Intrinsic emotion analysis system based on psychological information big data

Abstract

The present invention relates to the field of intrinsic data analysis technology, and in particular to an intrinsic emotion analysis system based on big data of psychological information. The system comprises: an emotion topic modeling module, a dynamic emotion driving factor analysis module, an intrinsic emotion abnormal trend identification module, and an abnormal emotion strategy adjustment module; the module is used to obtain psychological research data to perform emotion topic modeling and dynamic emotion driving factor analysis to obtain dynamic emotion driving factor data; based on the dynamic emotion driving factor data, the module analyzes the emotion transition states of different emotion stages to obtain emotion transition state data; the module identifies intrinsic emotion abnormal trends on the emotion transition state data and constructs a big data intrinsic emotion analysis model to obtain an intrinsic emotion abnormal analysis model and formulates abnormal intrinsic emotion strategy adjustments to obtain abnormal emotion strategy adjustment data. The present invention makes emotion analysis more accurate by optimizing intrinsic emotion analysis technology.

Description

Technical Field

[0001] The present invention relates to the technical field of intrinsic data analysis, and in particular to an intrinsic emotion analysis system based on psychology information big data. Background Art

[0002] Intrinsic emotions refer to the underlying emotional experiences within an individual, influenced and regulated by a variety of factors. Research based on big data from psychological information has revealed the diverse contexts of intrinsic emotions, including individual psychological characteristics, social environment, and life experiences. Regarding psychological characteristics, factors such as personality traits, emotion regulation ability, and cognitive style have a significant impact on intrinsic emotions. For example, optimistic individuals are more likely to experience positive internal emotions, while individuals with less emotional stability are more susceptible to negative emotions due to external events. The social environment is also a crucial context for intrinsic emotions. Factors such as family, work, and social relationships all influence individual emotional experiences. For example, individuals with harmonious and supportive families tend to experience greater happiness and satisfaction. Research based on big data from psychological information not only reveals the complex factors underlying intrinsic emotions but also provides important insights for understanding and intervening in individual emotional experiences. However, traditional intrinsic emotion analysis systems based on big data from psychological information often rely on static data, making it difficult to capture dynamic changes in emotions and unable to accurately analyze the emotional content. Summary of the Invention

[0003] Based on this, it is necessary to provide an intrinsic emotion analysis system based on psychological information big data to solve at least one of the above technical problems.

[0004] To achieve the above objectives, an intrinsic emotion analysis system based on psychology information big data is proposed, which includes the following modules:

[0005] The emotional topic modeling module is used to obtain psychological research data; perform semantic analysis on the psychological research data to obtain psychological semantic data; perform emotional topic modeling on the psychological research data based on the psychological semantic data to obtain emotional topic modeling data;

[0006] The dynamic emotion driving factor analysis module is used to perform dynamic emotion segmentation processing on the psychological semantic data based on the emotion topic modeling data to obtain dynamic emotion segmentation data; perform local emotion structure feature analysis on the dynamic emotion segmentation data to obtain local emotion structure feature data; and perform dynamic emotion driving factor analysis based on the local emotion structure feature data to obtain dynamic emotion driving factor data;

[0007] The module for identifying abnormal trends in internal emotions is used to analyze the local emotional stage evolution of local emotional structure feature data based on dynamic emotional driving factor data to obtain local emotional stage evolution data; analyze the emotional transition states of different emotional stages on the local emotional stage evolution data to obtain emotional transition state data; and identify abnormal trends in internal emotions on the emotional transition state data to obtain abnormal trends in internal emotions;

[0008] The abnormal emotion strategy adjustment module is used to use the long short-term memory network to construct a big data internal emotion analysis model for the internal emotion abnormal trend data, and obtain the internal emotion abnormal analysis model; according to the internal emotion abnormal analysis model, the abnormal internal emotion strategy adjustment is formulated to obtain the abnormal emotion strategy adjustment data.

[0009] By acquiring psychological research data and performing semantic analysis, the present invention can gain an in-depth understanding of the psychological state, emotional experience, and behavioral patterns of individuals or groups. Semantic analysis can help extract key information, including emotional vocabulary, behavioral descriptions, and psychological state, and provide data support for subsequent emotional topic modeling. Emotional topic modeling through psychological semantic data can systematically understand the association and transformation between different emotions, and reveal the emotional patterns and psychological mechanisms hidden behind the data. Emotional topic modeling helps to discover and understand the diversity and complexity of human emotional experience, and provide in-depth insights for psychological research. Dynamic emotion segmentation processing can divide psychological semantic data into different emotion fragments or stages, revealing the temporal changes and dynamic evolution of emotions. Local emotion structure feature analysis can deeply explore the microscopic features of emotion fragments or stages, such as emotion intensity, emotion type, emotion duration, etc., so as to understand the internal mechanism of emotions in more detail. According to dynamic emotion segmentation data and local emotion structure feature data, the drivers that affect emotion changes can be identified and analyzed. Factors include external events, internal psychological processes, and social interactions. The analysis of dynamic emotion driving factors helps to reveal the causes and regulatory mechanisms of emotions, and provides a theoretical basis and practical guidance for emotion management and psychological intervention; the analysis of local emotion stage evolution can identify the changing trends of individual emotions at different stages, including the rise and fall and stability of emotions. Through the analysis of emotion transition states, the transformation relationship between different emotions can be revealed, including sudden changes in emotions and smooth transition states. The identification of abnormal trends in internal emotions can discover abnormal patterns of individual emotion changes, including abnormal frequency, amplitude, and duration, thereby helping to identify possible mental health problems or difficulty in emotion regulation. The internal emotion abnormality analysis model is constructed using the long and short-term memory network, which can learn and model a large amount of internal emotion abnormality trend data, realize intelligent identification and analysis of abnormal emotions, and formulate abnormal emotion strategy adjustment plans based on the internal emotion abnormality analysis model, including personalized emotion management suggestions, emotion regulation skills, etc., to help individuals better understand and deal with their own emotional problems.

[0010] Preferably, the sentiment topic modeling module includes the following functions:

[0011] access to psychological research data;

[0012] Conduct semantic analysis on psychology research data to obtain psychology semantic data; conduct semantic tendency analysis on psychology semantic data to obtain semantic tendency data;

[0013] Emotional topic modeling is performed on the psychological research data based on the psychological semantic data and semantic tendency data to obtain emotional topic modeling data.

[0014] By acquiring psychological research data and performing semantic analysis, the present invention can deeply understand the content and information contained in the data, including emotional expression, behavioral description, and cognitive process. Semantic tendency analysis further deepens the understanding of the data, helps capture the emotional tendencies, attitudes, and emotional colors in the language, and makes the data more informative and emotionally expressive. Emotional topic modeling based on psychological semantic data and semantic tendency data helps to discover potential emotional themes and emotional patterns in the data. Through emotional topic modeling, the types, distribution, and correlations of different emotions in the data can be systematically analyzed, revealing the emotional structure and psychological characteristics behind the data. Emotional topic modeling data can provide in-depth insights for research in the field of psychology, helping researchers understand the emotional experience, mental health status, and social interaction of individuals or groups. Furthermore, these data can also be applied to practical scenarios, such as sentiment analysis, user experience research, and mental health assessment, to provide support for decision-making and practical applications. The application of technologies such as semantic analysis and emotional topic modeling helps to improve the processing efficiency and accuracy of psychological research data, reduce the workload and subjective bias of manual analysis, and provide a more reliable foundation for data-driven decision-making and applications, thereby improving work efficiency and decision quality.

[0015] Preferably, the dynamic emotion driving factor analysis module includes the following functions:

[0016] Perform dynamic emotion segmentation processing on the psychological semantic data according to the emotion topic modeling data to obtain dynamic emotion segmentation data;

[0017] Marking emotion transition points on the dynamic emotion segmentation data to obtain emotion transition point data; performing emotion transition pattern recognition on the emotion transition point data to obtain emotion transition pattern recognition data;

[0018] Perform local emotion structure feature analysis on the dynamic emotion segmentation data based on the emotion transition pattern recognition data to obtain local emotion structure feature data;

[0019] Dynamic emotion driving factor analysis is performed based on local emotion structure feature data and emotion transition pattern recognition data to obtain dynamic emotion driving factor data.

[0020] The dynamic emotion segmentation processing of the present invention can divide psychological semantic data into different emotion fragments, realize fine-grained analysis of emotion changes, and make emotion changes more clearly visible. The marking of emotion transition point data and emotion transition pattern recognition further deepen the understanding of emotion changes, reveal the critical points and patterns of emotion changes, and help to grasp the key moments and laws of emotion changes. Through local emotion structure feature analysis and emotion transition pattern recognition, the local emotion features and transition patterns in dynamic emotion segmentation data can be deeply explored, providing a rich data foundation for subsequent emotion driving factor analysis. Dynamic emotion driving factor analysis can systematically identify factors that affect emotion changes, including external events, internal psychological processes, and social interactions, thereby revealing the causes and regulatory mechanisms of emotion changes; through the analysis of dynamic emotion driving factors, it can help individuals better understand the fluctuations and changes of their own emotions, enhance their ability to manage and regulate emotions, and the results of emotion driving factor analysis can also provide a basis for individuals to formulate personalized emotion management strategies, helping individuals to more effectively cope with various emotional states and challenges.

[0021] Preferably, the performing local emotion structure feature analysis on the dynamic emotion segmentation data includes:

[0022] Performing dynamic emotion time series analysis on the dynamic emotion segmentation data to obtain dynamic emotion time series data; reconstructing the emotion pattern phase space of the dynamic emotion time series data according to the emotion transition pattern recognition data to obtain emotion pattern phase space reconstruction data;

[0023] Performing nonlinear spatial correlation analysis on the emotional pattern phase space reconstruction data to obtain nonlinear spatial correlation data;

[0024] According to the nonlinear spatial correlation data, the emotional pattern phase space reconstruction data is subjected to Poincare cross section analysis to obtain the nonlinear phase space Poincare cross section; the nonlinear phase space Poincare cross section is subjected to spatial trajectory periodicity analysis to obtain spatial trajectory periodicity data;

[0025] Based on the bifurcation theory technology, the trajectory structure data of the spatial trajectory is obtained by performing trajectory structure analysis on the periodicity of the spatial trajectory.

[0026] Local emotion structure feature data is obtained by performing local emotion structure feature analysis based on the spatial trajectory structure data and the spatial trajectory periodicity data.

[0027] Through dynamic emotion time series analysis, the present invention can gain an in-depth understanding of the temporal changes of emotions, including the fluctuation period, duration, and amplitude of emotions, which helps to identify the regularity and trend of emotion changes and provide basic data for subsequent analysis. Emotional pattern phase space reconstruction can map dynamic emotion time series data into high-dimensional phase space, revealing the multidimensional characteristics of emotion changes. Linear spatial correlation analysis can discover the nonlinear relationship between emotions and reveal the interaction and influence between emotions. Poincare cross-section analysis can capture the periodic structure of emotion changes in phase space, helping to understand the stability and periodicity of emotion changes. Spatial trajectory periodicity analysis further analyzes the periodic characteristics of emotion trajectories in phase space, helping to identify the stable period and trend of emotion changes. Trajectory structure analysis based on bifurcation theory technology can reveal bifurcation phenomena and critical points in the process of emotion changes, helping to understand the nonlinear dynamics of emotion changes. The final local emotion structure feature analysis comprehensively considers the temporal characteristics, phase space characteristics, and dynamic characteristics of emotions, extracts the key features of local emotion changes, and provides an in-depth understanding and basis for the analysis of emotion driving factors.

[0028] Preferably, the dynamic emotion driving factor analysis includes:

[0029] Performing pattern-context mapping on the emotion transition pattern recognition data to obtain emotion pattern-context data;

[0030] Performing situational distribution mapping on the emotion pattern situational data according to the local emotion structure feature data to obtain the local emotion structure situational distribution data;

[0031] Performing a structural situational distribution logistic regression analysis on the local emotion structure situational distribution data to obtain situational distribution logistic regression data; performing a logistic linear factor analysis on the local emotion structure situational distribution data based on the situational distribution logistic regression data to obtain situational emotion logical factor data;

[0032] Based on the recurrent neural network, a logical situation perception model is constructed for the situation emotion logic factor data to obtain the logical situation perception model;

[0033] According to the logical situation perception model, the situation emotion logical factor data is predicted with respect to the pre-logical factors to obtain the pre-logical factor data;

[0034] Dynamic emotion driving factor analysis is performed based on the pre-logical factor data and the situational emotion logical factor data to obtain dynamic emotion driving factor data.

[0035] The mapping of emotional pattern contextual data of the present invention will help to understand the manifestations and changes of different emotional patterns in different contexts. The mapping of local emotional structure contextual distribution data can further reveal the distribution and distribution patterns of emotions in various contexts, providing a basis for subsequent analysis. Structural contextual distribution logistic regression analysis and logistic linear factor analysis can identify the correlation and influencing factors between contexts and emotions, which is helpful to understand the changing patterns and driving factors of emotions in different contexts. These analysis results provide quantification and modeling of the relationship between emotions and contexts, and provide data support for subsequent analysis of emotional driving factors. The logical context perception model constructed based on recurrent neural networks can model and predict the complex relationship between contexts and emotions, and improve the understanding and perception of contextual factors. The prediction of pre-logical factor data can help predict the emotional states that may occur in future contexts, and provide early warning and guidance for emotional management and regulation. The final dynamic emotional driving factor data will comprehensively consider the complex relationship between emotions, contexts and other factors, which will help to deeply understand the inherent mechanism and dynamic evolution process of emotional changes.

[0036] Preferably, the module for identifying abnormal inner emotion trends includes the following functions:

[0037] Based on the dynamic emotion driving factor data, the local emotion structure feature data is analyzed for the local emotion stage evolution to obtain the local emotion stage evolution data;

[0038] Evaluate the emotion volatility of the local emotion phase evolution data to obtain emotion volatility evaluation data;

[0039] According to the emotion volatility evaluation data, the emotion transition state of the local emotion stage evolution data at different emotion stages is analyzed to obtain the emotion transition state data;

[0040] The abnormal inner emotion trend is identified on the emotion transition state data to obtain abnormal inner emotion trend data.

[0041] By performing local emotion stage evolution analysis on local emotion structure feature data, the present invention can identify different stages and stage characteristics of emotion changes, which helps to understand the time series pattern of emotion changes and reveal important nodes and trends in the emotion evolution process. The acquisition of emotion volatility assessment data can quantify the degree of fluctuation and frequency of emotion changes, help identify the stability and regularity of emotion fluctuations, which helps to judge the stability of emotion changes and emotion regulation ability, and provides an important reference for subsequent analysis; by analyzing the emotion transition state data, the transition patterns and trends between different emotion stages can be revealed, including states such as sudden changes and smooth transitions of emotions, which helps to discover abnormal emotion transition patterns and warn of existing mental health problems or emotion regulation difficulties. The final internal emotion abnormal trend data will comprehensively consider the results of local emotion stage evolution, emotion volatility assessment and emotion transition state analysis, and identify the emergence of abnormal emotion trends, which helps to detect and intervene in individual mental health problems or emotion management difficulties as early as possible.

[0042] Preferably, performing local emotion stage evolution analysis on local emotion structure feature data includes:

[0043] Based on the dynamic emotion driving factor data, the emotion driving factor emotion relationship adjacency table is constructed for the local emotion structure feature data to obtain the emotion driving relationship adjacency table;

[0044] Perform the same adjacency table structure depth traversal on the emotion-driven relationship adjacency table to obtain the adjacency table structure depth data;

[0045] According to the depth data of the adjacency table structure, the complexity of the relationships between different nodes in the emotion-driven relationship adjacency table is evaluated to obtain the table node relationship complexity data;

[0046] The emotion-driven relationship adjacency table is partitioned according to the table node relationship complexity data to obtain the emotion-driven relationship partitioned adjacency table;

[0047] Performing local implicit numerical calculation on the adjacency table of emotion-driven relationship partitioning to obtain local relationship implicit numerical data; performing adaptive step-size control on the adjacency table of emotion-driven relationship partitioning according to the local relationship implicit numerical data to obtain adjacency table step-size control data;

[0048] According to the local relationship implicit numerical data and the adjacency table step size control data, the local error control Euler-Cromer numerical calculation is performed on the emotion-driven relationship partition adjacency table to obtain the local error control relationship numerical data;

[0049] The local emotion stage evolution data is analyzed on the local emotion structure feature data according to the local error control relationship numerical data to obtain the local emotion stage evolution data.

[0050] By constructing an emotion-driven relationship adjacency table, the present invention can clearly present the driving relationship and influence degree between different emotions, which helps to understand the driving factors of emotion changes and their role in different emotion stages. By performing a structural depth traversal of the adjacency table, the relationship depth and path between different emotion nodes can be discovered. By evaluating the complexity of the relationship between different nodes, the complexity and key nodes of the emotion-driven relationship can be identified. Dividing the emotion-driven relationship adjacency table into different sub-tables helps to analyze the relationship patterns and characteristics between different sub-tables. Through adaptive step size control, the calculation step size in the analysis process can be effectively adjusted to improve calculation efficiency and accuracy. Using the Euler-Cromer numerical calculation method, the local emotion-driven relationship can be accurately simulated and calculated. Through local error control, the calculation error can be effectively controlled to ensure the accuracy and reliability of the analysis results. The final local emotion stage evolution data will comprehensively consider the complexity and dynamic changes of the emotion-driven relationship, revealing the evolution characteristics and laws of emotions at different stages, which helps to understand the process and mechanism of emotion change and provide a basis for further emotion management and intervention.

[0051] Preferably, performing emotional transition state analysis on local emotional phase evolution data includes:

[0052] Drawing a fluctuation curve on the emotion volatility assessment data to obtain an emotion fluctuation curve; extracting an increasing trend and a decreasing trend of the emotion fluctuation curve to obtain an increasing trend data set and a decreasing trend data set respectively;

[0053] Performing increasing rate calculation and decreasing rate calculation on the curve increasing trend data set and the curve decreasing trend data set respectively to obtain curve increasing rate data and curve decreasing rate data respectively;

[0054] Performing local emotion manifold learning mapping processing on the local emotion phase evolution data according to the curve increasing rate data and the curve decreasing rate data to obtain local emotion manifold learning data;

[0055] The local structure complexity of the local emotion manifold learning data is calculated using the manifold learning data point structure analysis algorithm to obtain the local data point complexity data; the local emotion manifold learning data is processed with local complexity constraints based on the local data point complexity data to obtain the local data point complexity constraint data;

[0056] Perform complexity-constrained local linear embedding processing on the local emotion manifold learning data according to the complexity constraint data of the local data points to obtain local constrained linear embedding data;

[0057] According to the local constrained linear embedding data, the local emotion stage evolution data is analyzed for the emotion transition state in different emotion stages to obtain the emotion transition state data.

[0058] The present invention draws the mood fluctuation curve to intuitively display the trend and fluctuation degree of mood changes. Extracting the increasing and decreasing trends of the curve can help identify the trend and law of mood changes and reveal the possibility of mood transitions. Calculating the increasing and decreasing rates helps to quantify the speed of mood changes and further analyze the speed and stability of mood changes, which helps to discover the emergence and trend of mood transition states. Through manifold learning and complexity calculation, mood data can be mapped to low-dimensional space to reveal the structural characteristics and complexity of mood changes, which helps to understand the implicit structure and dynamic characteristics of mood changes. Complexity constraints and local linear embedding can extract the key features of local mood changes and control the complexity of the data, which helps to reduce data dimensions, highlight the important features of mood transitions, and provide a clearer data basis for subsequent analysis. The final mood transition state data will comprehensively consider mood fluctuations, trends, rates and the results of manifold learning, and comprehensively analyze the state and trend of mood changes, which helps to identify the emergence and trend of mood transitions and provide an important reference for individual mood management and psychological intervention.

[0059] Preferably, the local structure complexity of the local emotion manifold learning data is calculated using a manifold learning data point structure analysis algorithm, wherein the manifold learning data point structure analysis algorithm is as follows:

[0060]

[0061] Where C represents the complexity result value of the local data point, α represents the nonlinear correlation coefficient of the local emotion manifold learning data, σ represents the data volume coefficient of the local emotion manifold learning data, x represents the estimated value of the data calculation difficulty of the local emotion manifold learning data, μ represents the dimension coefficient of the local emotion manifold learning data, t represents the maximum calculation time value, b represents the weight coefficient of the neighborhood data points in the local emotion manifold learning data, and R represents the error adjustment value of the manifold learning data point structure analysis algorithm.

[0062] The present invention constructs a manifold learning data point structure analysis algorithm, which provides a method for comprehensively analyzing the complexity of the local sentiment manifold learning data structure by considering factors such as nonlinear relationships, data volume, dimension, time series, neighborhood weights and error adjustment. It can capture the complex characteristics of the data set and help reveal the correlation and implicit patterns between data points. The nonlinear correlation coefficient α of the local sentiment manifold learning data, this parameter represents the strength of the nonlinear relationship between data points. A larger α value will enhance the nonlinear relationship between data points and help capture more complex data structures, which is very beneficial for processing data sets with highly nonlinear characteristics; the data volume coefficient σ of the local sentiment manifold learning data, this parameter represents the data volume size of the data set. A larger σ value will increase the distribution range of the data points and help cover a wider range of data features. When processing large-scale data sets, a larger σ value can provide a more comprehensive data structure analysis; the data calculation difficulty estimate x of the local sentiment manifold learning data, this parameter represents the calculation difficulty estimate of the data point. A larger x value means that the calculation difficulty of the data point is higher and requires more Complex models are used for modeling and analysis. This parameter can help determine the appropriate analysis method for different data points; the dimensionality coefficient μ of the local sentiment manifold learning data, this parameter represents the dimensionality coefficient of the dataset. A larger μ value will increase the dimension of the dataset, that is, the number of features of the data points. For high-dimensional datasets, a larger μ value can better capture the correlation between data features; the maximum calculation time value t, this parameter represents the maximum calculation time value of the algorithm. Increasing the t value will extend the calculation time of the algorithm, enabling it to analyze data within a longer time range, which is very beneficial for processing time series data or situations where long-term observation of data evolution is required; the weight coefficient b of the neighborhood data points in the local sentiment manifold learning data, this parameter represents the weight coefficient of the neighborhood data points. A larger b value will increase the contribution of neighborhood data points to the complexity of local data points, helping to more accurately describe the local relationship between data points; the error adjustment value R of the manifold learning data point structure analysis algorithm, this parameter represents the error adjustment value of the manifold learning data point structure analysis algorithm. By adjusting the R value, the results of the algorithm can be fine-tuned to obtain a better fitting effect.

[0063] Preferably, performing complexity local constraint processing on the local emotion manifold learning data includes:

[0064] Extracting data points from the local emotion manifold learning data according to the local data point complexity data to obtain local emotion manifold learning data points; performing local density calculation on the local emotion manifold learning data points to obtain local density data of the data points;

[0065] Establish the nearest neighbor data point relationship for the local sentiment manifold learning data based on the local density data of the data points to obtain the nearest neighbor data point relationship data;

[0066] Based on the nearest neighbor data point relationship data, the structural relationship normal distribution analysis is performed on the local emotion manifold learning data points to obtain the data point relationship normal distribution data; the dispersion degree of the data point relationship normal distribution data is calculated to obtain the relationship distribution dispersion degree data;

[0067] According to the relationship distribution discrete degree data, the data point relationship normal distribution data and the neighboring data point relationship data, the local sentiment manifold learning data points are subjected to structural connectivity distribution constraints to obtain structural connectivity distribution constraint data;

[0068] According to the structural connectivity distribution constraint data, the local emotion manifold learning data points are processed with local complexity constraints to obtain the local data point complexity constraint data.

[0069] The present invention extracts local emotion manifold learning data points and calculates local density, which can focus on local data points and reduce data dimension and complexity. The calculation of local density helps to determine the density around each data point, thereby highlighting key emotion change points, establishing nearest neighbor data point relationships and performing structural relationship analysis, which can reveal the interactions and connections between data points, which helps to understand the structural characteristics and degree of association between data points. Through relationship normal distribution analysis and discrete degree calculation, the distribution and stability of structural relationships between data points can be evaluated, which helps to identify the stability and reliability of the relationship between data points and provide a basis for subsequent analysis. Complexity constraint processing is performed based on structural connectivity distribution constraints, which can control the degree of connection and complexity between data points, which helps to extract key local emotion change characteristics and reduce data noise and interference, making the analysis results more accurate and reliable.

[0070] The beneficial effects of the present invention are as follows: the present invention provides an intrinsic emotion analysis system based on psychological information big data, which is an optimization process of the traditional intrinsic emotion analysis system based on psychological information big data. It solves the problem that the traditional intrinsic emotion analysis system based on psychological information big data is based on static data for analysis, which makes it difficult to capture the dynamic changes of emotions and cannot accurately analyze the emotions of the content. It can well capture the dynamic changes of emotions and improve the accuracy of the analysis of the emotions of the content. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 This is a module flow diagram of an intrinsic emotion analysis system based on big data of psychological information;

[0072] Figure 2 for Figure 1 Schematic diagram of the functional flow of the dynamic emotion driving factor analysis module;

[0073] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0074] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative work are within the scope of protection of the present invention.

[0075] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0076] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0077] To achieve this, please refer to Figures 1 to 2 , an intrinsic emotion analysis system based on psychology information big data, including the following modules:

[0078] The emotional topic modeling module is used to obtain psychological research data; perform semantic analysis on the psychological research data to obtain psychological semantic data; perform emotional topic modeling on the psychological research data based on the psychological semantic data to obtain emotional topic modeling data;

[0079] The dynamic emotion driving factor analysis module is used to perform dynamic emotion segmentation processing on the psychological semantic data based on the emotion topic modeling data to obtain dynamic emotion segmentation data; perform local emotion structure feature analysis on the dynamic emotion segmentation data to obtain local emotion structure feature data; and perform dynamic emotion driving factor analysis based on the local emotion structure feature data to obtain dynamic emotion driving factor data;

[0080] The module for identifying abnormal trends in internal emotions is used to analyze the local emotional stage evolution of local emotional structure feature data based on dynamic emotional driving factor data to obtain local emotional stage evolution data; analyze the emotional transition states of different emotional stages on the local emotional stage evolution data to obtain emotional transition state data; and identify abnormal trends in internal emotions on the emotional transition state data to obtain abnormal trends in internal emotions;

[0081] The abnormal emotion strategy adjustment module is used to use the long short-term memory network to construct a big data internal emotion analysis model for the internal emotion abnormal trend data, and obtain the internal emotion abnormal analysis model; according to the internal emotion abnormal analysis model, the abnormal internal emotion strategy adjustment is formulated to obtain the abnormal emotion strategy adjustment data.

[0082] In the embodiment of the present invention, reference Figure 1 The above is a module flow diagram of an intrinsic emotion analysis system based on psychology information big data of the present invention. In this example, the intrinsic emotion analysis system based on psychology information big data includes:

[0083] S1: Emotional topic modeling module, used to obtain psychology research data; perform semantic analysis on the psychology research data to obtain psychology semantic data; perform emotional topic modeling on the psychology research data based on the psychology semantic data to obtain emotional topic modeling data;

[0084] In an embodiment of the present invention, data sources in various forms, including various research documents, experimental data, and questionnaires in the field of psychology, are collected, and natural language processing (NLP) technology is used to perform semantic analysis on the psychology research data, and the text data is converted into structured semantic information, including processing steps such as part-of-speech tagging, entity recognition, and syntactic analysis, so as to better understand the meaning and content of the text data, and semantic information related to psychology, such as emotions, psychological states, psychological process keywords and topics, is extracted from the results of the semantic analysis. Based on the psychological semantic data, topic modeling algorithms (such as Latent Dirichlet Allocation, LDA) are used to perform emotion topic modeling on the psychology research data. These models can identify hidden topics and emotional tendencies in text data, decompose text data into a series of emotion-related topics, and convert the results of emotion topic modeling into a data set, in which each document or text data corresponds to the distribution of emotion topics, and each emotion topic is associated with a set of words that describe the emotion or emotional state represented by the topic.

[0085] S2: Dynamic emotion driving factor analysis module, used to perform dynamic emotion segmentation processing on psychological semantic data based on emotion topic modeling data to obtain dynamic emotion segmentation data; perform local emotion structure feature analysis on the dynamic emotion segmentation data to obtain local emotion structure feature data; perform dynamic emotion driving factor analysis based on the local emotion structure feature data to obtain dynamic emotion driving factor data;

[0086] In an embodiment of the present invention, based on the emotion topic modeling data, dynamic emotion segmentation processing is performed on the psychological semantic data, and the time series analysis method or window sliding technology is used to segment the semantic data into different time periods or emotion segments. The dynamic emotion segmentation data is subjected to local emotion structure feature analysis, and the characteristics of each emotion segment are analyzed, including the intensity, duration, transition frequency, etc. of the emotion, as well as the relationship between emotions. Based on the local emotion structure feature data, dynamic emotion driving factor analysis is performed, and statistical analysis, machine learning or deep learning methods are used to identify the key factors and driving mechanisms that affect emotion changes.

[0087] S3: Internal emotion abnormal trend identification module, used to perform local emotion stage evolution analysis on local emotion structure feature data based on dynamic emotion driving factor data to obtain local emotion stage evolution data; perform emotion transition state analysis on local emotion stage evolution data at different emotion stages to obtain emotion transition state data; perform internal emotion abnormal trend identification on emotion transition state data to obtain internal emotion abnormal trend data;

[0088] In an embodiment of the present invention, based on dynamic emotion driving factor data, a local emotion stage evolution analysis is performed on the local emotion structure feature data, and time series analysis or other related methods are used to identify the evolution trend and change law of the local emotion structure in different time periods. The local emotion stage evolution data is subjected to an emotion transition state analysis at different emotion stages, and the transition pattern, frequency and trend of emotions between different stages are analyzed to identify the occurrence and change of emotion transition states. Emotion transition state analysis refers to the transition from one emotion state to another. Based on the emotion transition state data, abnormal trends of internal emotions are identified, and statistical analysis, machine learning or pattern recognition and other methods are used to identify abnormal emotion change patterns and trends.

[0089] S4: Abnormal emotion strategy adjustment module, which is used to use the long short-term memory network to construct a big data intrinsic emotion analysis model for the intrinsic emotion abnormal trend data, and obtain the intrinsic emotion abnormal analysis model; adjust the abnormal intrinsic emotion strategy according to the intrinsic emotion abnormal analysis model, and obtain the abnormal emotion strategy adjustment data.

[0090] In an embodiment of the present invention, a long short-term memory network (LSTM) or other deep learning models suitable for processing time series data are used to construct a big data intrinsic emotion analysis model for intrinsic emotion abnormal trend data. The abnormal emotion trend data is used as input, and the model is trained to learn the changing patterns and trends of intrinsic emotions. Based on the constructed intrinsic emotion abnormality analysis model, an abnormal intrinsic emotion strategy adjustment plan is formulated. According to the output results of the model and the identification of abnormal emotions, targeted emotion regulation strategies are designed to help individuals cope with abnormal emotions. The formulated abnormal intrinsic emotion strategy adjustment plan is converted into a data format to form abnormal emotion strategy adjustment data, including specific methods, time arrangements, and behavioral guidance content for emotion regulation.

[0091] By acquiring psychological research data and performing semantic analysis, the present invention can gain an in-depth understanding of the psychological state, emotional experience, and behavioral patterns of individuals or groups. Semantic analysis can help extract key information, including emotional vocabulary, behavioral descriptions, and psychological state, and provide data support for subsequent emotional topic modeling. Emotional topic modeling through psychological semantic data can systematically understand the association and transformation between different emotions, and reveal the emotional patterns and psychological mechanisms hidden behind the data. Emotional topic modeling helps to discover and understand the diversity and complexity of human emotional experience, and provide in-depth insights for psychological research. Dynamic emotion segmentation processing can divide psychological semantic data into different emotion fragments or stages, revealing the temporal changes and dynamic evolution of emotions. Local emotion structure feature analysis can deeply explore the microscopic features of emotion fragments or stages, such as emotion intensity, emotion type, emotion duration, etc., so as to understand the internal mechanism of emotions in more detail. According to dynamic emotion segmentation data and local emotion structure feature data, the drivers that affect emotion changes can be identified and analyzed. Factors include external events, internal psychological processes, and social interactions. The analysis of dynamic emotion driving factors helps to reveal the causes and regulatory mechanisms of emotions, and provides a theoretical basis and practical guidance for emotion management and psychological intervention; the analysis of local emotion stage evolution can identify the changing trends of individual emotions at different stages, including the rise and fall and stability of emotions. Through the analysis of emotion transition states, the transformation relationship between different emotions can be revealed, including sudden changes in emotions and smooth transition states. The identification of abnormal trends in internal emotions can discover abnormal patterns of individual emotion changes, including abnormal frequency, amplitude, and duration, thereby helping to identify possible mental health problems or difficulty in emotion regulation. The internal emotion abnormality analysis model is constructed using the long and short-term memory network, which can learn and model a large amount of internal emotion abnormality trend data, realize intelligent identification and analysis of abnormal emotions, and formulate abnormal emotion strategy adjustment plans based on the internal emotion abnormality analysis model, including personalized emotion management suggestions, emotion regulation skills, etc., to help individuals better understand and deal with their own emotional problems.

[0092] Preferably, the sentiment topic modeling module includes the following functions:

[0093] access to psychological research data;

[0094] Conduct semantic analysis on psychology research data to obtain psychology semantic data; conduct semantic tendency analysis on psychology semantic data to obtain semantic tendency data;

[0095] Emotional topic modeling is performed on the psychological research data based on the psychological semantic data and semantic tendency data to obtain emotional topic modeling data.

[0096] In an embodiment of the present invention, data sources such as research literature, experimental data, and questionnaires in the field of psychology are collected, and natural language processing (NLP) technology is used to perform semantic analysis on the psychology research data, and the text data is converted into structured semantic information, including processing steps such as part-of-speech tagging, entity recognition, and syntactic analysis, so as to better understand the meaning and content of the text data. Sentiment analysis or tendency analysis is performed on the psychology semantic data obtained through semantic analysis, and sentiment analysis technology is used to identify emotional tendencies in the text, including positive, negative, or neutral emotions. Combined with semantic data and semantic tendency data, emotional topic modeling is performed on the psychology research data. Topic modeling algorithms, such as Latent Dirichlet Allocation (LDA), can be used to decompose the text data into a series of emotion-related topics, and the results of the emotional topic modeling are converted into a data set, wherein each document or text data corresponds to the distribution of emotional topics, and each emotional topic is associated with a set of words that describe the emotion or emotional state represented by the topic.

[0097] By acquiring psychological research data and performing semantic analysis, the present invention can deeply understand the content and information contained in the data, including emotional expression, behavioral description, and cognitive process. Semantic tendency analysis further deepens the understanding of the data, helps capture the emotional tendencies, attitudes, and emotional colors in the language, and makes the data more informative and emotionally expressive. Emotional topic modeling based on psychological semantic data and semantic tendency data helps to discover potential emotional themes and emotional patterns in the data. Through emotional topic modeling, the types, distribution, and correlations of different emotions in the data can be systematically analyzed, revealing the emotional structure and psychological characteristics behind the data. Emotional topic modeling data can provide in-depth insights for research in the field of psychology, helping researchers understand the emotional experience, mental health status, and social interaction of individuals or groups. Furthermore, these data can also be applied to practical scenarios, such as sentiment analysis, user experience research, and mental health assessment, to provide support for decision-making and practical applications. The application of technologies such as semantic analysis and emotional topic modeling helps to improve the processing efficiency and accuracy of psychological research data, reduce the workload and subjective bias of manual analysis, and provide a more reliable foundation for data-driven decision-making and applications, thereby improving work efficiency and decision quality.

[0098] Preferably, the dynamic emotion driving factor analysis module includes the following functions:

[0099] Perform dynamic emotion segmentation processing on the psychological semantic data according to the emotion topic modeling data to obtain dynamic emotion segmentation data;

[0100] Marking emotion transition points on the dynamic emotion segmentation data to obtain emotion transition point data; performing emotion transition pattern recognition on the emotion transition point data to obtain emotion transition pattern recognition data;

[0101] Perform local emotion structure feature analysis on the dynamic emotion segmentation data based on the emotion transition pattern recognition data to obtain local emotion structure feature data;

[0102] Dynamic emotion driving factor analysis is performed based on local emotion structure feature data and emotion transition pattern recognition data to obtain dynamic emotion driving factor data.

[0103] As an example of the present invention, refer to Figure 2 As shown, Figure 1 Schematic diagram of the functional flow of the dynamic emotion driving factor analysis module. In this example, the functions of the dynamic emotion driving factor analysis module include:

[0104] S201: performing dynamic emotion segmentation processing on the psychological semantic data according to the emotion topic modeling data to obtain dynamic emotion segmentation data;

[0105] In an embodiment of the present invention, the emotional topic modeling data obtained by the emotional topic modeling module is used as input, and time series analysis or window sliding technology is used to segment the psychological semantic data into different time periods or emotional fragments, and the emotional transition point, that is, the location of the emotional switch, is determined in the time series to obtain dynamic emotional segmentation data.

[0106] S202: Marking emotion transition points on the dynamic emotion segmentation data to obtain emotion transition point data; performing emotion transition pattern recognition on the emotion transition point data to obtain emotion transition pattern recognition data;

[0107] In an embodiment of the present invention, dynamic emotion segmentation data is traversed, sample by sample is analyzed, and predefined emotion transition rules are used to detect emotion transition points in the samples. The determination of emotion transition points can be based on changes in emotion intensity, conversion of emotion polarity, or emotion fluctuation indicators. For the discovered emotion transition points, the timestamp or sample index of their occurrence is recorded, and machine learning or statistical analysis methods are used to perform pattern recognition on the marked emotion transition point data. The emotion transition point data is used as input features. The goal is to identify different emotion transition patterns. Common emotion transition patterns include gradual changes, sudden changes, and periodic changes. Clustering, classification, or time series analysis techniques can be used for pattern recognition to discover potential emotion transition patterns in the data.

[0108] S203: performing local emotion structure feature analysis on the dynamic emotion segmentation data according to the emotion transition pattern recognition data to obtain local emotion structure feature data;

[0109] In an embodiment of the present invention, emotion transition pattern recognition data is obtained, including different identified emotion transition patterns and their corresponding time periods or sample indexes, dynamic emotion segmentation data is prepared, ensuring that each time period or sample has a corresponding emotion tag or emotional value, and the dynamic emotion segmentation data is divided into different local emotion structures according to the emotion transition pattern recognition data. For each local emotion structure, emotion feature analysis is performed, including the intensity, polarity, and duration of the emotion. The average emotion intensity, emotion change amplitude, and emotion stability of each local emotion structure can be calculated, and the emotion features obtained for each local emotion structure are integrated to form local emotion structure feature data. Each local emotion structure feature data includes the emotion feature information within the time period or sample.

[0110] S204: Perform dynamic emotion driving factor analysis based on the local emotion structure feature data and the emotion transition pattern recognition data to obtain dynamic emotion driving factor data.

[0111] In an embodiment of the present invention, local emotion structure feature data and emotion transition pattern recognition data are obtained to ensure the consistency and integrity of the data. According to the emotion transition pattern recognition data, the dynamic emotion segmentation data is divided into different local emotion structures. For each local emotion structure, the emotion driving factors therein are analyzed using the local emotion structure feature data. Correlation analysis, factor analysis, machine learning and other methods can be used to identify the driving factors affecting the local emotion structure. The identified dynamic emotion driving factors are integrated into a data set. Each sample represents a local emotion structure and its corresponding driving factor. Each driving factor can be a factor that affects emotion changes, such as a situation, an external event, and an individual characteristic. The dynamic emotion driving factor data is statistically analyzed to explore the driving factor characteristics under different emotion transition modes. The analysis results may include information such as the main driving factors, the weights or the degree of influence of the driving factors under different emotion transition modes, and dynamic emotion driving factor data is obtained.

[0112] The dynamic emotion segmentation processing of the present invention can divide psychological semantic data into different emotion fragments, realize fine-grained analysis of emotion changes, and make emotion changes more clearly visible. The marking of emotion transition point data and emotion transition pattern recognition further deepen the understanding of emotion changes, reveal the critical points and patterns of emotion changes, and help to grasp the key moments and laws of emotion changes. Through local emotion structure feature analysis and emotion transition pattern recognition, the local emotion features and transition patterns in dynamic emotion segmentation data can be deeply explored, providing a rich data foundation for subsequent emotion driving factor analysis. Dynamic emotion driving factor analysis can systematically identify factors that affect emotion changes, including external events, internal psychological processes, and social interactions, thereby revealing the causes and regulatory mechanisms of emotion changes; through the analysis of dynamic emotion driving factors, it can help individuals better understand the fluctuations and changes of their own emotions, enhance their ability to manage and regulate emotions, and the results of emotion driving factor analysis can also provide a basis for individuals to formulate personalized emotion management strategies, helping individuals to more effectively cope with various emotional states and challenges.

[0113] Preferably, the performing local emotion structure feature analysis on the dynamic emotion segmentation data includes:

[0114] Performing dynamic emotion time series analysis on the dynamic emotion segmentation data to obtain dynamic emotion time series data; reconstructing the emotion pattern phase space of the dynamic emotion time series data according to the emotion transition pattern recognition data to obtain emotion pattern phase space reconstruction data;

[0115] Performing nonlinear spatial correlation analysis on the emotional pattern phase space reconstruction data to obtain nonlinear spatial correlation data;

[0116] According to the nonlinear spatial correlation data, the emotional pattern phase space reconstruction data is subjected to Poincare cross section analysis to obtain the nonlinear phase space Poincare cross section; the nonlinear phase space Poincare cross section is subjected to spatial trajectory periodicity analysis to obtain spatial trajectory periodicity data;

[0117] Based on the bifurcation theory technology, the trajectory structure data of the spatial trajectory is obtained by performing trajectory structure analysis on the periodicity of the spatial trajectory.

[0118] Local emotion structure feature data is obtained by performing local emotion structure feature analysis based on the spatial trajectory structure data and the spatial trajectory periodicity data.

[0119] In an embodiment of the present invention, the data is segmented dynamically by emotions to determine time series data points, and time series analysis techniques such as sliding windows and smoothing are applied to capture the changing trend of emotions over time. Statistical methods or machine learning techniques such as ARIMA models and LSTM neural networks are used to model and predict time series data in order to understand the dynamic changes of emotions. The emotion time series data is converted into emotion pattern phase space data, and an appropriate phase space reconstruction technique such as delayed coordinate method or time delay embedding is used to retain the dynamic features in the data, and appropriate delay time and embedding dimension are determined to ensure that the reconstructed phase space can accurately reflect the dynamic characteristics of the original data. Nonlinear spatial correlation analysis methods such as cross-correlation analysis of phase space reconstruction or maximum Lyapunov exponent calculation are used to measure the degree of mutual correlation between emotion patterns, and appropriate correlation measurement indicators are determined to evaluate the correlation between different emotion patterns, and possible nonlinear correlation structures are found to reconstruct the emotion pattern phase space. Poincare section analysis is performed on the reconstructed data to capture the important geometric features of the phase space trajectory. By drawing the Poincare section in the phase space, the intersection points and intersection structures of the phase space trajectory can be visualized, which helps to understand the dynamic interactive relationship between emotional patterns. Periodic analysis is performed on the Poincare section data to detect possible periodic changes between emotional patterns. Spectral analysis or periodicity detection algorithms, such as Fourier transform or autocorrelation function, are applied to identify periodic components in the spatial trajectory. Based on the bifurcation theory technology, the spatial trajectory periodic data is analyzed to reveal the bifurcation structure between emotional patterns. By analyzing the bifurcation points and bifurcation paths of the trajectories, we can gain an in-depth understanding of the transition patterns and evolution trends between emotional patterns. Based on the spatial trajectory structure data and the spatial trajectory periodicity data, the local emotional structure characteristics are analyzed; the local features between emotional patterns, such as local maximum points, local minimum points, etc., and the correlation between these feature points are determined.

[0120] Through dynamic emotion time series analysis, the present invention can gain an in-depth understanding of the temporal changes of emotions, including the fluctuation period, duration, and amplitude of emotions, which helps to identify the regularity and trend of emotion changes and provide basic data for subsequent analysis. Emotional pattern phase space reconstruction can map dynamic emotion time series data into high-dimensional phase space, revealing the multidimensional characteristics of emotion changes. Linear spatial correlation analysis can discover the nonlinear relationship between emotions and reveal the interaction and influence between emotions. Poincare cross-section analysis can capture the periodic structure of emotion changes in phase space, helping to understand the stability and periodicity of emotion changes. Spatial trajectory periodicity analysis further analyzes the periodic characteristics of emotion trajectories in phase space, helping to identify the stable period and trend of emotion changes. Trajectory structure analysis based on bifurcation theory technology can reveal bifurcation phenomena and critical points in the process of emotion changes, helping to understand the nonlinear dynamics of emotion changes. The final local emotion structure feature analysis comprehensively considers the temporal characteristics, phase space characteristics, and dynamic characteristics of emotions, extracts the key features of local emotion changes, and provides an in-depth understanding and basis for the analysis of emotion driving factors.

[0121] Preferably, the dynamic emotion driving factor analysis includes:

[0122] Performing pattern-context mapping on the emotion transition pattern recognition data to obtain emotion pattern-context data;

[0123] Performing situational distribution mapping on the emotion pattern situational data according to the local emotion structure feature data to obtain the local emotion structure situational distribution data;

[0124] Performing a structural situational distribution logistic regression analysis on the local emotion structure situational distribution data to obtain situational distribution logistic regression data; performing a logistic linear factor analysis on the local emotion structure situational distribution data based on the situational distribution logistic regression data to obtain situational emotion logical factor data;

[0125] Based on the recurrent neural network, a logical situation perception model is constructed for the situation emotion logic factor data to obtain the logical situation perception model;

[0126] According to the logical situation perception model, the situation emotion logical factor data is predicted with respect to the pre-logical factors to obtain the pre-logical factor data;

[0127] Dynamic emotion driving factor analysis is performed based on the pre-logical factor data and the situational emotion logical factor data to obtain dynamic emotion driving factor data.

[0128] In the embodiment of the present invention, the emotion transition pattern recognition data is used to map the emotion pattern to a specific situation or background to form emotion pattern situation data. The situation types or situation characteristics corresponding to different emotion patterns are identified through data mining or pattern recognition technology. Based on the local emotion structure feature data, the situation distribution is mapped to the emotion pattern situation data. The correlation between the local emotion structure feature and the situation distribution is analyzed. The distribution characteristics of the local emotion structure in different situations are identified. Logistic regression analysis is performed on the local emotion structure situation distribution data. The logistic regression model is used to explore the distribution law of the local emotion structure in different situations. The situation factors that affect the local emotion structure are identified. According to the situation distribution logistic regression data, the local emotion structure situation distribution data is subjected to logical linear factor analysis. The emotion distribution is determined through factor analysis. Situational emotional logical factors are used to reveal the potential correlation and causal relationship between situations and emotional patterns. Based on recurrent neural networks or other appropriate models, a logical situational awareness model is constructed. Using situational emotional logical factor data, the model is trained to learn the complex relationship between situations and emotions, and a situational awareness model is established. According to the logical situational awareness model, the situational emotional logical factor data is predicted with pre-logical factors. The trained model is used to predict the emotional logical factors in future or unknown situations, thereby revealing the impact of situational changes on emotions. Combined with pre-logical factor data and situational emotional logical factor data, dynamic emotion driving factor analysis is conducted to analyze the driving effect of situational changes on emotional patterns, identify the emotional change patterns and driving factors in different situations, and thus deeply understand the dynamic characteristics of emotions and the causal relationship behind them.

[0129] The mapping of emotional pattern contextual data of the present invention will help to understand the manifestations and changes of different emotional patterns in different contexts. The mapping of local emotional structure contextual distribution data can further reveal the distribution and distribution patterns of emotions in various contexts, providing a basis for subsequent analysis. Structural contextual distribution logistic regression analysis and logistic linear factor analysis can identify the correlation and influencing factors between contexts and emotions, which is helpful to understand the changing patterns and driving factors of emotions in different contexts. These analysis results provide quantification and modeling of the relationship between emotions and contexts, and provide data support for subsequent analysis of emotional driving factors. The logical context perception model constructed based on recurrent neural networks can model and predict the complex relationship between contexts and emotions, and improve the understanding and perception of contextual factors. The prediction of pre-logical factor data can help predict the emotional states that may occur in future contexts, and provide early warning and guidance for emotional management and regulation. The final dynamic emotional driving factor data will comprehensively consider the complex relationship between emotions, contexts and other factors, which will help to deeply understand the inherent mechanism and dynamic evolution process of emotional changes.

[0130] Preferably, the module for identifying abnormal inner emotion trends includes the following functions:

[0131] Based on the dynamic emotion driving factor data, the local emotion structure feature data is analyzed for the local emotion stage evolution to obtain the local emotion stage evolution data;

[0132] Evaluate the emotion volatility of the local emotion phase evolution data to obtain emotion volatility evaluation data;

[0133] According to the emotion volatility evaluation data, the emotion transition state of the local emotion stage evolution data at different emotion stages is analyzed to obtain the emotion transition state data;

[0134] The abnormal inner emotion trend is identified on the emotion transition state data to obtain abnormal inner emotion trend data.

[0135] In an embodiment of the present invention, dynamic emotion driving factor data is used to analyze local emotion structure feature data, and by identifying and analyzing the changing trend of the local emotion structure, the emotion state in different time periods is determined. Based on the dynamic emotion driving factor data, a stage evolution analysis is performed on the local emotion structure feature data, the emotion structure feature data is grouped by time period, and the evolution process of the emotion structure in each time period is analyzed to obtain local emotion stage evolution data. An emotion volatility assessment is performed on the local emotion stage evolution data, and a statistical method or indicator is used to assess the degree of emotion fluctuation in different stages to quantify the amplitude and frequency of emotion changes and obtain emotion volatility assessment data. Based on the emotion volatility assessment data, an emotion transition state analysis is performed on the local emotion stage evolution data, and the transition process and state between different emotion stages are analyzed to identify the transition of emotion from one state to another and obtain emotion transition state data. Based on the emotion volatility assessment data and the emotion transition state data, abnormal trends of internal emotions are identified, and abnormal emotion fluctuation patterns or transition states are identified using machine learning or statistical analysis methods, thereby determining abnormal trends of internal emotions and obtaining abnormal trend data of internal emotions.

[0136] By performing local emotion stage evolution analysis on local emotion structure feature data, the present invention can identify different stages and stage characteristics of emotion changes, which helps to understand the time series pattern of emotion changes and reveal important nodes and trends in the emotion evolution process. The acquisition of emotion volatility assessment data can quantify the degree of fluctuation and frequency of emotion changes, help identify the stability and regularity of emotion fluctuations, which helps to judge the stability of emotion changes and emotion regulation ability, and provides an important reference for subsequent analysis; by analyzing the emotion transition state data, the transition patterns and trends between different emotion stages can be revealed, including states such as sudden changes and smooth transitions of emotions, which helps to discover abnormal emotion transition patterns and warn of existing mental health problems or emotion regulation difficulties. The final internal emotion abnormal trend data will comprehensively consider the results of local emotion stage evolution, emotion volatility assessment and emotion transition state analysis, and identify the emergence of abnormal emotion trends, which helps to detect and intervene in individual mental health problems or emotion management difficulties as early as possible.

[0137] Preferably, performing local emotion stage evolution analysis on local emotion structure feature data includes:

[0138] Based on the dynamic emotion driving factor data, the emotion driving factor emotion relationship adjacency table is constructed for the local emotion structure feature data to obtain the emotion driving relationship adjacency table;

[0139] Perform the same adjacency table structure depth traversal on the emotion-driven relationship adjacency table to obtain the adjacency table structure depth data;

[0140] According to the depth data of the adjacency table structure, the complexity of the relationships between different nodes in the emotion-driven relationship adjacency table is evaluated to obtain the table node relationship complexity data;

[0141] The emotion-driven relationship adjacency table is partitioned according to the table node relationship complexity data to obtain the emotion-driven relationship partitioned adjacency table;

[0142] Performing local implicit numerical calculation on the adjacency table of emotion-driven relationship partitioning to obtain local relationship implicit numerical data; performing adaptive step-size control on the adjacency table of emotion-driven relationship partitioning according to the local relationship implicit numerical data to obtain adjacency table step-size control data;

[0143] According to the local relationship implicit numerical data and the adjacency table step size control data, the local error control Euler-Cromer numerical calculation is performed on the emotion-driven relationship partition adjacency table to obtain the local error control relationship numerical data;

[0144] The local emotion stage evolution data is analyzed on the local emotion structure feature data according to the local error control relationship numerical data to obtain the local emotion stage evolution data.

[0145] In the embodiment of the present invention, based on the dynamic emotion driving factor data, an emotion driving relationship adjacency table is constructed, and the relationship between the emotion driving factors is expressed in the form of an adjacency table for subsequent analysis. The constructed emotion driving relationship adjacency table is subjected to depth-first traversal to obtain the adjacency table structure depth data. The depth traversal can help understand the relationship depth and structure between the emotion factors. According to the adjacency table structure depth data, the relationship complexity between different nodes is evaluated. The degree, clustering coefficient and other indicators in graph theory can be used to evaluate the relationship complexity between nodes. According to the complexity data of the relationship between nodes, the emotion driving relationship adjacency table is divided. The division can be performed according to the complexity threshold or other standards. For subsequent analysis and calculation, the divided emotion driving relationship adjacency table is subjected to local implicit numerical calculation, and the numerical calculation method can be used to derive And calculate the implicit relationship between emotional factors to better understand the evolution process of emotions. According to the implicit numerical data of local relationships, the adjacency list of emotion-driven relationship partition is adaptively controlled by step size control. The step size control can adjust the step size of the calculation according to the calculation results to improve the calculation efficiency and accuracy. Combined with the implicit numerical data of local relationships and the step size control data, the adjacency list of emotion-driven relationship partition is subjected to local error control Euler-Cromer numerical calculation. Through the numerical calculation method, the emotion evolution process is simulated and analyzed to obtain the local error control relationship numerical data. Based on the local error control relationship numerical data, the local emotion structure feature data is analyzed for the local emotion stage evolution. By analyzing the relationship changes between local emotion factors, the emotion evolution at different stages is obtained, thereby obtaining the local emotion stage evolution data.

[0146] By constructing an emotion-driven relationship adjacency table, the present invention can clearly present the driving relationship and influence degree between different emotions, which helps to understand the driving factors of emotion changes and their role in different emotion stages. By performing a structural depth traversal of the adjacency table, the relationship depth and path between different emotion nodes can be discovered. By evaluating the complexity of the relationship between different nodes, the complexity and key nodes of the emotion-driven relationship can be identified. Dividing the emotion-driven relationship adjacency table into different sub-tables helps to analyze the relationship patterns and characteristics between different sub-tables. Through adaptive step size control, the calculation step size in the analysis process can be effectively adjusted to improve calculation efficiency and accuracy. Using the Euler-Cromer numerical calculation method, the local emotion-driven relationship can be accurately simulated and calculated. Through local error control, the calculation error can be effectively controlled to ensure the accuracy and reliability of the analysis results. The final local emotion stage evolution data will comprehensively consider the complexity and dynamic changes of the emotion-driven relationship, revealing the evolution characteristics and laws of emotions at different stages, which helps to understand the process and mechanism of emotion change and provide a basis for further emotion management and intervention.

[0147] Preferably, performing emotional transition state analysis on local emotional phase evolution data includes:

[0148] Drawing a fluctuation curve on the emotion volatility assessment data to obtain an emotion fluctuation curve; extracting an increasing trend and a decreasing trend of the emotion fluctuation curve to obtain an increasing trend data set and a decreasing trend data set respectively;

[0149] Performing increasing rate calculation and decreasing rate calculation on the curve increasing trend data set and the curve decreasing trend data set respectively to obtain curve increasing rate data and curve decreasing rate data respectively;

[0150] Performing local emotion manifold learning mapping processing on the local emotion phase evolution data according to the curve increasing rate data and the curve decreasing rate data to obtain local emotion manifold learning data;

[0151] The local structure complexity of the local emotion manifold learning data is calculated using the manifold learning data point structure analysis algorithm to obtain the local data point complexity data; the local emotion manifold learning data is processed with local complexity constraints based on the local data point complexity data to obtain the local data point complexity constraint data;

[0152] Perform complexity-constrained local linear embedding processing on the local emotion manifold learning data according to the complexity constraint data of the local data points to obtain local constrained linear embedding data;

[0153] According to the local constrained linear embedding data, the local emotion stage evolution data is analyzed for the emotion transition state in different emotion stages to obtain the emotion transition state data.

[0154] In an embodiment of the present invention, the emotion volatility assessment data is plotted into a fluctuation curve graph in chronological order. Time can be divided into hours, days or other appropriate units. An appropriate time interval is selected according to the time span of the data. By observing the fluctuation curve, the increasing and decreasing segments of the curve are identified. The increasing segment indicates that the emotion is in an enhanced state, while the decreasing segment indicates that the emotion tends to be stable or weakened. The increasing segment data and the decreasing segment data on the curve are extracted to form a curve increasing trend data set and a curve decreasing trend data set, respectively. The rate of the curve increasing trend data set and the curve decreasing trend data set are calculated respectively. The speed of change of the curve is calculated by linear regression or difference method. The curve increasing rate data and the curve decreasing rate data are used to perform local emotion manifold learning mapping processing. Manifold learning algorithms, such as Locally Linear Embedding (LLE) or t-distributed stochastic neighbor embedding (t-SNE), can be used to map high-dimensional emotion data into low-dimensional space; manifold learning data point structure analysis algorithms, such as Locally Linear Embedding (LLE) or t-distributed stochastic neighbor embedding (t-SNE) are used. embedding (t-SNE), which calculates the structural complexity of each data point in the local sentiment manifold learning data. This can include local density-based methods, such as Local Reachability Density (LRD) or Local Outlier Factor (LOF). Based on the local data point complexity data, the local emotion manifold learning data is subjected to complexity constraint processing. Data points with higher or lower complexity can be retained, or other constraint strategies can be formulated according to specific application requirements. The local data point complexity constraint data is used to perform complexity constrained local linear embedding processing on the local emotion manifold learning data. The local linear embedding algorithm can map high-dimensional data to low-dimensional space by maintaining the local linear relationship between data points, while considering complexity constraints to retain the structural characteristics of the data points. Based on the local constrained linear embedding data, the local emotion stage evolution data is subjected to emotion transition state analysis at different emotion stages. Clustering algorithms, such as k-means clustering or DBSCAN, can be used to perform cluster analysis on the local constrained linear embedding data to identify data point clusters at different emotion stages. For further analysis of the emotion transition state, time series analysis methods, such as state transition models or event-driven models, can be considered to capture the transition and change patterns of emotions between different stages.

[0155] The present invention draws the mood fluctuation curve to intuitively display the trend and fluctuation degree of mood changes. Extracting the increasing and decreasing trends of the curve can help identify the trend and law of mood changes and reveal the possibility of mood transitions. Calculating the increasing and decreasing rates helps to quantify the speed of mood changes and further analyze the speed and stability of mood changes, which helps to discover the emergence and trend of mood transition states. Through manifold learning and complexity calculation, mood data can be mapped to low-dimensional space to reveal the structural characteristics and complexity of mood changes, which helps to understand the implicit structure and dynamic characteristics of mood changes. Complexity constraints and local linear embedding can extract the key features of local mood changes and control the complexity of the data, which helps to reduce data dimensions, highlight the important features of mood transitions, and provide a clearer data basis for subsequent analysis. The final mood transition state data will comprehensively consider mood fluctuations, trends, rates and the results of manifold learning, and comprehensively analyze the state and trend of mood changes, which helps to identify the emergence and trend of mood transitions and provide an important reference for individual mood management and psychological intervention.

[0156] Preferably, the local structure complexity of the local emotion manifold learning data is calculated using a manifold learning data point structure analysis algorithm, wherein the manifold learning data point structure analysis algorithm is as follows:

[0157]

[0158] Where C represents the complexity result value of the local data point, α represents the nonlinear correlation coefficient of the local emotion manifold learning data, σ represents the data volume coefficient of the local emotion manifold learning data, x represents the estimated value of the data calculation difficulty of the local emotion manifold learning data, μ represents the dimension coefficient of the local emotion manifold learning data, t represents the maximum calculation time value, b represents the weight coefficient of the neighborhood data points in the local emotion manifold learning data, and R represents the error adjustment value of the manifold learning data point structure analysis algorithm.

[0159] The present invention constructs a manifold learning data point structure analysis algorithm, which provides a method for comprehensively analyzing the complexity of the local sentiment manifold learning data structure by considering factors such as nonlinear relationships, data volume, dimension, time series, neighborhood weights and error adjustment. It can capture the complex characteristics of the data set and help reveal the correlation and implicit patterns between data points. The nonlinear correlation coefficient α of the local sentiment manifold learning data, this parameter represents the strength of the nonlinear relationship between data points. A larger α value will enhance the nonlinear relationship between data points and help capture more complex data structures, which is very beneficial for processing data sets with highly nonlinear characteristics; the data volume coefficient σ of the local sentiment manifold learning data, this parameter represents the data volume size of the data set. A larger σ value will increase the distribution range of the data points and help cover a wider range of data features. When processing large-scale data sets, a larger σ value can provide a more comprehensive data structure analysis; the data calculation difficulty estimate x of the local sentiment manifold learning data, this parameter represents the calculation difficulty estimate of the data point. A larger x value means that the calculation difficulty of the data point is higher and requires more Complex models are used for modeling and analysis. This parameter can help determine the appropriate analysis method for different data points; the dimensionality coefficient μ of the local sentiment manifold learning data, this parameter represents the dimensionality coefficient of the dataset. A larger μ value will increase the dimension of the dataset, that is, the number of features of the data points. For high-dimensional datasets, a larger μ value can better capture the correlation between data features; the maximum calculation time value t, this parameter represents the maximum calculation time value of the algorithm. Increasing the t value will extend the calculation time of the algorithm, enabling it to analyze data within a longer time range, which is very beneficial for processing time series data or situations where long-term observation of data evolution is required; the weight coefficient b of the neighborhood data points in the local sentiment manifold learning data, this parameter represents the weight coefficient of the neighborhood data points. A larger b value will increase the contribution of neighborhood data points to the complexity of local data points, helping to more accurately describe the local relationship between data points; the error adjustment value R of the manifold learning data point structure analysis algorithm, this parameter represents the error adjustment value of the manifold learning data point structure analysis algorithm. By adjusting the R value, the results of the algorithm can be fine-tuned to obtain a better fitting effect.

[0160] Preferably, performing complexity local constraint processing on the local emotion manifold learning data includes:

[0161] Extracting data points from the local emotion manifold learning data according to the local data point complexity data to obtain local emotion manifold learning data points; performing local density calculation on the local emotion manifold learning data points to obtain local density data of the data points;

[0162] Establish the nearest neighbor data point relationship for the local sentiment manifold learning data based on the local density data of the data points to obtain the nearest neighbor data point relationship data;

[0163] Based on the nearest neighbor data point relationship data, the structural relationship normal distribution analysis is performed on the local emotion manifold learning data points to obtain the data point relationship normal distribution data; the dispersion degree of the data point relationship normal distribution data is calculated to obtain the relationship distribution dispersion degree data;

[0164] According to the relationship distribution discrete degree data, the data point relationship normal distribution data and the neighboring data point relationship data, the local sentiment manifold learning data points are subjected to structural connectivity distribution constraints to obtain structural connectivity distribution constraint data;

[0165] According to the structural connectivity distribution constraint data, the local emotion manifold learning data points are processed with local complexity constraints to obtain the local data point complexity constraint data.

[0166] In the embodiment of the present invention, the local data point complexity data is used to extract data points from the local emotion manifold learning data to obtain local emotion manifold learning data points, and the local density of the extracted local emotion manifold learning data points is calculated. The local density data of each data point is calculated using density estimation methods, such as the K-nearest neighbor method or kernel density estimation, and the nearest neighbor data point relationship is established based on the local density data of the data points. For each data point, its nearest K neighbors are determined, and these neighbors will become part of its local relationship. The nearest neighbor data point relationship data is used to perform structural relationship normal distribution analysis on the local emotion manifold learning data points. Statistical methods, such as normal distribution fitting, can be used to capture the data point relationships. The normal property of the system is used to calculate the degree of discreteness of the normal distribution data of the data point relationship. The standard deviation of the normal distribution of the data point relationship or other discreteness indicators can be calculated to quantify the distribution shape of the data point relationship. Combined with the discrete degree data of the relationship distribution, the normal distribution data of the data point relationship and the neighboring data point relationship data, the structural connectivity distribution constraints are performed on the local sentiment manifold learning data points. Constraints can be formulated, such as keeping the data point relationship distribution within a certain normal range or adjusting the connection density between data points. According to the structural connectivity distribution constraint data, the local complexity constraint processing is performed on the local sentiment manifold learning data points. According to the constraint conditions, the position or connection relationship of the data points is adjusted to meet the local complexity constraints.

[0167] The present invention extracts local emotion manifold learning data points and calculates local density, which can focus on local data points and reduce data dimension and complexity. The calculation of local density helps to determine the density around each data point, thereby highlighting key emotion change points, establishing nearest neighbor data point relationships and performing structural relationship analysis, which can reveal the interactions and connections between data points, which helps to understand the structural characteristics and degree of association between data points. Through relationship normal distribution analysis and discrete degree calculation, the distribution and stability of structural relationships between data points can be evaluated, which helps to identify the stability and reliability of the relationship between data points and provide a basis for subsequent analysis. Complexity constraint processing is performed based on structural connectivity distribution constraints, which can control the degree of connection and complexity between data points, which helps to extract key local emotion change characteristics and reduce data noise and interference, making the analysis results more accurate and reliable.

[0168] The beneficial effects of the present invention are as follows: the present invention provides an intrinsic emotion analysis system based on psychological information big data, which is an optimization process of the traditional intrinsic emotion analysis system based on psychological information big data. It solves the problem that the traditional intrinsic emotion analysis system based on psychological information big data is based on static data for analysis, which makes it difficult to capture the dynamic changes of emotions and cannot accurately analyze the emotions of the content. It can well capture the dynamic changes of emotions and improve the accuracy of the analysis of the emotions of the content.

[0169] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0170] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. An intrinsic emotion analysis system based on psychology information big data, characterized by: Includes the following modules: The emotional topic modeling module is used to obtain psychological research data, which includes various research documents, experimental data, and questionnaires in the field of psychology. It performs semantic analysis on the psychological research data to extract semantic information related to psychology, and obtains psychological semantic data including emotion, psychological state, psychological process keywords and themes. It performs emotional topic modeling on the psychological research data based on the psychological semantic data. The model can identify the hidden themes and emotional tendencies in the text data, decompose the text data into a series of emotion-related topics, and convert the results of the emotional topic modeling into a dataset, in which each document or text data corresponds to the distribution of emotional topics. Each emotional topic is associated with a set of words that describe the emotion or emotional state represented by the emotional topic, thus obtaining emotional topic modeling data. The dynamic emotion driving factor analysis module is used to perform dynamic emotion segmentation processing on psychological semantic data based on emotion topic modeling data, dividing the psychological semantic data into different emotion segments or stages, revealing the temporal changes and dynamic evolution process of emotions, and obtaining dynamic emotion segmentation data; performing local emotion structure feature analysis on the dynamic emotion segmentation data to obtain local emotion structure feature data. Local emotion structure feature analysis can deeply explore the micro-features of emotion segments or stages, including emotion intensity, emotion type, and emotion duration, and understand the internal mechanism of emotions; performing dynamic emotion driving factor analysis based on local emotion structure feature data to obtain dynamic emotion driving factor data. Based on the dynamic emotion segmentation data and local emotion structure feature data, the driving factors that affect emotion changes can be identified and analyzed, including external events, internal psychological processes, and social interactions. Dynamic emotion driving factor analysis helps to reveal the causes and regulatory mechanisms of emotions; The module for identifying abnormal trends in internal emotions is used to perform local emotion stage evolution analysis on local emotion structure feature data based on dynamic emotion driving factor data. The local emotion stage evolution analysis can identify the changing trends of individual emotions at different stages, including the rise and fall and stability of emotions, and obtain local emotion stage evolution data. The module also performs emotion transition state analysis at different emotion stages on the local emotion stage evolution data to obtain emotion transition state data. Through the emotion transition state analysis, the transition relationship between different emotions can be revealed, including sudden changes in emotions and smooth transition states. The module also performs abnormal trend identification on the emotion transition state data. The abnormal trend identification can discover abnormal patterns of individual emotion changes, including abnormal frequency, amplitude, and duration, and obtain abnormal trend data on internal emotions. The abnormal emotion strategy adjustment module is used to construct a big data internal emotion analysis model based on the internal emotion abnormal trend data using the long short-term memory network to obtain the internal emotion abnormal analysis model; adjust the abnormal internal emotion strategy based on the internal emotion abnormal analysis model to obtain the abnormal emotion strategy adjustment data; Emotional transition analysis of local emotion phase evolution data includes: Drawing a fluctuation curve on the emotion volatility assessment data to obtain an emotion fluctuation curve; extracting an increasing trend and a decreasing trend of the emotion fluctuation curve to obtain an increasing trend data set and a decreasing trend data set respectively; Performing increasing rate calculation and decreasing rate calculation on the curve increasing trend data set and the curve decreasing trend data set respectively to obtain curve increasing rate data and curve decreasing rate data respectively; Performing local emotion manifold learning mapping processing on the local emotion phase evolution data according to the curve increasing rate data and the curve decreasing rate data to obtain local emotion manifold learning data; The local structure complexity of the local emotion manifold learning data is calculated using the manifold learning data point structure analysis algorithm to obtain the local data point complexity data; the local emotion manifold learning data is processed with local complexity constraints based on the local data point complexity data to obtain the local data point complexity constraint data; Perform complexity-constrained local linear embedding processing on the local emotion manifold learning data according to the complexity constraint data of the local data points to obtain local constrained linear embedding data; According to the local constrained linear embedding data, the local emotion stage evolution data is analyzed for the emotion transition state in different emotion stages to obtain the emotion transition state data.

2. The intrinsic emotion analysis system based on psychology information big data according to claim 1 is characterized in that: The sentiment topic modeling module includes the following functions: access to psychological research data; Conduct semantic analysis on psychology research data to obtain psychology semantic data; conduct semantic tendency analysis on psychology semantic data to obtain semantic tendency data; Emotional topic modeling is performed on the psychological research data based on the psychological semantic data and semantic tendency data to obtain emotional topic modeling data.

3. The intrinsic emotion analysis system based on psychology information big data according to claim 1 is characterized in that: The Dynamic Sentiment Drivers Analysis module includes the following features: Perform dynamic emotion segmentation processing on the psychological semantic data according to the emotion topic modeling data to obtain dynamic emotion segmentation data; Marking emotion transition points on the dynamic emotion segmentation data to obtain emotion transition point data; performing emotion transition pattern recognition on the emotion transition point data to obtain emotion transition pattern recognition data; Perform local emotion structure feature analysis on the dynamic emotion segmentation data based on the emotion transition pattern recognition data to obtain local emotion structure feature data; Dynamic emotion driving factor analysis is performed based on local emotion structure feature data and emotion transition pattern recognition data to obtain dynamic emotion driving factor data.

4. The intrinsic emotion analysis system based on psychology information big data according to claim 3 is characterized in that: The analysis of local emotion structure characteristics of dynamic emotion segmentation data includes: Performing dynamic emotion time series analysis on the dynamic emotion segmentation data to obtain dynamic emotion time series data; reconstructing the emotion pattern phase space of the dynamic emotion time series data according to the emotion transition pattern recognition data to obtain emotion pattern phase space reconstruction data; Performing nonlinear spatial correlation analysis on the emotional pattern phase space reconstruction data to obtain nonlinear spatial correlation data; According to the nonlinear spatial correlation data, the emotional pattern phase space reconstruction data is subjected to Poincare cross section analysis to obtain the nonlinear phase space Poincare cross section; the nonlinear phase space Poincare cross section is subjected to spatial trajectory periodicity analysis to obtain spatial trajectory periodicity data; Based on the bifurcation theory technology, the trajectory structure data of the spatial trajectory is obtained by performing trajectory structure analysis on the periodicity of the spatial trajectory. Local emotion structure feature data is obtained by performing local emotion structure feature analysis based on the spatial trajectory structure data and the spatial trajectory periodicity data.

5. The inner emotion analysis system based on psychology information big data according to claim 3 is characterized in that: The dynamic sentiment driver analysis includes: Performing pattern-context mapping on the emotion transition pattern recognition data to obtain emotion pattern-context data; Based on the local emotion structure feature data, the emotion pattern situation data is mapped to the situation distribution to obtain the local emotion structure situation distribution data. The mapping of the local emotion structure situation distribution data can reveal the distribution and distribution rules of emotions in various situations. A structural contextual distribution logistic regression analysis was performed on the local emotion structure contextual distribution data to obtain contextual distribution logistic regression data. Based on the contextual distribution logistic regression data, a logistic linear factor analysis was performed on the local emotion structure contextual distribution data to obtain contextual emotion logical factor data. The structural contextual distribution logistic regression analysis and logistic linear factor analysis can identify the correlation and influencing factors between context and emotion, and help understand the changing patterns and driving factors of emotions in different contexts. Based on the recurrent neural network, a logical situation perception model is constructed for the situation emotion logical factor data to obtain a logical situation perception model. The logical situation perception model constructed based on the recurrent neural network can model and predict the complex relationship between situation and emotion. Based on the logical situation perception model, the pre-logical factor prediction of the situational emotional logical factor data is performed to obtain the pre-logical factor data. The prediction of the pre-logical factor data can help predict the emotional state that may occur in future situations and provide early warning and guidance for emotional management and regulation; Dynamic emotion driving factor analysis is conducted based on the pre-logical factor data and the situational emotion logical factor data to obtain dynamic emotion driving factor data. The final dynamic emotion driving factor data will comprehensively consider the complex relationship between emotions, situations and other factors, which will help to deeply understand the internal mechanism and dynamic evolution process of emotion changes.

6. The intrinsic emotion analysis system based on psychology information big data according to claim 1 is characterized in that: The module for identifying abnormal trends in internal emotions includes the following functions: Based on the dynamic emotion driving factor data, the local emotion structure feature data is analyzed for the local emotion stage evolution to obtain the local emotion stage evolution data; Evaluate the emotion volatility of the local emotion phase evolution data to obtain emotion volatility evaluation data; According to the emotion volatility evaluation data, the emotion transition state of the local emotion stage evolution data at different emotion stages is analyzed to obtain the emotion transition state data; The abnormal inner emotion trend is identified on the emotion transition state data to obtain abnormal inner emotion trend data.

7. The inner emotion analysis system based on psychology information big data according to claim 6 is characterized in that: The local emotion stage evolution analysis of local emotion structure feature data includes: Based on the dynamic emotion driving factor data, the emotion driving factor emotion relationship adjacency table is constructed for the local emotion structure feature data to obtain the emotion driving relationship adjacency table; Perform the same adjacency table structure depth traversal on the emotion-driven relationship adjacency table to obtain the adjacency table structure depth data; According to the depth data of the adjacency table structure, the complexity of the relationships between different nodes in the emotion-driven relationship adjacency table is evaluated to obtain the table node relationship complexity data; The emotion-driven relationship adjacency table is partitioned according to the table node relationship complexity data to obtain the emotion-driven relationship partitioned adjacency table; Performing local implicit numerical calculation on the adjacency table of emotion-driven relationship partitioning to obtain local relationship implicit numerical data; performing adaptive step-size control on the adjacency table of emotion-driven relationship partitioning according to the local relationship implicit numerical data to obtain adjacency table step-size control data; According to the local relationship implicit numerical data and the adjacency table step size control data, the local error control Euler-Cromer numerical calculation is performed on the adjacency table of the emotion-driven relationship partition to obtain the local error control relationship numerical data; The local emotion stage evolution data is analyzed on the local emotion structure feature data according to the local error control relationship numerical data to obtain the local emotion stage evolution data.

8. The inner emotion analysis system based on psychology information big data according to claim 1 is characterized in that: The local structure complexity of the local emotion manifold learning data is calculated using the manifold learning data point structure analysis algorithm, where the manifold learning data point structure analysis algorithm is as follows: ； Where, Represents the complexity result value of the local data point, represents the nonlinear correlation coefficient of the local sentiment manifold learning data, represents the data volume coefficient of the local sentiment manifold learning data, Represents the estimated difficulty of computing data for local sentiment manifold learning data, The dimensionality coefficient of the local sentiment manifold learning data, Indicates the calculation of the maximum time value, Represents the weight coefficient of the neighborhood data point in the local sentiment manifold learning data, Represents the error adjustment value for the manifold learning data point structure analysis algorithm.

9. The inner emotion analysis system based on psychology information big data according to claim 1 is characterized in that: The complexity local constraint processing of local emotion manifold learning data includes: Extracting data points from the local emotion manifold learning data according to the local data point complexity data to obtain local emotion manifold learning data points; performing local density calculation on the local emotion manifold learning data points to obtain local density data of the data points; Establish the nearest neighbor data point relationship of the local sentiment manifold learning data based on the local density data of the data points to obtain the nearest neighbor data point relationship data; Based on the nearest neighbor data point relationship data, the structural relationship normal distribution analysis is performed on the local emotion manifold learning data points to obtain the data point relationship normal distribution data; the dispersion degree of the data point relationship normal distribution data is calculated to obtain the relationship distribution dispersion degree data; According to the relationship distribution discrete degree data, the data point relationship normal distribution data and the neighboring data point relationship data, the local sentiment manifold learning data points are subjected to structural connectivity distribution constraints to obtain structural connectivity distribution constraint data; According to the structural connectivity distribution constraint data, the local emotion manifold learning data points are processed with local complexity constraints to obtain the local data point complexity constraint data.

Images