user's emotion changes, emotion change events, and emotion optimization suggestions for the user in the first period. The first report is generated based on at least one of the emotion type sequence, the proportion of the emotion type, and the emotion change analysis text.
[0107] Based on any of the above embodiments, the generating unit is further configured to: Based on the sentiment types of multiple rounds of interaction data within the second period, the percentage of sentiment types in the second period is calculated. Based on the emotion types and emotion expression events of the multi-round interaction data in the second cycle, representative events for each emotion type are selected from the emotion expression events corresponding to each emotion type. Based on the representative emotion type, the percentage of each emotion type, and at least one representative event for each emotion type in the second period, a second report is generated that reflects the user's emotions in the second period.
[0108] Based on any of the above embodiments, the analysis unit is specifically used for: For each round of interaction data, determine the functional category of the agent corresponding to the interaction data; When the function category is communication, emotion recognition and event extraction are performed on the interaction data.
[0109] Based on any of the above embodiments, the generating unit is further configured to: Also includes: Based on the multi-round interaction data, the interaction duration of the user for each intelligent agent is calculated. Based on the user's interaction duration for each agent, agent interaction statistics are generated and added to the first report.
[0110] Based on any of the above embodiments, the generating unit is further configured to: The communication information between the user and the monitoring terminal during the first period is statistically analyzed, and the communication information is added to the first report.
[0111] Based on any of the above embodiments, the device further includes a display unit, used for: The first report is presented in chart form.
[0112] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7 As shown, the electronic device may include: a processor 710, a communications interface 720, a memory 730, and a communication bus 740, wherein the processor 710, the communications interface 720, and the memory 730 communicate with each other via the communication bus 740. The processor 710 can call logical instructions in the memory 730 to execute a report generation method, which includes: Acquire user interaction data from multiple rounds within the first cycle; Emotion recognition and event extraction are performed on the interaction data of each round to obtain the emotion type and emotion expression events of each round of interaction data; Based on the emotion type, emotion expression event, and interaction timestamp of each round of interaction data, a first report is generated that shows the user's emotional changes during the first period.
[0113] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to related technologies, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0114] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, wherein when the computer program is executed by a processor, the computer is able to execute the report generation method provided by the above methods, the method comprising: Acquire user interaction data from multiple rounds within the first cycle; Emotion recognition and event extraction are performed on the interaction data of each round to obtain the emotion type and emotion expression events of each round of interaction data; Based on the emotion type, emotion expression event, and interaction timestamp of each round of interaction data, a first report is generated that shows the user's emotional changes during the first period.
[0115] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the report generation methods provided by the methods described above, the method comprising: Acquire user interaction data from multiple rounds within the first cycle; Emotion recognition and event extraction are performed on the interaction data of each round to obtain the emotion type and emotion expression events of each round of interaction data; Based on the emotion type, emotion expression event, and interaction timestamp of each round of interaction data, a first report is generated that shows the user's emotional changes during the first period.
[0116] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0117] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of software products. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A report generation method, characterized in that, include: Acquire user interaction data from multiple rounds within the first cycle; Emotion recognition and event extraction are performed on the interaction data of each round to obtain the emotion type and emotion expression events of each round of interaction data; Based on the emotion type, emotion expression event, and interaction timestamp of each round of interaction data, a first report is generated that shows the user's emotional changes during the first period.
2. The report generation method according to claim 1, characterized in that, The first report, generated based on the emotion type, emotion expression events, and interaction timestamps of each round of interaction data, represents the user's emotional changes within the first period, including: Based on the interaction timestamp of each round of interaction data, the second cycle in which each round of interaction data is located is determined, and the first cycle includes multiple second cycles; Based on the sentiment types of multi-round interaction data within each second cycle, the representative sentiment type of each second cycle is determined. Based on the order of the multiple second periods within the first period, the representative emotion type of each second period, and the candidate performance events within each second period, a first report is generated that shows the user's emotional changes within the first period; the candidate performance events are emotional performance events of the interaction data with the emotion type of the representative emotion type.
3. The report generation method according to claim 2, characterized in that, The process of generating a first report on the user's emotional changes within the first period, based on the order of the multiple second periods within the first period, the representative emotion type of each second period, and the candidate performance events within each second period, includes: Based on the arrangement order of the multiple second cycles in the first cycle, the representative emotion types of each second cycle are sorted to obtain the emotion type sequence of the first cycle. Based on the representative emotion types of the multiple second periods, or based on the emotion types of all interaction data within the first period, the percentage of emotion types in the first period is calculated. Based on the order of the multiple second periods in the first period, the representative emotion type of each second period, the candidate performance events in each second period, and the emotion intensity corresponding to the candidate performance events, an emotion change analysis text for the first period is generated; the emotion change analysis text is used to describe the user's emotion changes, emotion change events, and emotion optimization suggestions for the user in the first period. The first report is generated based on at least one of the emotion type sequence, the proportion of the emotion type, and the emotion change analysis text.
4. The report generation method according to claim 2, characterized in that, For each second cycle, it also includes: Based on the sentiment types of multiple rounds of interaction data within the second period, the percentage of sentiment types in the second period is calculated. Based on the emotion types and emotion expression events of the multi-round interaction data in the second cycle, representative events for each emotion type are selected from the emotion expression events corresponding to each emotion type. Based on the representative emotion type, the percentage of each emotion type, and at least one representative event for each emotion type in the second period, a second report is generated that reflects the user's emotions in the second period.
5. The report generation method according to any one of claims 1 to 4, characterized in that, The process of performing emotion recognition and event extraction on each round of interaction data includes: For each round of interaction data, determine the functional category of the agent corresponding to the interaction data; When the function category is communication, emotion recognition and event extraction are performed on the interaction data.
6. The report generation method according to any one of claims 1 to 4, characterized in that, Also includes: Based on the multi-round interaction data, the interaction duration of the user for each intelligent agent is calculated. Based on the user's interaction duration for each agent, agent interaction statistics are generated and added to the first report.
7. The report generation method according to any one of claims 1 to 4, characterized in that, Also includes: The communication information between the user and the monitoring terminal during the first period is statistically analyzed, and the communication information is added to the first report.
8. The report generation method according to any one of claims 1 to 4, characterized in that, Also includes: The first report is presented in chart form.
9. A report generation device, characterized in that, include: The acquisition unit is used to acquire user interaction data from multiple rounds within the first cycle. The analysis unit is used to perform emotion recognition and event extraction on the interaction data of each round, and to obtain the emotion type and emotion expression events of each round of interaction data. The generation unit is used to generate a first report showing the user's emotional changes in the first period based on the emotion type, emotion expression event and interaction timestamp of the interaction data in each round.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the report generation method as described in any one of claims 1 to 8.
11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the report generation method as described in any one of claims 1 to 8.
12. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the report generation method as described in any one of claims 1 to 8.
