Psychological potential risk assessment method based on multi-modal grading model

Abstract

The invention discloses a psychological potential risk assessment method based on a multi-modal grading model. The method comprises the following steps: S1, collecting various basic data and psychological risk types of a target psychology; the method has the beneficial effects that the multi-modal grading model can be constructed through the multi-modal algorithm and the collected information data, so that the evaluation report of the psychological potential risk can be more intuitively and comprehensively displayed through the multi-modal grading model; different evaluation modes can be adopted to obtain required evaluation results and evaluation risk levels under different modes by using a multi-mode grading model and a multi-mode algorithm, and a required monitoring mode and a required monitoring threshold can be set, so that a latest evaluation report and a latest risk level can be generated in time once the risk threshold is exceeded; further, reasonable suggestions and intervention measures are given, and the accuracy of an evaluation result can be guaranteed while the potential psychological risk can be evaluated through the settings.

Description

Technical Field

[0001] The present invention relates to the field of potential psychological risks, and in particular to a potential psychological risk assessment method based on a multimodal grading model. Background Art

[0002] Psychological risk refers to the risk factors that increase the chances of risk accidents and the extent of losses due to people's carelessness and indifference. It is an intangible factor related to people's psychological state. It refers to the factors that increase the probability of risk accidents and the extent of losses due to people's carelessness, indifference, luck or dependence on insurance.

[0003] With the development of society, potential psychological risks are becoming increasingly prominent in today's society. However, the assessment of potential psychological risks in existing technologies is not comprehensive and specific enough, and it is impossible to establish a connection with the multimodal grading model, resulting in the inability to accurately assess potential psychological risks. This makes the psychological risks in daily life always in a potential hidden state, and once the potential psychological risk level exceeds the threshold, the assessment results will be confused. Summary of the invention

[0004] The purpose of the present invention is to provide a psychological potential risk assessment method based on a multimodal grading model to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for assessing potential psychological risks based on a multimodal grading model, comprising the following contents:

[0006] S1: Collect basic data and psychological risk types of the target psychology;

[0007] S2: Complete preprocessing work after acquiring the collected data;

[0008] S3: Construct a multimodal classification model based on multimodal algorithms and using preprocessed data;

[0009] S4: Make risk assessment on target psychology based on multimodal grading model and generate risk level;

[0010] S5: Determine monitoring methods and set monitoring thresholds based on the psychological risk assessment conclusions and risk levels;

[0011] S6: When the target psychological risk level exceeds the threshold, timely warning is issued, and psychological risk and level assessment is conducted again based on the multimodal grading model;

[0012] S7: Based on the newly derived psychological risk and level assessment conclusions, re-collect data in combination with the goals to construct a new multimodal grading model.

[0013] Preferably, the step S1 specifically includes the following:

[0014] S11: Determine potential psychological risk assessment targets;

[0015] S12: Collect various psychological data and personal situation data for the identified risk targets;

[0016] S13: Classify the potential psychological risk types of risk targets;

[0017] S14: Collect relevant risk type data from the website for the psychological risk type of the risk target.

[0018] Preferably, the step S2 specifically includes the following:

[0019] S21: pre-processing the data collected in step S1;

[0020] S22: Segment and classify the collected data, screen out the complex data, and after completing the segmentation and classification of the remaining practical data, collect relevant data for additional supplementation.

[0021] Preferably, the step S3 specifically includes the following:

[0022] S31: Data alignment based on multimodal algorithms to ensure consistency in time and content of data in different modalities;

[0023] S32: Integrate multimodal data and make full use of information data of each modality;

[0024] S33: Use multimodal algorithms to build a unified representation space so that data from different modalities can be understood and combined with each other;

[0025] S34: Construct a multimodal classification model based on multimodal algorithms and using collected data.

[0026] Preferably, the step S4 specifically includes the following:

[0027] S41: Using the constructed multimodal classification model and based on the target data, multi-scenario evaluation is performed through the model;

[0028] S42: Generate descriptive text from images based on a multimodal hierarchical model;

[0029] S43: Answer text questions based on images based on a multimodal hierarchical model, fusing image and text information;

[0030] S44: recognize and synthesize speech, convert speech into text or convert text into speech, and combine speech with other modalities for multimodal interaction;

[0031] S45: Based on the above steps, the potential psychological risks are scored and analyzed by using scale assessment method, clinical interview method, behavioral observation method, case analysis method and psychological test method, and the assessment conclusion is drawn. At the same time, the target potential psychological risk level is classified.

[0032] Preferably, the step S5 specifically includes the following:

[0033] S51: Based on the conclusions and target levels obtained through the multimodal grading model, a multimodal algorithm is used to set a monitoring method for the target psychological risk, and at the same time, a monitoring threshold is set according to the target parameter data;

[0034] S52: The monitoring method adopts a monitoring method set based on a multimodal grading model using multimodal emotion recognition.

[0035] Preferably, the step S6 specifically includes the following:

[0036] S61: Once the monitoring exceeds the set threshold, timely warning work will be carried out, and the results of the multimodal hierarchical model analysis will generate a psychological cognitive assessment report;

[0037] S62: At the same time, the target's current psychological potential risk is recalculated and classified into risk levels through models and algorithms;

[0038] S63: Based on the above assessment report and the classification level, conduct an in-depth analysis of potential psychological risks and provide reasonable suggestions and intervention measures in a timely manner.

[0039] Preferably, the step S7 specifically includes the following:

[0040] S71: After taking intervention measures for the target's potential psychological risks, re-collect the latest data of the target based on the new potential psychological risk assessment report and the new potential psychological risk level derived from the multimodal grading model;

[0041] S72: After preprocessing the latest collected data, a multimodal grading model is reconstructed based on the multimodal algorithm, the new assessment report and the potential psychological risk level to re-evaluate the target's potential psychological risk.

[0042] Compared with the prior art, the beneficial effects of the present invention are: a multimodal grading model can be constructed through a multimodal algorithm and the collected information data, and then the multimodal grading model can more intuitively and comprehensively display the assessment report of potential psychological risks. Moreover, by using the multimodal grading model and the multimodal algorithm, different assessment methods can be adopted to obtain the required assessment results and assessment risk levels under different modes, and the required monitoring methods and monitoring thresholds can be set, and once the risk threshold is exceeded, the latest assessment report and the latest risk level can be generated in time, and reasonable suggestions and intervention measures can be given. Through the above settings, the potential psychological risks can be assessed while ensuring the accuracy of the assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the process flow of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 The present invention provides a technical solution: a method for assessing potential psychological risks based on a multimodal grading model, comprising the following contents:

[0046] S1: Collect basic data and psychological risk types of the target psychology;

[0047] S2: Complete preprocessing work after acquiring the collected data;

[0048] S3: Construct a multimodal classification model based on multimodal algorithms and using preprocessed data;

[0049] S4: Make risk assessment on target psychology based on multimodal grading model and generate risk level;

[0050] S5: Determine monitoring methods and set monitoring thresholds based on the psychological risk assessment conclusions and risk levels;

[0051] S6: When the target psychological risk level exceeds the threshold, timely warning is issued, and psychological risk and level assessment is conducted again based on the multimodal grading model;

[0052] S7: Based on the newly derived psychological risk and level assessment conclusions, re-collect data in combination with the goals to construct a new multimodal grading model.

[0053] Wherein, the step S1 specifically includes the following:

[0054] S11: Determine potential psychological risk assessment targets;

[0055] S12: Collect various psychological data and personal situation data for the identified risk targets;

[0056] S13: Classify the potential psychological risk types of risk targets;

[0057] S14: Collect relevant risk type data from the website for the psychological risk type of the risk target;

[0058] Wherein, the step S2 specifically includes the following:

[0059] S21: pre-processing the data collected in step S1;

[0060] S22: Segment and classify the collected data, screen and eliminate the complex data, and collect relevant data for additional supplementation after segmenting and classifying the remaining practical data;

[0061] Wherein, the step S3 specifically includes the following:

[0062] S31: Data alignment based on multimodal algorithms to ensure consistency in time and content of data in different modalities;

[0063] S32: Integrate multimodal data and make full use of information data of each modality;

[0064] S33: Use multimodal algorithms to build a unified representation space so that data from different modalities can be understood and combined with each other;

[0065] S34: Construct a multimodal classification model based on multimodal algorithms and using collected data;

[0066] Wherein, the step S4 specifically includes the following:

[0067] S41: Using the constructed multimodal classification model and based on the target data, multi-scenario evaluation is performed through the model;

[0068] S42: Generate descriptive text from images based on a multimodal hierarchical model;

[0069] S43: Answer text questions based on images based on a multimodal hierarchical model, fusing image and text information;

[0070] S44: recognize and synthesize speech, convert speech into text or convert text into speech, and combine speech with other modalities for multimodal interaction;

[0071] S45: Based on the above steps, the potential psychological risks are scored and analyzed by using the scale assessment method, clinical interview method, behavioral observation method, case analysis method and psychological test method, and the assessment conclusion is drawn, and the target potential psychological risk level is graded;

[0072] Wherein, the step S5 specifically includes the following:

[0073] S51: Based on the conclusions and target levels obtained through the multimodal grading model, a multimodal algorithm is used to set a monitoring method for the target psychological risk, and at the same time, a monitoring threshold is set according to the target parameter data;

[0074] S52: wherein the monitoring method adopts a monitoring method set by using multimodal emotion recognition based on a multimodal classification model;

[0075] Wherein, the step S6 specifically includes the following:

[0076] S61: Once the monitoring exceeds the set threshold, timely warning work will be carried out, and the results of the multimodal hierarchical model analysis will generate a psychological cognitive assessment report;

[0077] S62: At the same time, the target's current psychological potential risk is recalculated and classified into risk levels through models and algorithms;

[0078] S63: Based on the above assessment report and the classification level, conduct an in-depth analysis of potential psychological risks and provide reasonable suggestions and intervention measures in a timely manner;

[0079] Wherein, the step S7 specifically includes the following:

[0080] S71: After taking intervention measures for the target's potential psychological risks, re-collect the latest data of the target based on the new potential psychological risk assessment report and the new potential psychological risk level derived from the multimodal grading model;

[0081] S72: After preprocessing the latest collected data, a multimodal grading model is reconstructed based on the multimodal algorithm, the new assessment report and the potential psychological risk level to re-evaluate the target's potential psychological risk.

[0082] Specifically, when using the present invention, determine the potential psychological risk assessment target, collect various psychological data and personal situation data for the determined risk target, classify the potential psychological risk type of the risk target, collect relevant risk type data from the website for the psychological risk type of the risk target, perform preprocessing on the various data collected in step S1, subdivide and classify the collected data, screen and eliminate the complex data, and at the same time, collect relevant data for additional supplementation after completing the subdivision and classification of the remaining practical data, and align the multimodal data based on the multimodal algorithm to ensure the consistency of data in different modes in time and content. Integrate and make full use of the information data of each modality. Use multimodal algorithms to build a unified representation space so that data of different modalities can understand and combine each other. Based on multimodal algorithms and using collected data to build a multimodal classification model, use the constructed multimodal classification model and based on target data to conduct multi-scenario evaluation through the model. Generate descriptive text based on the image based on the multimodal classification model. Answer text questions based on the image based on the multimodal classification model, fuse image and text information, recognize and synthesize speech, and convert speech into text or convert text into speech, combine speech with other modalities for multimodal interaction. Based on the above steps, match at the same time The scale assessment method, clinical interview method, behavioral observation method, case analysis method and psychological test method are used to score and analyze the potential psychological risks, and the assessment conclusion is obtained. At the same time, the target psychological potential risk level is divided into levels. Based on the multimodal grading model, the conclusions obtained from the assessment and the target level are used to set the monitoring method for the target psychological risk using the multimodal algorithm, and at the same time, the monitoring threshold is set according to the target parameter data. The monitoring method is based on the multimodal grading model and the multimodal emotion recognition is used to set the monitoring method. Once the monitoring exceeds the set threshold, an early warning is carried out in time. The results of the multimodal grading model analysis generate a psychological cognitive assessment report and at the same time, the model is used to The model and algorithm recalculate the target's current psychological potential risks and divide the risk levels. Based on the above-derived assessment report and the divided levels, an in-depth analysis of the psychological potential risks is conducted, and reasonable suggestions and intervention measures are given in a timely manner. After completing the intervention measures for the target's psychological potential risks, based on the multimodal grading model's newly derived psychological potential risk assessment report and the newly divided psychological potential risk levels, the target's latest data is collected again. After preprocessing the latest collected data again, the multimodal grading model is rebuilt based on the multimodal algorithm, the new assessment report and psychological potential risk levels, to re-evaluate the target's potential psychological risks.

[0083] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A psychological potential risk assessment method based on a multimodal grading model, characterized in that: Includes the following: S1: Collect basic data and psychological risk types of the target psychology; S2: Complete preprocessing work after acquiring the collected data; S3: Construct a multimodal classification model based on multimodal algorithms and using preprocessed data; S4: Make risk assessment on target psychology based on multimodal grading model and generate risk level; S5: Determine monitoring methods and set monitoring thresholds based on the psychological risk assessment conclusions and risk levels; S6: When the target psychological risk level exceeds the threshold, timely warning is issued, and psychological risk and level assessment is conducted again based on the multimodal grading model; S7: Based on the newly derived psychological risk and level assessment conclusions, re-collect data in combination with the goals to construct a new multimodal grading model.

2. The method for assessing psychological potential risks based on a multimodal grading model according to claim 1, characterized in that: The step S1 specifically includes the following: S11: Determine potential psychological risk assessment targets; S12: Collect various psychological data and personal situation data for the identified risk targets; S13: Classify the potential psychological risk types of risk targets; S14: Collect relevant risk type data from the website for the psychological risk type of the risk target.

3. The method for assessing psychological potential risks based on a multimodal grading model according to claim 2, characterized in that: The step S2 specifically includes the following: S21: pre-processing the data collected in step S1; S22: Segment and classify the collected data, screen out the complex data, and after completing the segmentation and classification of the remaining practical data, collect relevant data for additional supplementation.

4. The method for assessing potential psychological risks based on a multimodal grading model according to claim 3, characterized in that: The step S3 specifically includes the following: S31: Data alignment based on multimodal algorithms to ensure consistency in time and content of data in different modalities; S32: Integrate multimodal data and make full use of information data of each modality; S33: Use multimodal algorithms to build a unified representation space so that data from different modalities can be understood and combined with each other; S34: Construct a multimodal classification model based on multimodal algorithms and using collected data.

5. The method for assessing potential psychological risks based on a multimodal grading model according to claim 4, characterized in that: The step S4 specifically includes the following: S41: Using the constructed multimodal classification model and based on the target data, multi-scenario evaluation is performed through the model; S42: Generate descriptive text from images based on a multimodal hierarchical model; S43: Answer text questions based on images based on a multimodal hierarchical model, fusing image and text information; S44: recognize and synthesize speech, convert speech into text or convert text into speech, and combine speech with other modalities for multimodal interaction; S45: Based on the above steps, the potential psychological risks are scored and analyzed by using scale assessment method, clinical interview method, behavioral observation method, case analysis method and psychological test method, and the assessment conclusion is drawn. At the same time, the target potential psychological risk level is classified.

6. The method for assessing potential psychological risks based on a multimodal grading model according to claim 5, characterized in that: The step S5 specifically includes the following: S51: Based on the conclusions and target levels obtained through the multimodal grading model, a multimodal algorithm is used to set a monitoring method for the target psychological risk, and at the same time, a monitoring threshold is set according to the target parameter data; S52: The monitoring method adopts a monitoring method set based on a multimodal grading model using multimodal emotion recognition.

7. The method for assessing potential psychological risks based on a multimodal grading model according to claim 6, characterized in that: The step S6 specifically includes the following: S61: Once the monitoring exceeds the set threshold, timely warning work will be carried out, and the results of the multimodal hierarchical model analysis will generate a psychological cognitive assessment report; S62: At the same time, the target's current psychological potential risk is recalculated and classified into risk levels through models and algorithms; S63: Based on the above assessment report and the classification level, conduct an in-depth analysis of potential psychological risks and provide reasonable suggestions and intervention measures in a timely manner.

8. The method for assessing potential psychological risks based on a multimodal grading model according to claim 7, characterized in that: The step S7 specifically includes the following: S71: After taking intervention measures for the target's potential psychological risks, re-collect the latest data of the target based on the new potential psychological risk assessment report and the new potential psychological risk level derived from the multimodal grading model; S72: After preprocessing the latest collected data, a multimodal grading model is reconstructed based on the multimodal algorithm, the new assessment report and the potential psychological risk level to re-evaluate the target's potential psychological risk.

Images