Delayed symptom identification and intervention system

Abstract

The invention discloses a delayed symptom recognition and intervention system which comprises a delayed symptom recognition system and a delayed symptom intervention system. The delay symptom identification system comprises a multi-modal data acquisition module used for collecting data; the data analysis and mining module is used for preprocessing the data and evaluating the characteristics and level of the delay behavior; and the delayed symptom recognition result module is used for obtaining a delayed symptom recognition result and the level of related core cognitive competence. The delay symptom intervention system comprises a personalized intervention module used for generating a personalized cognition and nerve regulation intervention scheme; the intervention guidance feedback module is used for performing delay symptom intervention training; and the intervention data management module is used for recording intervention training data. A multi-modal data acquisition module is designed in the delayed symptom identification system, and the delayed symptom of an individual is comprehensively evaluated through multiple dimensions, so that actual actions can be conveniently captured to evaluate the delayed symptom; by setting a personalized intervention module, a cognition and nerve regulation intervention scheme is customized for a specific individual.

Description

Technical Field

[0001] The present invention relates to an identification and intervention system, and in particular to a procrastination identification and intervention system. Background Art

[0002] Procrastination is a failure of self-regulation. It is a behavior that postpones planned things even when harmful consequences are foreseeable. Procrastination not only seriously affects people's academic performance and work efficiency, but also reduces a person's subjective sense of well-being, and thus damages physical and mental health.

[0003] Existing methods for identifying procrastination mainly rely on a combination of scales and verbal reports. However, this assessment tool has the problem of difficulty in capturing the actual behavior of individuals. Existing intervention methods are also greatly limited in terms of the effectiveness and long-term nature of improving procrastination, especially in meeting the differentiated needs of individuals. Summary of the invention

[0004] In order to solve the deficiencies of the above-mentioned technologies, the present invention provides a procrastination identification and intervention system.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a procrastination identification and intervention system, including a procrastination identification system and a procrastination intervention system;

[0006] The procrastination recognition system includes a multimodal data acquisition module, a data analysis and mining module, and a procrastination recognition result module; the multimodal data acquisition module is used to collect cognitive ability data, skin electricity data, pulse wave data, voice data, and expression data; the data analysis and mining module is used to pre-process the data and analyze the data in combination with the deep learning algorithm of CNN and bi-RNN to evaluate the characteristics and level of procrastination behavior. The convolution layer in CNN is followed by a ReLU activation function, which is a nonlinear output result output by the neuron after linear transformation. The formula of the ReLU activation function is:

[0007] f(x)=max(0,x);

[0008] Among them, x is the input vector from the previous layer of neural network entering the neuron;

[0009] The procrastination identification result module is used to obtain the procrastination identification results and the level of related core cognitive abilities based on the data analysis results;

[0010] The procrastination intervention system includes a personalized intervention module, an intervention guidance and feedback module, and an intervention data management module; the personalized intervention module is used to generate personalized cognitive and neural regulation intervention plans based on the procrastination identification results; the intervention guidance and feedback module is used to conduct procrastination intervention training; and the intervention data management module is used to record intervention training data.

[0011] Further, the multimodal data collection module includes an application association unit for guiding the user through the association authorization process through a user-friendly interface and clearly stating the purpose and scope of use of the data;

[0012] The cognitive variable collection unit is used to collect cognitive variables by using the Emotion Regulation Difficulties Scale, the Youth Self-Control Dual Systems Scale, and the Anticipatory Imagination Scale;

[0013] A skin electricity and pulse wave acquisition unit, used for acquiring skin electricity and pulse waves using a physiological sensor device;

[0014] The voice and facial expression acquisition unit is used to capture voice and facial expressions using a high-precision camera to collect emotion-related expression data.

[0015] Furthermore, the data analysis and mining module includes a data preprocessing unit for collating and cleaning the data collected by the multimodal data collection module to ensure the input quality of the data;

[0016] The algorithm model evaluation unit is used to optimize the algorithm model for accurately identifying and classifying procrastination behavior characteristics by using a deep learning algorithm that combines CNN and bi-RNN, that is, based on the spatial feature extraction capability of CNN and the dynamic characteristic capture capability of bi-RNN for time series data.

[0017] Furthermore, the training of the algorithm model is based on the mean square error loss function, which measures the average square error between the model prediction value and the actual value and optimizes the classification accuracy of the algorithm model. Based on the Adam optimizer, the accuracy and stability of the algorithm model prediction are improved. The formula of the mean square error loss function MSE is:

[0018]

[0019] Where n is the number of samples; x i is the true value of the i-th sample; y i is the predicted value of the model for the i-th sample.

[0020] Furthermore, the data are organized and cleaned, including data denoising, missing value processing, outlier detection and data standardization.

[0021] Furthermore, the procrastination identification result module includes a procrastination identification result unit, which is used to obtain the procrastination identification result through the evaluation result of the algorithm model evaluation unit in the data analysis and mining module;

[0022] The core cognitive ability unit is used to evaluate the development of core cognitive abilities by analyzing cognitive variables and comparing the analyzed cognitive variables with the norms of different age groups.

[0023] Further, the personalized intervention module includes a cognitive intervention unit for specifying a personalized cognitive intervention process according to the core cognitive ability level in the procrastination identification result module;

[0024] A neural regulatory unit is used to regulate a triple-structure neural network model based on procrastination behavior to improve severe procrastination behavior.

[0025] Furthermore, the cognitive intervention unit includes a cognitive strategy learning subunit and a cognitive ability training subunit; the cognitive strategy learning subunit includes a starting difficulty strategy, an emotion regulation strategy, a self-control strategy, and an anticipatory imagination strategy, which are used from the task initiation to the task execution stage;

[0026] The triple-structure neural network model based on procrastination behavior includes the parahippocampus and ventromedial prefrontal cortex of the anticipatory imagination network, the dorsolateral prefrontal cortex and anterior cingulate gyrus of the self-control network, and the orbitofrontal cortex and insula of emotion regulation.

[0027] Furthermore, starting difficulty strategies were designed based on the SMART principles and self-monitoring theory; emotion regulation strategies were designed based on the process model of emotion regulation; self-control strategies were designed based on the dual-system theory of self-control; and anticipatory imagery strategies were designed based on the 2×2 anticipatory imagery model of procrastination.

[0028] Further, the intervention data management module includes a training duration unit for recording the duration of training;

[0029] The training effect unit is used to record the effect of training; combined with the training duration analysis to evaluate and optimize the training plan to ensure the personalization and timeliness of the intervention.

[0030] The present invention discloses a procrastination identification and intervention system. A multimodal data acquisition module is designed in the procrastination identification system. Through sensors, cognitive ability, skin electricity, pulse wave, language and expression data are monitored to comprehensively evaluate the procrastination symptoms of individuals from multiple dimensions, making it easier to capture actual actions for procrastination assessment. By setting up a personalized intervention module, cognitive and neural regulation intervention plans can be customized for specific individuals, taking into account their respective core cognitive abilities and specific needs, providing more accurate and effective treatments, meeting individual differentiated needs, and maintaining effectiveness and long-term effectiveness in improving procrastination. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0032] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0033] Embodiment 1:

[0034] like Figure 1 A procrastination identification and intervention system shown includes a procrastination identification system and a procrastination intervention system;

[0035] The procrastination recognition system includes a multimodal data acquisition module, a data analysis and mining module, and a procrastination recognition result module; the multimodal data acquisition module is used to collect cognitive ability data, skin electricity data, pulse wave data, voice data, and expression data; the data analysis and mining module is used to pre-process the data and analyze the data in combination with the deep learning algorithm of CNN and bi-RNN to evaluate the characteristics and level of procrastination behavior. The convolution layer in CNN is followed by a ReLU activation function, which is a nonlinear output result output by the neuron after linear transformation. The formula of the ReLU activation function is:

[0036] f(x)=max(0,x);

[0037] Among them, x is the input vector from the previous layer of neural network entering the neuron;

[0038] The procrastination identification result module is used to obtain the procrastination identification results and the level of related core cognitive abilities based on the data analysis results;

[0039] The procrastination intervention system includes a personalized intervention module, an intervention guidance feedback module, and an intervention data management module; the personalized intervention module is used to generate personalized cognitive and neural regulation intervention plans based on the procrastination identification results; the intervention guidance feedback module is used to conduct procrastination intervention training; and the intervention data management module is used to record intervention training data. At the same time, the intervention guidance feedback module provides real-time feedback and guidance to effectively reduce individual procrastination behavior.

[0040] The multimodal data collection module includes an application association unit, which is used to guide users to complete the association authorization process through a user-friendly interface and clearly explain the purpose and scope of data use; the data is collected using a multi-source data acquisition method, which includes online acquisition, sensor data acquisition, and mobile application data acquisition. Since the application platforms used by users are diverse, when performing application association, integrating online platforms and mobile applications at the same time can obtain more comprehensive data.

[0041] The cognitive variable collection unit, i.e., online data collection, is used to collect cognitive variables by using the Emotion Regulation Difficulties Scale, the Youth Self-Control Dual Systems Scale, and the Anticipatory Imagination Scale;

[0042] The skin electricity and pulse wave acquisition unit is used to collect skin electricity and pulse waves using physiological sensing devices; that is, sensor data acquisition is to obtain the user's physiological data through the use of heart rate monitors, skin monitors and motion trackers, and the hardware equipment can connect to the user's mobile phones, tablets and computer devices via Bluetooth to transmit the data to the system; and mobile application data collection is that the user connects his mobile application account to the system through associated authorization, so that the system obtains the user's usage data in the mobile application, and the mobile application is integrated through the software development kit to realize the regular transmission of data.

[0043] The voice and facial expression acquisition unit is used to capture voice and facial expressions using a high-precision camera to collect emotion-related expression data.

[0044] The data analysis and mining module includes a data preprocessing unit, a feature engineering processing unit and an algorithm model evaluation unit; the data preprocessing unit is used to organize and clean the data collected by the multimodal data acquisition module to ensure the input quality of the data; the feature engineering processing unit is used to extract features from the cleaned data;

[0045] The data is preprocessed, including questionnaire data preprocessing, namely cognitive ability data preprocessing, skin electricity data preprocessing, pulse wave data preprocessing, voice data preprocessing and expression data preprocessing; questionnaire data preprocessing includes the following steps:

[0046] Step 1: Clean the data to remove incomplete or invalid questionnaires;

[0047] Step 2: Perform coding conversion, digitize qualitative data, and use principal component analysis (PCA) or factor analysis to reduce the dimension of the data;

[0048] Step 3: Integrate multiple cognitive ability measurements into a small number of principal components or factors that represent key cognitive ability characteristics, helping to reduce data redundancy and improve algorithm model efficiency.

[0049] The preprocessing of skin electrodermal data includes denoising and data segmentation.

[0050] The preprocessing of pulse wave data includes signal correction and noise removal.

[0051] The preprocessing of speech data includes background noise removal and sound segmentation.

[0052] The preprocessing of expression data includes facial detection, etc.

[0053] The algorithm model evaluation unit is used to optimize the algorithm model for accurately identifying and classifying procrastination behavior characteristics by using a deep learning algorithm that combines CNN and bi-RNN, that is, based on CNN's spatial feature extraction capability and bi-RNN's ability to capture the dynamic characteristics of time series data. That is, the integrated cognitive ability features are encoded into a one-dimensional vector and spliced ​​with other pre-processed data feature vectors into a comprehensive feature matrix for algorithm model training.

[0054] Operations to organize and clean data include data denoising, missing value processing, outlier detection, and data standardization.

[0055] The procrastination identification result module includes a procrastination identification result unit, which is used to obtain the procrastination identification result through the evaluation result of the algorithm model evaluation unit in the data analysis and mining module;

[0056] The core cognitive ability unit is used to evaluate the development of core cognitive abilities by analyzing cognitive variables and comparing the analyzed cognitive variables with the norms of different age groups. The norm is the general level or distribution of levels of a certain population in the characteristics measured by the test.

[0057] The personalized intervention module includes a cognitive intervention unit, which is used to specify a personalized cognitive intervention process according to the core cognitive ability level in the procrastination identification result module;

[0058] The neural control unit is used to adjust the triple structure neural network model based on procrastination behavior to improve severe procrastination behavior. Among them, transcranial direct current stimulation technology is used to adjust the triple structure neural network model based on procrastination behavior.

[0059] The cognitive intervention unit includes a cognitive strategy learning subunit and a cognitive ability training subunit; the cognitive strategy learning subunit includes start-up difficulty strategy, emotion regulation strategy, self-control strategy and anticipatory imagination strategy, which are used to effectively improve procrastination behavior from task initiation to task execution.

[0060] The triple-structure neural network model based on procrastination behavior includes the parahippocampus and ventromedial prefrontal cortex of the anticipatory imagination network, the dorsolateral prefrontal cortex and anterior cingulate gyrus of the self-control network, and the orbitofrontal cortex and insula of emotion regulation.

[0061] Design initiation difficulty strategies based on SMART principles and self-monitoring theory; design emotion regulation strategies based on the process model of emotion regulation; design self-control strategies based on the dual-system theory of self-control; design anticipatory imagery strategies based on the 2×2 anticipatory imagery model of procrastination.

[0062] The intervention data management module includes a training duration unit, which is used to record the duration of training;

[0063] The training effect unit is used to record the effect of training; combined with the training duration analysis to evaluate and optimize the training plan to ensure the personalization and timeliness of the intervention.

[0064] Embodiment 2:

[0065] CNN is a convolutional neural network, which is a deep feedforward neural network with local connection and weight sharing. A convolutional neural network mainly consists of 5 layers, namely, input layer, convolution layer, pooling layer, fully connected layer and output layer. In this embodiment, CNN is composed of 10 convolution layers, each layer uses a 3×3 convolution kernel, followed by a ReLU activation function and a 2×2 maximum pooling layer to extract the spatial features of the data; the ReLU activation function is used as the activation function of the neuron, which is the nonlinear output result of the neuron after linear transformation; for the input vector x from the previous neural network entering the neuron, the neuron using the ReLU function will output max(0, w T x+b) to the next layer of neurons or as the output of the entire neural network; the formula of the ReLU activation function is:

[0066] f(x)=max(0,x);

[0067] Among them, x is the input vector from the previous layer of neural network entering the neuron.

[0068] RNN is a recurrent neural network, a type of neural network used to process sequence data. Bi-RNN is a bidirectional recurrent neural network. The core theorem is that it introduces two independent RNN branches, forward and reverse, to propagate information forward and backward along the timeline respectively. The two branches, forward propagation along the timeline and backward propagation along the timeline, share the same weights, but process the input from both ends of the sequence until they meet in the middle. This enables the hidden state at each moment to not only contain past input information, but also integrate the impact of future input, thus achieving a comprehensive capture of the global context of the entire sequence.

[0069] The forward RNN propagates forward along the time axis, and the hidden state is determined by the input at the current moment and the hidden state at the previous moment; the reverse RNN propagates backward along the time axis, and the hidden state is determined by the input at the current moment and the state at the next moment; the comprehensive hidden state of the bi-directional bi-RNN at each moment is spliced ​​by the forward hidden state and the reverse hidden state.

[0070] In this embodiment, the bi-RNN includes two layers of bidirectional long short-term memory networks LSTM, each layer of which includes 128 units, which are optimized for long-term dependencies of time series data. In order to reduce overfitting and improve the prediction ability of unknown data, a dropout layer is added between the two layers of bidirectional long short-term memory networks, where the dropout rate is set to 0.3; the dropout layer is used to randomly discard a part of neurons during the training process, so as to simulate multiple different network structures for training, thereby reducing the complexity of the algorithm model and enhancing the generalization ability.

[0071] The entire network integrates features through a densely connected layer containing 256 neurons and generates prediction results through an output layer with a linear activation function. The algorithm model is trained using the mean square error loss function and the Adam optimizer, which ensures efficient learning and parameter optimization in the prediction task, thereby significantly improving the generalization ability of the algorithm model in various operating environments and data sets.

[0072] The mean square error loss function is used to measure the average square error between the predicted value and the actual value of the algorithm model, and optimize the classification accuracy of the algorithm model. The formula of the mean square error loss function MSE is:

[0073]

[0074] Where n is the number of samples; x i is the true value of the i-th sample; y i is the predicted value of the algorithm model for the i-th sample.

[0075] The Adam optimizer is used to simultaneously calculate the first-order moment of the gradient, i.e. the mean, and the second-order moment, i.e. the exponential moving average of the uncentered variance, and perform bias correction on them to ensure that the gradient estimate does not tend to 0 in the early stages of training. The gradient descent method with adaptive learning rate adjusts the network parameters to further improve the accuracy and stability of the algorithm model prediction.

[0076] The workflow of the present invention is:

[0077] 1. Collect data from online platforms, sensors, and mobile applications through multimodal data acquisition modules;

[0078] Second, the data analysis and mining module uses the trained support vector machine and neural network technology to identify the individual's procrastination level based on the collected data; and determines the individual's key cognitive characteristics based on the collected data;

[0079] 3. Personalized intervention module selects personalized intervention plans according to individual cognitive characteristics, and conducts cognitive intervention and neuromodulation intervention respectively;

[0080] 4. The intervention data management module records the individual's data during the training process and adjusts the intervention plan in real time.

[0081] The method of collecting multi-source data through the multimodal data acquisition module, namely cognitive ability data, sensor data, skin electricity data, pulse wave data, voice data and expression data, can comprehensively evaluate the individual's procrastination symptoms from multiple dimensions, making it easier to capture actual actions for procrastination assessment; by setting up a personalized intervention module, it is possible to customize cognitive and neural regulation intervention plans for specific individuals, taking into account their core cognitive abilities and specific needs, and provide more accurate and effective treatments to meet the needs of individual differentiation and ensure significant effectiveness and long-term effectiveness in improving procrastination. Not only can it improve procrastination in a short period of time, but in the long run, it can form an effective self-management and adjustment mechanism to achieve lasting improvement effects.

[0082] The above implementation modes are not limitations of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the scope of the technical solution of the present invention also belong to the protection scope of the present invention.

Claims

1. A procrastination identification and intervention system, characterized in that: It includes a procrastination identification system and a procrastination intervention system; the procrastination identification system includes a multimodal data acquisition module, a data analysis and mining module, and a procrastination identification result module; the multimodal data acquisition module is used to collect cognitive ability data, skin electricity data, pulse wave data, voice data, and expression data; The data analysis and mining module is used to pre-process the data and analyze the data by combining the deep learning algorithms of CNN and bi-RNN to evaluate the characteristics and level of procrastination behavior. The convolution layer in CNN is followed by the ReLU activation function, which is the nonlinear output result of the neuron after linear transformation. The formula of the ReLU activation function is: f(x)=max(0,x); Among them, x is the input vector from the previous layer of neural network entering the neuron; The procrastination identification result module is used to obtain the procrastination identification results and the level of related core cognitive abilities based on the data analysis results; The procrastination intervention system includes a personalized intervention module, an intervention guidance and feedback module, and an intervention data management module; the personalized intervention module is used to generate personalized cognitive and neural regulation intervention plans based on the procrastination identification results; the intervention guidance and feedback module is used to conduct procrastination intervention training; and the intervention data management module is used to record intervention training data.

2. The procrastination identification and intervention system according to claim 1, characterized in that: The multimodal data acquisition module includes an application association unit, which is used to guide the user to complete the association authorization process through a user-friendly interface and clearly explain the purpose and scope of data use; The cognitive variable collection unit is used to collect cognitive variables by using the Emotion Regulation Difficulties Scale, the Youth Self-Control Dual Systems Scale, and the Anticipatory Imagination Scale; A skin electricity and pulse wave acquisition unit, used for acquiring skin electricity and pulse waves using a physiological sensor device; The voice and facial expression acquisition unit is used to capture voice and facial expressions using a high-precision camera to collect emotion-related expression data.

3. The procrastination identification and intervention system according to claim 1, characterized in that: The data analysis and mining module includes a data preprocessing unit, which is used to organize and clean the data collected by the multimodal data collection module to ensure the input quality of the data; The algorithm model evaluation unit is used to optimize the algorithm model for accurately identifying and classifying procrastination behavior characteristics by using a deep learning algorithm that combines CNN and bi-RNN, that is, based on the spatial feature extraction capability of CNN and the dynamic characteristic capture capability of bi-RNN for time series data.

4. The procrastination identification and intervention system according to claim 3, characterized in that: The training of the algorithm model is based on the mean square error loss function, which measures the average square error between the model prediction value and the actual value and optimizes the classification accuracy of the algorithm model. Based on the Adam optimizer, the accuracy and stability of the algorithm model prediction are improved. The formula of the mean square error loss function MSE is: Where n is the number of samples; x i is the true value of the i-th sample; y i is the predicted value of the model for the i-th sample.

5. The procrastination identification and intervention system according to claim 4, characterized in that: The data are sorted and cleaned, including data denoising, missing value processing, outlier detection and data standardization.

6. The procrastination identification and intervention system according to claim 1, characterized in that: The procrastination identification result module includes a procrastination identification result unit, which is used to obtain the procrastination identification result through the evaluation result of the algorithm model evaluation unit in the data analysis and mining module; The core cognitive ability unit is used to evaluate the development of core cognitive abilities by analyzing cognitive variables and comparing the analyzed cognitive variables with the norms of different age groups.

7. The procrastination identification and intervention system according to claim 1, characterized in that: The personalized intervention module includes a cognitive intervention unit, which is used to specify a personalized cognitive intervention process according to the core cognitive ability level in the procrastination identification result module; A neural regulatory unit is used to regulate a triple-structure neural network model based on procrastination behavior to improve severe procrastination behavior.

8. The procrastination identification and intervention system according to claim 7, characterized in that: The cognitive intervention unit includes a cognitive strategy learning subunit and a cognitive ability training subunit; The cognitive strategy learning subunit includes initiation difficulty strategies, emotion regulation strategies, self-control strategies, and anticipatory imagery strategies, which are used from task initiation to task execution stage; The triple-structure neural network model based on procrastination behavior includes the parahippocampus and ventromedial prefrontal cortex of the anticipatory imagination network, the dorsolateral prefrontal cortex and anterior cingulate gyrus of the self-control network, and the orbitofrontal cortex and insula of emotion regulation.

9. The procrastination identification and intervention system according to claim 8, characterized in that: The initiation difficulty strategy is designed based on the SMART principle and self-monitoring theory; Design emotion regulation strategies based on the process model of emotion regulation; Design self-control strategies based on the dual-system theory of self-control; Design anticipatory imagery strategies based on the 2×2 anticipatory imagery model of procrastination.

10. The procrastination identification and intervention system according to claim 1, characterized in that: The intervention data management module includes a training duration unit for recording the duration of training; The training effect unit is used to record the effect of training; combined with the training duration analysis to evaluate and optimize the training plan to ensure the personalization and timeliness of the intervention.

Images