Psychological consultation method and system based on artificial intelligence

Abstract

The invention discloses a psychological counseling method and system based on artificial intelligence, relates to the field of artificial intelligence, and solves the problem of poor effect of the existing psychological counseling method.The psychological counseling method comprises the steps that S1, a plurality of psychological counseling records are collected, question and answer concentration evaluation and screening are conducted on a model user corresponding to each psychological counseling record, and a psychological counseling result is obtained; the method comprises the following steps: S1, obtaining psychological counseling screening data, S2, carrying out vocabulary negative feature analysis on each effective psychological counseling record, obtaining a vocabulary negative index according to an analysis result, and obtaining interactive vocabulary analysis data, and S3, carrying out scene interactive psychological counseling on model users corresponding to the effective psychological counseling records. And S4, feeding back a consulting result to the model user, carrying out brain wave monitoring on the model user in the consulting process, and obtaining interactive behavior monitoring data according to the monitoring result. The psychological consulting method can improve the accuracy and objectivity of the psychological consulting method.

Description

Technical Field

[0001] The present invention belongs to the field of psychological counseling and relates to artificial intelligence technology, specifically to a psychological counseling method and system based on artificial intelligence. Background Art

[0002] The existing psychological counseling methods have the following specific defects when conducting psychological counseling:

[0003] 1. Although existing psychological counseling methods use artificial intelligence models to interact with users, their interaction process is unable to identify and screen the user's concentration during the interaction, thus failing to guarantee the accuracy of psychological counseling interaction data, resulting in poor counseling results;

[0004] 2. Existing psychological counseling methods provide psychological counseling to users through questionnaires or dialogues. The counseling results often depend on the questionnaire scores provided by the psychological counseling scale or the subjective judgment of the doctor being consulted, which leads to a lack of pertinence and objectivity in the counseling process.

[0005] To this end, we propose a psychological counseling method and system based on artificial intelligence. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a psychological consultation method and system based on artificial intelligence. The present invention aims to improve the pertinence and objectivity of the psychological consultation method.

[0007] The purpose of the present invention can be achieved through the following technical solution: a psychological consultation method based on artificial intelligence, comprising the following specific steps:

[0008] Step S1: Obtain several psychological consultation records, conduct question-answering concentration assessment on the model user in each psychological consultation record, screen the psychological consultations based on the assessment results, and obtain psychological consultation screening data;

[0009] Step S2: performing a vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtaining a vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and obtaining interactive vocabulary analysis data;

[0010] Step S3: Based on the psychological consultation screening data, scenario-based interactive psychological consultation is conducted on the model users corresponding to the valid psychological consultation records, and brainwave monitoring is performed on the model users during the consultation process. Based on the monitoring results, the scenario interaction abnormality corresponding to each model user is obtained to obtain interactive behavior monitoring data;

[0011] Step S4: Provide consultation result feedback to the model user based on the interactive vocabulary analysis data and interactive behavior monitoring data.

[0012] Furthermore, the step S1 further includes the following specific steps:

[0013] Step S11: In the process of psychological monitoring of the model user, a psychological consultation history period is marked, and historical question-and-answer interaction data between the model user and the psychological consultation model in the psychological consultation history period is obtained. Several different batches of psychological consultation records in the historical question-and-answer interaction data are selected, and a sample psychological consultation record is randomly selected from the multiple psychological consultation records obtained;

[0014] Step S12: monitoring physiological indicators of the model user corresponding to the sample psychological consultation record, and obtaining the consultation focus time ratio corresponding to the sample psychological consultation based on the monitoring results;

[0015] Step S13: Obtain the consultation focus time ratio corresponding to each psychological consultation record to obtain multiple consultation focus time ratios;

[0016] Step S14: Obtaining a reference interval for the consultation focus time ratio. If the consultation focus time ratio is within the reference interval, marking the corresponding psychological consultation record as a valid psychological consultation record; if the consultation focus time ratio is not within the reference interval, marking the corresponding psychological consultation record as an invalid psychological consultation record.

[0017] Step S15: defining the obtained multiple valid psychological counseling records as psychological counseling screening data.

[0018] Furthermore, the step S12 further includes the following specific steps:

[0019] Step S121: obtaining the question and answer content involved in the sample psychological consultation record, and marking the obtained question and answer content as J1 question and answer content to Ja question and answer content in the order of interaction time;

[0020] Step S122: Acquire the time point at which the psychological counseling model completes outputting the question portion of the J1 question-and-answer content, obtaining the J1 question time point; acquire the time point at which the model user completes outputting the reply portion of the J1 question-and-answer content, obtaining the J1 reply time point; and acquire the interval between the J1 question time point and the J1 reply time point, obtaining the J1 user reply time point;

[0021] Step S123: Similarly, the time point at which the psychological counseling model completes outputting the question part of the Ja Q&A content is obtained to obtain the Ja question time point, the time point at which the model user completes outputting the reply part of the Ja Q&A content is obtained to obtain the Ja reply time point, and the interval between the Ja question time point and the Ja reply time point is obtained to obtain the Ja user reply time point;

[0022] Step S124: monitoring physiological indicators of the model user who interacts with the J1 question-and-answer content, and obtaining the J1 concentration time based on the monitoring results;

[0023] The consultation focus time ratio corresponding to the sample psychological consultation is obtained by calculating the time between J1's focus time and Ja's focus time, as well as the time between J1's user response time and Ja's user response time.

[0024] The specific formula for calculating the consultation focus duration ratio corresponding to the sample psychological consultation is as follows:

[0025]

[0026] Among them, Zxb is the consultation focus time ratio corresponding to the sample psychological consultation, Zzji is Ji's focus time, Zcji is Ji's user response time, and a is the quantity value corresponding to the question and answer content in the sample psychological consultation record.

[0027] Furthermore, the step S124 further includes the following specific steps:

[0028] Conduct three-dimensional modeling of the spatial environment where the model user and the psychological counseling model output terminal are located to obtain a three-dimensional model of the terminal environment;

[0029] In the three-dimensional model of the terminal environment, the display interface corresponding to the output terminal of the psychological counseling model is marked as the first model interaction plane. A pixel point is arbitrarily selected in the first model interaction plane as the coordinate origin, and a straight line is arbitrarily drawn through the coordinate origin to obtain the first model interaction line. In the first model interaction plane, a perpendicular line to the first model interaction line is drawn through the coordinate origin to obtain the second model interaction line. A perpendicular line to the first model interaction plane is drawn through the coordinate origin to obtain the third model interaction line.

[0030] The first model interaction line is marked as the coordinate x-axis, the second model interaction line is marked as the coordinate y-axis, and the third model interaction line is marked as the coordinate z-axis. The spatial coordinate system composed of the coordinate origin, the coordinate x-axis, the coordinate y-axis, and the coordinate z-axis is marked as the scene space coordinate system.

[0031] In the scene space coordinate system, mark the center point of the model user's left pupil as the first pupil marker point, and mark the center point of the model user's right pupil as the second pupil marker point;

[0032] Using the ELG algorithm to obtain the gaze ray starting from the first pupil mark point to obtain the first pupil ray, and using the ELG algorithm to obtain the gaze ray starting from the second pupil mark point to obtain the second pupil ray;

[0033] Marking a period in which the first pupil ray intersects the first model interaction plane as a first gaze period, acquiring the area of ​​the left pupil of the target user in the first target period to obtain a real-time area value of the left pupil, obtaining a left pupil focus area value, and marking the first target period in which the real-time area value of the left pupil is greater than or equal to the left pupil focus area value as the first pupil focus period;

[0034] Marking the period during which the second pupil ray intersects the second model interaction plane as the second gaze period, acquiring the area of ​​the right pupil of the target user in the second target period to obtain a real-time area value of the right pupil, obtaining a right pupil focus area value, and marking the second target period during which the real-time area value of the right pupil is greater than or equal to the right pupil focus area value as the second pupil focus period;

[0035] The common time period of the first pupil focus period and the second pupil focus period is obtained to obtain the model user focus period, and the duration of the model user focus period is obtained to obtain the J1 focus duration;

[0036] Repeat the process of obtaining the j1 concentration time, and obtain the concentration time corresponding to the J2 question and answer content to the Ja question and answer content, and obtain the J2 concentration time to the Ja concentration time.

[0037] Furthermore, the step S2 further includes the following specific steps:

[0038] Step S21: Acquire psychological consultation screening data, acquire each valid psychological consultation record according to the psychological consultation screening data, and arbitrarily select a characteristic psychological consultation record from the multiple valid psychological consultation records acquired;

[0039] Step S22: performing a lexical depth analysis on the characteristic psychological counseling record to obtain a lexical negativity index corresponding to the characteristic psychological counseling record;

[0040] Step S23: Obtain the vocabulary negativity index corresponding to each valid psychological counseling record to obtain interactive vocabulary analysis data.

[0041] Furthermore, the step S22 further includes the following specific steps:

[0042] Step S221: obtaining the output content of the model user in the characteristic psychological consultation record to obtain the user output text, using a word recognition algorithm to obtain negative words in the user output text, and re-labeling the obtained multiple negative words as C1 negative words to Cb negative words;

[0043] Step S222: using the SO-PMI algorithm to obtain the sentiment intensities corresponding to the negative words C1 to Cb respectively, and obtaining sentiment intensity values ​​of the words C1 to Cb;

[0044] Step S223: Analyze the frequency of occurrence of C1 negative words in the user output text, and obtain the C1 word frequency index value according to the analysis result;

[0045] The step S223 further includes the following specific steps:

[0046] If the C1 negative word only appears once in the user output text, the value "1" will be used to assign the C1 word frequency index value to the parameter;

[0047] If the number of occurrences of the C1 negative word in the user output text is greater than 1, the number of occurrences of the C1 negative word in the user output text is counted to obtain the C1 negative word occurrence quantity value, and the interval length between each two consecutive occurrences of the C1 negative word is obtained to obtain multiple negative word occurrence interval lengths, and the obtained multiple negative word occurrence interval lengths are averaged to obtain the C1 negative word occurrence interval length, and the product of the C1 negative word occurrence quantity value and the C1 negative word occurrence interval length is calculated to obtain the C1 word frequency index value;

[0048] Step S224: Acquire the word frequency index values ​​corresponding to the negative words C2 to Cb to obtain the word frequency index values ​​C2 to Cb;

[0049] Step S225: Calculating the vocabulary negativity index corresponding to the characteristic psychological counseling record by combining the vocabulary sentiment intensity value C1 to the vocabulary sentiment intensity value Cb and the vocabulary frequency index value C1 to the vocabulary frequency index value Cb;

[0050] The vocabulary negativity index corresponding to the characteristic psychological counseling records is calculated using the following formula:

[0051]

[0052] Among them, Cxz is the vocabulary negativity index corresponding to the characteristic psychological counseling record, Qqdi is the Ci vocabulary emotion intensity value, Cpzi is the Ci word frequency index value, and b is the quantity value corresponding to the negative vocabulary.

[0053] Furthermore, the step S3 further includes the following specific steps:

[0054] Step S31: Acquire psychological consultation screening data, acquire model users corresponding to valid psychological consultation records based on the psychological consultation screening data, and select a sample model user from the acquired multiple model patients;

[0055] Step S32: When providing psychological consultation to the model user during interaction, p different types of psychological consultation scenarios are set, and a sample consultation scenario is randomly selected from the set psychological consultation scenarios;

[0056] Step S33: performing brain wave monitoring on the sample model user in the sample consultation scenario, and judging the abnormality of scene interaction of the sample consultation scenario based on the monitoring;

[0057] Step S34: performing brainwave monitoring on the sample model users in each psychological counseling scenario, judging the abnormality of scene interaction in the psychological counseling scenario based on the monitoring, and counting the number of abnormal interaction scenarios to obtain the abnormal number value of the interaction scenarios;

[0058] Step S35: Calculate the ratio of the number of interaction scene anomalies to p to obtain the scene interaction anomaly degree corresponding to the sample model user;

[0059] Step S36: Obtain the scene interaction abnormality corresponding to each model user to obtain interaction behavior monitoring data;

[0060] The step S33 further includes the following specific steps:

[0061] Step S331: Marking a number of scene feature points in the sample consultation scene, and marking the scene feature points as T1 scene feature points to Td scene feature points in chronological order;

[0062] Step S332: performing real-time brain wave monitoring on the sample model user at the T1 scene feature point, and obtaining a T1 brain wave monitoring index value based on the monitoring result;

[0063] Step S333: Obtain the brainwave monitoring index values ​​corresponding to the scene feature points T2 to Td of the sample model user, and obtain the brainwave monitoring index values ​​T2 to Ta;

[0064] Step S334: obtaining the brain wave reference index interval corresponding to the T1 scene feature point to the Td scene feature point, and obtaining the T1 brain wave reference index interval to the Td brain wave reference index interval;

[0065] Step S335: If the T1 EEG monitoring index value is within the T1 EEG benchmark index interval, the T1 scene feature point is marked as a normal scene feature point; if the T1 EEG monitoring index value is not within the T1 EEG benchmark index interval, the T1 scene feature point is marked as an abnormal scene feature point;

[0066] Step S336: Similarly, if the Ta brainwave monitoring index value is within the Ta brainwave benchmark index interval, then the Ta scene feature point is marked as a normal scene feature point; if the Ta brainwave monitoring index value is not within the Ta brainwave benchmark index interval, then the Ta scene feature point is marked as an abnormal scene feature point;

[0067] Step S337: Count the number of normal scene feature points and abnormal scene feature points respectively to obtain the number of normal feature points and the number of abnormal feature points. If the number of normal feature points is greater than the number of abnormal feature points, the sample consultation scene is judged to be a normal interaction scene. If the number of normal feature points is less than or equal to the number of abnormal feature points, the sample consultation scene is judged to be an abnormal interaction scene.

[0068] Furthermore, the step S332 further includes the following specific steps:

[0069] The T1 scene feature point is used as an intermediate time node to mark a T1 scene monitoring cycle, and the T1 scene monitoring cycle is divided into multiple monitoring time points, and a feature monitoring time point is arbitrarily selected from the multiple monitoring time points obtained;

[0070] Performing real-time brainwave monitoring on the sample model user at the characteristic monitoring time point, obtaining the alpha wave power corresponding to the sample model user at the characteristic monitoring time point, obtaining a first brainwave index value, obtaining the beta wave power corresponding to the sample model user at the characteristic monitoring time point, obtaining a second brainwave index value, calculating the ratio of the first brainwave index value to the second brainwave index value, and obtaining the brainwave monitoring index value corresponding to the characteristic monitoring time point;

[0071] Acquire the brainwave monitoring index value corresponding to each monitoring time point respectively to obtain multiple brainwave monitoring index values;

[0072] The monitoring time point mark is used as the horizontal axis, and the brain wave monitoring index value is marked as the vertical axis to obtain multiple brain wave monitoring coordinate points, and the obtained multiple brain wave monitoring coordinate points are connected to obtain a brain wave waveform graph;

[0073] Acquire brain wave waveform segments in the same monotonic state in the brain wave waveform graph to obtain monotonically identical waveform segments, and obtain the peak value, trough value, and peak-to-trough interval duration between the peak value and the trough value of each monotonically identical waveform segment;

[0074] The peak value, trough value and peak-to-trough interval duration corresponding to the same monotonous waveform segment are calculated by calculating the waveform change slope;

[0075] Calculate the slope of the waveform change. The specific formula is as follows:

[0076]

[0077] Among them, Bbx is the slope of the waveform change, Bfz is the peak value, Bgz is the trough value, and Jfg is the peak-to-trough interval length;

[0078] Obtaining the waveform change slope corresponding to each monotonically identical waveform segment respectively, and averaging the obtained multiple waveform change slopes to obtain the waveform average change slope;

[0079] Calculating the difference between the peak value and the trough value corresponding to each monotonic identical waveform segment, taking the absolute value of the obtained difference to obtain multiple monotonic waveform amplitudes, and averaging the obtained multiple monotonic waveform amplitudes to obtain the waveform average amplitude;

[0080] The ratio of the average waveform amplitude to the average waveform change slope is calculated to obtain the T1 EEG monitoring index value.

[0081] Furthermore, the step S4 further includes the following specific steps:

[0082] Obtaining interactive vocabulary analysis data, obtaining multiple vocabulary negativity indexes corresponding to model users based on the interactive vocabulary analysis data, and averaging the obtained multiple vocabulary negativity indexes to obtain an average vocabulary negativity index;

[0083] Obtain interactive behavior monitoring data, and obtain the scene interaction abnormality corresponding to the model user based on the interactive behavior monitoring data;

[0084] Obtaining the preset interval of vocabulary negativity index and the preset interval of scene interaction abnormality;

[0085] If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then it is judged that the model user's psychological state is normal;

[0086] If the average vocabulary negativity index is not within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then the model user's psychological state is judged to be abnormal;

[0087] If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is not within the preset scene interaction abnormality range, it is judged that the model user's psychological state is abnormal.

[0088] A psychological consultation system based on artificial intelligence, comprising:

[0089] Question and answer status module: obtains several psychological consultation records, evaluates the question and answer concentration of the model user in each psychological consultation record, and screens the psychological consultations based on the evaluation results to obtain psychological consultation screening data;

[0090] Interactive vocabulary module: Perform vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtain the vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and obtain interactive vocabulary analysis data;

[0091] Interactive behavior module: Based on the psychological consultation screening data, scenario-based interactive psychological consultation is conducted on the model users corresponding to the valid psychological consultation records. The model users' brain waves are monitored during the consultation process. Based on the monitoring results, the scenario interaction abnormality corresponding to each model user is obtained to obtain the interactive behavior monitoring data;

[0092] Consultation feedback module: Provides consultation result feedback to model users based on interactive vocabulary analysis data and interactive behavior monitoring data.

[0093] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0094] 1. The present invention collects psychological consultation records of several model users and psychological consultation models, evaluates the model user's question-answering concentration in each psychological consultation record, and screens psychological consultations based on the evaluation results, thereby improving the accuracy of psychological consultation interaction data and thus improving consultation efficiency and the accuracy of consultation results;

[0095] 2. The present invention performs a vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtains the vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and performs scene interactive psychological counseling on the model user corresponding to the valid psychological counseling record. During the consultation process, the model user's brain waves are monitored, and the scene interaction abnormality corresponding to each model user is obtained based on the monitoring results. The model user is fed back the consultation results based on the interactive vocabulary analysis data and the interactive behavior monitoring data, which can effectively ensure the pertinence of the consultation process and the objectivity of the consultation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0097] Figure 1 It is a diagram of the implementation steps of the present invention;

[0098] Figure 2 is a block diagram of the overall system of the present invention;

[0099] Figure 3 This is the brain wave waveform diagram of the present invention. DETAILED DESCRIPTION

[0100] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0101] Example 1

[0102] See also Figure 1 The present invention provides a technical solution: a psychological consultation method based on artificial intelligence, comprising the following specific steps:

[0103] Step S1: Obtain several psychological consultation records, conduct question-answering concentration assessment on the model user in each psychological consultation record, screen the psychological consultations based on the assessment results, and obtain psychological consultation screening data;

[0104] The step S1 further includes the following specific steps:

[0105] Step S11: In the process of psychological monitoring of the model user, a psychological consultation history period is marked, and historical question-and-answer interaction data between the model user and the psychological consultation model in the psychological consultation history period is obtained. Several different batches of psychological consultation records in the historical question-and-answer interaction data are selected, and a sample psychological consultation record is randomly selected from the multiple psychological consultation records obtained;

[0106] Step S12: monitoring physiological indicators of the model user corresponding to the sample psychological consultation record, and obtaining the consultation focus time ratio corresponding to the sample psychological consultation based on the monitoring results;

[0107] The step S12 further includes the following specific steps:

[0108] Step S121: obtaining the question and answer content involved in the sample psychological consultation record, and marking the obtained question and answer content as J1 question and answer content to Ja question and answer content in the order of interaction time;

[0109] Step S122: Acquire the time point at which the psychological counseling model completes outputting the question portion of the J1 question-and-answer content, obtaining the J1 question time point; acquire the time point at which the model user completes outputting the reply portion of the J1 question-and-answer content, obtaining the J1 reply time point; and acquire the interval between the J1 question time point and the J1 reply time point, obtaining the J1 user reply time point;

[0110] Step S123: Similarly, the time point at which the psychological counseling model completes outputting the question part of the Ja Q&A content is obtained to obtain the Ja question time point, the time point at which the model user completes outputting the reply part of the Ja Q&A content is obtained to obtain the Ja reply time point, and the interval between the Ja question time point and the Ja reply time point is obtained to obtain the Ja user reply time point;

[0111] Step S124: monitoring physiological indicators of the model user who interacts with the J1 question-and-answer content, and obtaining the J1 concentration time based on the monitoring results;

[0112] The step S124 further includes the following specific steps:

[0113] Conduct three-dimensional modeling of the spatial environment where the model user and the psychological counseling model output terminal are located to obtain a three-dimensional model of the terminal environment;

[0114] In the three-dimensional model of the terminal environment, the display interface corresponding to the output terminal of the psychological counseling model is marked as the first model interaction plane. A pixel point is arbitrarily selected in the first model interaction plane as the coordinate origin, and a straight line is arbitrarily drawn through the coordinate origin to obtain the first model interaction line. In the first model interaction plane, a perpendicular line to the first model interaction line is drawn through the coordinate origin to obtain the second model interaction line. A perpendicular line to the first model interaction plane is drawn through the coordinate origin to obtain the third model interaction line.

[0115] The first model interaction line is marked as the coordinate x-axis, the second model interaction line is marked as the coordinate y-axis, and the third model interaction line is marked as the coordinate z-axis. The spatial coordinate system composed of the coordinate origin, the coordinate x-axis, the coordinate y-axis, and the coordinate z-axis is marked as the scene space coordinate system.

[0116] In the scene space coordinate system, mark the center point of the model user's left pupil as the first pupil marker point, and mark the center point of the model user's right pupil as the second pupil marker point;

[0117] Using the ELG algorithm to obtain the gaze ray starting from the first pupil mark point to obtain the first pupil ray, and using the ELG algorithm to obtain the gaze ray starting from the second pupil mark point to obtain the second pupil ray;

[0118] Marking a period in which the first pupil ray intersects the first model interaction plane as a first gaze period, acquiring the area of ​​the left pupil of the target user in the first target period to obtain a real-time area value of the left pupil, obtaining a left pupil focus area value, and marking the first target period in which the real-time area value of the left pupil is greater than or equal to the left pupil focus area value as the first pupil focus period;

[0119] Marking the period during which the second pupil ray intersects the second model interaction plane as the second gaze period, acquiring the area of ​​the right pupil of the target user in the second target period to obtain a real-time area value of the right pupil, obtaining a right pupil focus area value, and marking the second target period during which the real-time area value of the right pupil is greater than or equal to the right pupil focus area value as the second pupil focus period;

[0120] The common time period of the first pupil focus period and the second pupil focus period is obtained to obtain the model user focus period, and the duration of the model user focus period is obtained to obtain the J1 focus duration;

[0121] Repeat the process of obtaining the concentration time of j1, and obtain the concentration time corresponding to the question and answer content of J2 to the question and answer content of Ja, and obtain the concentration time of J2 to Ja;

[0122] Step S125: Calculate the consultation focus duration ratio corresponding to the sample psychological consultation by dividing the concentration duration of J1 to the concentration duration of Ja and the response duration of J1 to the response duration of Ja;

[0123] The specific formula for calculating the consultation focus duration ratio corresponding to the sample psychological consultation is as follows:

[0124]

[0125] Among them, Zxb is the consultation focus time ratio corresponding to the sample psychological consultation, Zzji is Ji’s focus time, Zcji is Ji’s user response time, and a is the number of question and answer contents in the sample psychological consultation record;

[0126] Step S13: Obtain the consultation focus time ratio corresponding to each psychological consultation record to obtain multiple consultation focus time ratios;

[0127] Step S14: Obtaining a reference interval for the consultation focus time ratio. If the consultation focus time ratio is within the reference interval, marking the corresponding psychological consultation record as a valid psychological consultation record; if the consultation focus time ratio is not within the reference interval, marking the corresponding psychological consultation record as an invalid psychological consultation record.

[0128] Step S15: defining the obtained multiple valid psychological counseling records as psychological counseling screening data.

[0129] Step S2: performing a vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtaining a vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and obtaining interactive vocabulary analysis data;

[0130] The step S2 further includes the following specific steps:

[0131] Step S21: Acquire psychological consultation screening data, acquire each valid psychological consultation record according to the psychological consultation screening data, and arbitrarily select a characteristic psychological consultation record from the multiple valid psychological consultation records acquired;

[0132] Step S22: performing a lexical depth analysis on the characteristic psychological counseling record to obtain a lexical negativity index corresponding to the characteristic psychological counseling record;

[0133] The step S22 further includes the following specific steps:

[0134] Step S221: obtaining the output content of the model user in the characteristic psychological consultation record to obtain the user output text, using a word recognition algorithm to obtain negative words in the user output text, and re-labeling the obtained multiple negative words as C1 negative words to Cb negative words;

[0135] Step S222: using the SO-PMI algorithm to obtain the sentiment intensities corresponding to the negative words C1 to Cb respectively, and obtaining sentiment intensity values ​​of the words C1 to Cb;

[0136] Step S223: Analyze the frequency of occurrence of C1 negative words in the user output text, and obtain the C1 word frequency index value according to the analysis result;

[0137] The step S223 further includes the following specific steps:

[0138] If the C1 negative word only appears once in the user output text, the value "1" will be used to assign the C1 word frequency index value to the parameter;

[0139] If the number of occurrences of the C1 negative word in the user output text is greater than 1, the number of occurrences of the C1 negative word in the user output text is counted to obtain the C1 negative word occurrence quantity value, and the interval length between each two consecutive occurrences of the C1 negative word is obtained to obtain multiple negative word occurrence interval lengths, and the obtained multiple negative word occurrence interval lengths are averaged to obtain the C1 negative word occurrence interval length, and the product of the C1 negative word occurrence quantity value and the C1 negative word occurrence interval length is calculated to obtain the C1 word frequency index value;

[0140] Step S224: Acquire the word frequency index values ​​corresponding to the negative words C2 to Cb to obtain the word frequency index values ​​C2 to Cb;

[0141] Step S225: Calculating the vocabulary negative index corresponding to the characteristic psychological counseling record by combining the vocabulary sentiment intensity value C1 to the vocabulary sentiment intensity value Cb and the vocabulary frequency index value C1 to the vocabulary frequency index value Cb;

[0142] The vocabulary negativity index corresponding to the characteristic psychological counseling records is calculated using the following formula:

[0143]

[0144] Among them, Cxz is the vocabulary negativity index corresponding to the characteristic psychological counseling record, Qqdi is the Ci vocabulary emotional intensity value, Cpzi is the Ci word frequency index value, and b is the number value corresponding to the negative vocabulary;

[0145] Step S23: Obtaining the vocabulary negativity index corresponding to each valid psychological consultation record to obtain interactive vocabulary analysis data;

[0146] Step S3: Based on the psychological consultation screening data, scenario-based interactive psychological consultation is conducted on the model users corresponding to the valid psychological consultation records, and brainwave monitoring is performed on the model users during the consultation process. Based on the monitoring results, the scenario interaction abnormality corresponding to each model user is obtained to obtain interactive behavior monitoring data;

[0147] The step S3 further includes the following specific steps:

[0148] Step S31: Acquire psychological consultation screening data, acquire model users corresponding to valid psychological consultation records based on the psychological consultation screening data, and select a sample model user from the acquired multiple model patients;

[0149] Step S32: When providing psychological consultation to the model user during interaction, p different types of psychological consultation scenarios are set, and a sample consultation scenario is randomly selected from the set psychological consultation scenarios;

[0150] Step S33: performing brain wave monitoring on the sample model user in the sample consultation scenario, and judging the abnormality of scene interaction of the sample consultation scenario based on the monitoring;

[0151] The step S33 further includes the following specific steps:

[0152] Step S331: Marking a number of scene feature points in the sample consultation scene, and marking the scene feature points as T1 scene feature points to Td scene feature points in chronological order;

[0153] Step S332: performing real-time brain wave monitoring on the sample model user at the T1 scene feature point, and obtaining a T1 brain wave monitoring index value based on the monitoring result;

[0154] The step S332 further includes the following specific steps:

[0155] The T1 scene feature point is used as an intermediate time node to mark a T1 scene monitoring cycle, and the T1 scene monitoring cycle is divided into multiple monitoring time points, and a feature monitoring time point is arbitrarily selected from the multiple monitoring time points obtained;

[0156] Performing real-time brainwave monitoring on the sample model user at the characteristic monitoring time point, obtaining the alpha wave power corresponding to the sample model user at the characteristic monitoring time point, obtaining a first brainwave index value, obtaining the beta wave power corresponding to the sample model user at the characteristic monitoring time point, obtaining a second brainwave index value, calculating the ratio of the first brainwave index value to the second brainwave index value, and obtaining the brainwave monitoring index value corresponding to the characteristic monitoring time point;

[0157] Acquire the brainwave monitoring index value corresponding to each monitoring time point respectively to obtain multiple brainwave monitoring index values;

[0158] The monitoring time point mark is used as the horizontal axis, and the brain wave monitoring index value is marked as the vertical axis to obtain multiple brain wave monitoring coordinate points, and the obtained multiple brain wave monitoring coordinate points are connected to obtain a brain wave waveform graph;

[0159] Acquire brain wave waveform segments in the same monotonic state in the brain wave waveform graph to obtain monotonically identical waveform segments, and obtain the peak value, trough value, and peak-to-trough interval duration between the peak value and the trough value of each monotonically identical waveform segment;

[0160] The peak value, trough value and peak-to-trough interval duration corresponding to the same monotonous waveform segment are calculated by calculating the waveform change slope;

[0161] Calculate the slope of the waveform change. The specific formula is as follows:

[0162]

[0163] Among them, Bbx is the slope of the waveform change, Bfz is the peak value, Bgz is the trough value, and Jfg is the peak-to-trough interval length;

[0164] Obtaining the waveform change slope corresponding to each monotonically identical waveform segment respectively, and averaging the obtained multiple waveform change slopes to obtain the waveform average change slope;

[0165] Calculating the difference between the peak value and the trough value corresponding to each monotonic identical waveform segment, taking the absolute value of the obtained difference to obtain multiple monotonic waveform amplitudes, and averaging the obtained multiple monotonic waveform amplitudes to obtain the waveform average amplitude;

[0166] Calculate the ratio of the waveform average amplitude to the waveform average change slope to obtain the T1 EEG monitoring index value;

[0167] Step S333: Obtain the brainwave monitoring index values ​​corresponding to the scene feature points T2 to Td of the sample model user, and obtain the brainwave monitoring index values ​​T2 to Ta;

[0168] Step S334: obtaining the brain wave reference index interval corresponding to the T1 scene feature point to the Td scene feature point, and obtaining the T1 brain wave reference index interval to the Td brain wave reference index interval;

[0169] Step S335: If the T1 EEG monitoring index value is within the T1 EEG benchmark index interval, the T1 scene feature point is marked as a normal scene feature point; if the T1 EEG monitoring index value is not within the T1 EEG benchmark index interval, the T1 scene feature point is marked as an abnormal scene feature point;

[0170] Step S336: Similarly, if the Ta brainwave monitoring index value is within the Ta brainwave benchmark index interval, then the Ta scene feature point is marked as a normal scene feature point; if the Ta brainwave monitoring index value is not within the Ta brainwave benchmark index interval, then the Ta scene feature point is marked as an abnormal scene feature point;

[0171] Step S337: Count the number of normal scene feature points and abnormal scene feature points respectively to obtain the number of normal feature points and the number of abnormal feature points. If the number of normal feature points is greater than the number of abnormal feature points, the sample consultation scene is judged to be a normal interaction scene. If the number of normal feature points is less than or equal to the number of abnormal feature points, the sample consultation scene is judged to be an abnormal interaction scene.

[0172] Step S34: performing brainwave monitoring on the sample model users in each psychological counseling scenario, judging the abnormality of scene interaction in the psychological counseling scenario based on the monitoring, and counting the number of abnormal interaction scenarios to obtain the abnormal number value of the interaction scenarios;

[0173] Step S35: Calculate the ratio of the number of interaction scene anomalies to p to obtain the scene interaction anomaly degree corresponding to the sample model user;

[0174] Step S36: Obtain the scene interaction abnormality corresponding to each model user to obtain interaction behavior monitoring data;

[0175] Step S4: providing consultation result feedback to the model user based on the interactive vocabulary analysis data and the interactive behavior monitoring data;

[0176] The step S4 further includes the following specific steps:

[0177] Obtaining interactive vocabulary analysis data, obtaining multiple vocabulary negativity indexes corresponding to model users based on the interactive vocabulary analysis data, and averaging the obtained multiple vocabulary negativity indexes to obtain an average vocabulary negativity index;

[0178] Obtain interactive behavior monitoring data, and obtain the scene interaction abnormality corresponding to the model user based on the interactive behavior monitoring data;

[0179] Obtaining the preset interval of vocabulary negativity index and the preset interval of scene interaction abnormality;

[0180] If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then it is judged that the model user's psychological state is normal;

[0181] If the average vocabulary negativity index is not within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then the model user's psychological state is judged to be abnormal;

[0182] If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is not within the preset scene interaction abnormality range, it is judged that the model user's psychological state is abnormal.

[0183] In this application, if a corresponding calculation formula appears, the above calculation formula is dimensionless and its numerical calculation is performed. The weight coefficient, proportional coefficient and other coefficients in the formula are set to a result value obtained by quantifying each parameter. Regarding the size of the weight coefficient and the proportional coefficient, as long as it does not affect the proportional relationship between the parameter and the result value, it is acceptable.

[0184] Example 2

[0185] See also Figure 2 Based on another concept of the same invention, an artificial intelligence-based psychological counseling system is now proposed, including a question-and-answer status module, an interactive vocabulary module, an interactive behavior module, a consultation feedback module, and a server. The question-and-answer status module, the interactive vocabulary module, the interactive behavior module, and the consultation feedback module are respectively connected to the server, and the server controls the question-and-answer status module, the interactive vocabulary module, the interactive behavior module, and the consultation feedback module respectively;

[0186] The question-answering status module obtains several psychological consultation records, evaluates the question-answering concentration of the model user in each psychological consultation record, and screens the psychological consultations based on the evaluation results to obtain psychological consultation screening data;

[0187] The details are as follows:

[0188] In the process of psychological monitoring of the model user, a psychological consultation history cycle is marked, and the historical question-and-answer interaction data between the model user and the psychological consultation model in the psychological consultation history cycle is obtained. Several different batches of psychological consultation records in the historical question-and-answer interaction data are analyzed, and a sample psychological consultation record is randomly selected from the multiple psychological consultation records obtained;

[0189] It should be noted here that:

[0190] In this application, the psychological counseling model involved here is specifically a psychological counseling artificial intelligence dialogue model, and the psychological counseling artificial intelligence dialogue model involved here can conduct interactive psychological counseling dialogues with users;

[0191] In this application, the duration of a single psychological consultation interaction between the psychological consultation model and each user is 6-8 minutes. The interaction duration corresponding to several different batches of psychological consultation records selected here is greater than 6 minutes.

[0192] The sample psychological consultation involved here is a real-time psychological consultation, that is, the model user and the psychological consultation model are conducting real-time question-and-answer interactions;

[0193] Monitor the physiological indicators of the model users corresponding to the sample psychological consultation records, and obtain the consultation focus time ratio corresponding to the sample psychological consultation based on the monitoring results;

[0194] The details are as follows:

[0195] The question and answer content involved in the sample psychological consultation record is obtained, and the obtained question and answer content is marked as J1 question and answer content to Ja question and answer content according to the interaction time sequence;

[0196] It should be noted here that:

[0197] In this application, J is designed here to be the sign symbol corresponding to the question and answer content, and a is the quantity value corresponding to the question and answer content in the sample psychological counseling record.

[0198] The time point at which the psychological counseling model completes outputting the question part of the J1 question and answer content is obtained to obtain the J1 question time point; the time point at which the model user completes outputting the reply part of the J1 question and answer content is obtained to obtain the J1 reply time point; the interval between the J1 question time point and the J1 reply time point is obtained to obtain the J1 user reply time point;

[0199] Similarly, the time point when the psychological counseling model finishes outputting the question part of the Ja Q&A content is obtained to obtain the Ja question time point, the time point when the model user finishes outputting the reply part of the Ja Q&A content is obtained to obtain the Ja reply time point, and the interval between the Ja question time point and the Ja reply time point is obtained to obtain the Ja user reply time point;

[0200] Physiological indicators of model users interacting with J1 question-and-answer content are monitored, and J1 focus duration is obtained based on the monitoring results;

[0201] The details are as follows:

[0202] Conduct three-dimensional modeling of the spatial environment where the model user and the psychological counseling model output terminal are located to obtain a three-dimensional model of the terminal environment;

[0203] It should be noted here that:

[0204] In this application, the psychological counseling model output terminal involved here is specifically the display of the PC device where the psychological counseling model is deployed.

[0205] In the three-dimensional model of the terminal environment, the display interface corresponding to the output terminal of the psychological counseling model is marked as the first model interaction plane. A pixel point is arbitrarily selected in the first model interaction plane as the coordinate origin, and a straight line is arbitrarily drawn through the coordinate origin to obtain the first model interaction line. In the first model interaction plane, a perpendicular line to the first model interaction line is drawn through the coordinate origin to obtain the second model interaction line. A perpendicular line to the first model interaction plane is drawn through the coordinate origin to obtain the third model interaction line.

[0206] The first model interaction line is marked as the coordinate x-axis, the second model interaction line is marked as the coordinate y-axis, and the third model interaction line is marked as the coordinate z-axis. The spatial coordinate system composed of the coordinate origin, the coordinate x-axis, the coordinate y-axis, and the coordinate z-axis is marked as the scene space coordinate system.

[0207] In the scene space coordinate system, mark the center point of the model user's left pupil as the first pupil marker point, and mark the center point of the model user's right pupil as the second pupil marker point;

[0208] Using the ELG algorithm to obtain the gaze ray starting from the first pupil mark point to obtain the first pupil ray, and using the ELG algorithm to obtain the gaze ray starting from the second pupil mark point to obtain the second pupil ray;

[0209] It should be noted here that:

[0210] In this application, the ELG algorithm: a line of sight estimation method based on eye landmarks can obtain the line of sight of the model user;

[0211] In the present application, the first pupil ray and the second pupil ray involved here both require the model user to face the first model interaction plane, that is, the model user faces the psychological counseling model output terminal.

[0212] Marking a period in which the first pupil ray intersects the first model interaction plane as a first gaze period, acquiring the area of ​​the left pupil of the target user in the first target period to obtain a real-time area value of the left pupil, obtaining a left pupil focus area value, and marking the first target period in which the real-time area value of the left pupil is greater than or equal to the left pupil focus area value as the first pupil focus period;

[0213] Marking the period during which the second pupil ray intersects the second model interaction plane as the second gaze period, acquiring the area of ​​the right pupil of the target user in the second target period to obtain a real-time area value of the right pupil, obtaining a right pupil focus area value, and marking the second target period during which the real-time area value of the right pupil is greater than or equal to the right pupil focus area value as the second pupil focus period;

[0214] The common time period of the first pupil focus period and the second pupil focus period is obtained to obtain the model user focus period, and the duration of the model user focus period is obtained to obtain the J1 focus duration;

[0215] Repeat the process of obtaining the concentration time of j1, and obtain the concentration time corresponding to the question and answer content of J2 to the question and answer content of Ja, and obtain the concentration time of J2 to Ja;

[0216] The consultation focus time ratio corresponding to the sample psychological consultation is obtained by calculating the time between J1's focus time and Ja's focus time, as well as the time between J1's user response time and Ja's user response time.

[0217] The specific formula for calculating the consultation focus duration ratio corresponding to the sample psychological consultation is as follows:

[0218]

[0219] Among them, Zxb is the consultation focus time ratio corresponding to the sample psychological consultation, Zzji is Ji’s focus time, Zcji is Ji’s user response time, and a is the number of question and answer contents in the sample psychological consultation record;

[0220] It should be noted here that:

[0221] When analyzing the consultation focus time ratio, the following experimental data are available:

[0222] If J1's concentration time is 15s, J1 user's reply time is 27s, J2's concentration time is 18s, J2 user's reply time is 34s, J3's concentration time is 19s, and J3 user's reply time is 30s, then the consultation concentration time ratio can be calculated to be 0.57.

[0223] Repeat the process of obtaining the consultation focus time ratio corresponding to the sample psychological consultations, obtain the consultation focus time ratio corresponding to each psychological consultation record, and obtain multiple consultation focus time ratios;

[0224] Obtain a reference interval for the consultation focus time ratio. If the consultation focus time ratio is within the reference interval, mark the corresponding psychological consultation record as a valid psychological consultation record. If the consultation focus time ratio is not within the reference interval, mark the corresponding psychological consultation record as an invalid psychological consultation record.

[0225] It should be noted here that:

[0226] In this application, the valid psychological counseling records involved include situations where the counseling focus time ratio is at the boundary of the counseling focus time ratio benchmark interval;

[0227] In this application, the upper limit of the interval corresponding to the reference interval of the consultation focus duration involved here is 100%, that is, the model user remains focused during the period between the question time point and the reply time point;

[0228] The upper limit of the interval corresponding to the reference interval of the consultation focus time involved is obtained as follows:

[0229] According to historical data, several valid psychological counseling records are obtained, and the counseling focus time ratio corresponding to each valid psychological counseling record is obtained respectively. The average of the obtained multiple counseling focus time ratios is calculated to obtain the upper limit of the interval corresponding to the counseling focus time ratio benchmark interval, and the numerical range between the upper limit of the interval and the upper limit of the interval is marked as the counseling focus time ratio benchmark interval.

[0230] The multiple valid psychological consultation records obtained are defined as psychological consultation screening data;

[0231] The question-answer status module acquires the psychological consultation screening data and transmits it to the interactive vocabulary module and the interactive behavior module;

[0232] The interactive vocabulary module performs a vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtains the vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and obtains the interactive vocabulary analysis data;

[0233] The details are as follows:

[0234] Acquire psychological consultation screening data, acquire each valid psychological consultation record according to the psychological consultation screening data, and arbitrarily select a characteristic psychological consultation record from the multiple valid psychological consultation records acquired;

[0235] Conducting a deep vocabulary analysis on the characteristic psychological counseling records to obtain the vocabulary negativity index corresponding to the characteristic psychological counseling records;

[0236] The details are as follows:

[0237] The output content of the model user in the characteristic psychological consultation record is obtained to obtain the user output text, the negative words in the user output text are obtained using a word recognition algorithm, and the obtained multiple negative words are re-labeled as C1 negative words to Cb negative words;

[0238] It should be noted here that:

[0239] In this application, the output content of the model user involved here can be text content and voice content. If the output content is voice content, the voice content is converted into text content;

[0240] In this application, C is the sign corresponding to the negative word, b is the number value corresponding to the negative word, and b is an integer greater than 0;

[0241] In this application, the negative words involved here are specifically the negative words included in the HowNet sentiment dictionary;

[0242] Use the SO-PMI algorithm to obtain the sentiment intensity corresponding to the negative words C1 to Cb respectively, and obtain the sentiment intensity value of the C1 word to the Cb word;

[0243] It should be noted here that:

[0244] In this application, the SO-PMI algorithm evaluates the emotional intensity of a word by calculating the mutual information between the word and the seed emotional words (such as "happy" and "sad"). The emotional intensity value is specifically divided into 1-5 levels. For example, level 1 may indicate mild negative emotions, such as "a little tired"; level 2 indicates moderate negative emotions, such as "anxiety"; level 3 indicates strong negative emotions, such as "despair"; level 4 indicates strong negative emotions, such as "suicidal tendencies"; level 5 indicates extreme negative emotions, such as "life is worse than death."

[0245] Analyze the frequency of occurrence of C1 negative words in the user's output text, and obtain the C1 word frequency index value based on the analysis results;

[0246] The details are as follows:

[0247] If the C1 negative word only appears once in the user output text, the value "1" will be used to assign the C1 word frequency index value to the parameter;

[0248] If the number of occurrences of the C1 negative word in the user output text is greater than 1, the number of occurrences of the C1 negative word in the user output text is counted to obtain the C1 negative word occurrence quantity value, and the interval length between each two consecutive occurrences of the C1 negative word is obtained to obtain multiple negative word occurrence interval lengths, and the obtained multiple negative word occurrence interval lengths are averaged to obtain the C1 negative word occurrence interval length, and the product of the C1 negative word occurrence quantity value and the C1 negative word occurrence interval length is calculated to obtain the C1 word frequency index value;

[0249] It should be noted here that:

[0250] The time point of appearance of negative words involved here shall be based on the time point of the corresponding negative words output by the model user. The output methods involved here include but are not limited to text output and voice output;

[0251] It should be noted here that:

[0252] If the C1 negative word appears 5 times in the user output text, and the interval between the first and second appearances is 5 seconds, the interval between the second and third appearances is 9 seconds, the interval between the third and fourth appearances is 11 seconds, and the interval between the fourth and fifth appearances is 20 seconds, it can be calculated that the interval between the appearances of the C1 negative word is 13.75 seconds, and the C1 word frequency index value can be calculated to be 68.75.

[0253] Repeat the process of obtaining the C1 word frequency index value, and obtain the word frequency index values ​​corresponding to the C2 negative words to the Cb negative words respectively, to obtain the C2 word frequency index values ​​to the Cb word frequency index values;

[0254] The word negativity index corresponding to the characteristic psychological counseling record is obtained by calculating the word sentiment intensity value of C1 to the word sentiment intensity value of Cb and the word frequency index value of C1 to the word frequency index value of Cb;

[0255] The vocabulary negativity index corresponding to the characteristic psychological counseling records is calculated using the following formula:

[0256]

[0257] Among them, Cxz is the vocabulary negativity index corresponding to the characteristic psychological counseling record, Qqdi is the Ci vocabulary emotional intensity value, Cpzi is the Ci word frequency index value, and b is the number value corresponding to the negative vocabulary;

[0258] It should be noted here that:

[0259] In this application, the Ci vocabulary sentiment intensity value involved here can be any vocabulary sentiment intensity value from the C1 vocabulary sentiment intensity value to the Cb vocabulary sentiment intensity value, and the Ci word frequency index value involved here can be any word frequency index value from the C1 word frequency index value to the Cb word frequency index value;

[0260] In the specific implementation, there are the following experimental data:

[0261] When the C1 lexical sentiment intensity value is 3, the C1 word frequency index value is 1, the C2 lexical sentiment intensity value is 2, the C2 word frequency index value is 65.8, the C3 lexical sentiment intensity value is 5, the C3 word frequency index value is 25.8, and b = 3, the lexical negativity index can be calculated to be 87.87;

[0262] In this application, the vocabulary negativity index involved here is specifically an indicator value that represents the negativity of the interactive content output by the model user during psychological counseling, specifically calculating the average value of the word frequency index value and the word sentiment intensity value corresponding to each negative word.

[0263] Repeat the process of obtaining the vocabulary negativity index corresponding to the characteristic psychological counseling record, obtain the vocabulary negativity index corresponding to each valid psychological counseling record, and obtain the interactive vocabulary analysis data;

[0264] The interactive vocabulary module acquires the interactive vocabulary analysis data and transmits it to the consultation feedback module;

[0265] The interactive behavior module conducts scenario-based interactive psychological consultations on model users corresponding to valid psychological consultation records based on the psychological consultation screening data, and monitors the model users' brain waves during the consultation process. Based on the monitoring results, the scenario interaction abnormality corresponding to each model user is obtained to obtain interactive behavior monitoring data;

[0266] The details are as follows:

[0267] Obtain psychological consultation screening data, obtain model users corresponding to valid psychological consultation records based on the psychological consultation screening data, and select a sample model user from the multiple model patients obtained;

[0268] When conducting interactive psychological consultation for the model user, p different types of psychological consultation scenarios are set, and a sample consultation scenario is randomly selected from the set psychological consultation scenarios;

[0269] It should be noted here that:

[0270] p is the quantity value corresponding to the psychological counseling scenario, and p is an integer greater than 0;

[0271] In this application, the psychological counseling scenario set here is specifically an exposure therapy scenario. The psychological counseling scenarios involved here include but are not limited to "conversing with a stranger for two minutes continuously", "looking at yourself in the mirror for one minute" and "participating in a three-person meeting discussion for three minutes".

[0272] Conduct brainwave monitoring on sample model users in sample consultation scenarios, and determine abnormalities in scene interaction in the sample consultation scenarios based on the monitoring;

[0273] The details are as follows:

[0274] Mark several scene feature points in the sample consultation scene, and mark the scene feature points as T1 scene feature points to Td scene feature points in chronological order;

[0275] It should be noted here that:

[0276] In this application, T referred to herein is the sign symbol corresponding to the scene feature point, d referred to herein is the number value corresponding to the scene feature point, and d is an integer greater than 0;

[0277] In this application, if the sample consultation scenario is a two-minute continuous conversation with a stranger, the scene feature point is the time point when the sample model user starts speaking;

[0278] It is important to note that:

[0279] For different model users, the interaction scenarios corresponding to the T1 scene feature points to the Td scene feature points in the psychological counseling scenario are the same. For example, if the T1 scene feature point is the first time the model user speaks.

[0280] Perform real-time brainwave monitoring on the sample model user at the T1 scene feature point, and obtain the T1 brainwave monitoring index value based on the monitoring results;

[0281] The details are as follows:

[0282] The T1 scene feature point is used as an intermediate time node to mark a T1 scene monitoring cycle, and the T1 scene monitoring cycle is divided into multiple monitoring time points, and a feature monitoring time point is arbitrarily selected from the multiple monitoring time points obtained;

[0283] Performing real-time brainwave monitoring on the sample model user at the characteristic monitoring time point, obtaining the alpha wave power corresponding to the sample model user at the characteristic monitoring time point, obtaining a first brainwave index value, obtaining the beta wave power corresponding to the sample model user at the characteristic monitoring time point, obtaining a second brainwave index value, calculating the ratio of the first brainwave index value to the second brainwave index value, and obtaining the brainwave monitoring index value corresponding to the characteristic monitoring time point;

[0284] It should be noted here that:

[0285] In this application, alpha wave power: In brain wave monitoring, the power of alpha waves (8-13 Hz) reflects the activity intensity of the brain in a wakeful and relaxed state. The greater the alpha wave power, the more relaxed the brain is.

[0286] Beta wave power: The power of beta waves (above 14Hz) is related to the brain's excitatory activities. The greater the beta wave power, the more excited and alert the brain is. At this time, the person is in a normal wakeful state and the thinking activities are active. The alpha wave power and beta wave power in different scenarios can reflect the human body's adaptability to the surrounding environment and the psychological state in the environment, which can improve the authenticity of the psychological counseling process.

[0287] Acquire the brainwave monitoring index value corresponding to each monitoring time point respectively to obtain multiple brainwave monitoring index values;

[0288] See also Figure 3 , mark the monitoring time point as the horizontal axis, mark the brain wave monitoring index value as the vertical axis, obtain multiple brain wave monitoring coordinate points, and connect the obtained multiple brain wave monitoring coordinate points to obtain a brain wave waveform diagram;

[0289] Acquire brain wave waveform segments in the same monotonic state in the brain wave waveform graph to obtain monotonically identical waveform segments, and obtain the peak value, trough value, and peak-to-trough interval duration between the peak value and the trough value of each monotonically identical waveform segment;

[0290] The peak value, trough value and peak-to-trough interval duration corresponding to the same monotonous waveform segment are calculated by calculating the waveform change slope;

[0291] Calculate the slope of the waveform change. The specific formula is as follows:

[0292]

[0293] Among them, Bbx is the slope of the waveform change, Bfz is the peak value, Bgz is the trough value, and Jfg is the peak-to-trough interval length;

[0294] Obtaining the waveform change slope corresponding to each monotonically identical waveform segment respectively, and averaging the obtained multiple waveform change slopes to obtain the waveform average change slope;

[0295] Calculating the difference between the peak value and the trough value corresponding to each monotonic identical waveform segment, taking the absolute value of the obtained difference to obtain multiple monotonic waveform amplitudes, and averaging the obtained multiple monotonic waveform amplitudes to obtain the waveform average amplitude;

[0296] Calculate the ratio of the waveform average amplitude to the waveform average change slope to obtain the T1 EEG monitoring index value;

[0297] It should be noted here that:

[0298] In this application, the T1 EEG monitoring index value involved here represents the average amplitude output efficiency under unit slope change, reflecting the energy concentration degree in the dynamic change of the waveform;

[0299] The EEG monitoring index value provides a standardized tool for evaluating the user's psychological state by integrating the dynamic characteristics of the waveform and energy parameters. This indicator conforms to the neuroelectrophysiological mechanism and has clinical translation value. In actual application, it needs to be combined with comprehensive judgment such as age and lead position.

[0300] Repeat the process of obtaining the T1 brainwave monitoring index value, and obtain the brainwave monitoring index values ​​corresponding to the sample model user at the T2 scene feature point to the Td scene feature point, and obtain the T2 brainwave monitoring index value to the Ta brainwave monitoring index value;

[0301] Obtain the brain wave benchmark index interval corresponding to the T1 scene feature point to the Td scene feature point, and obtain the T1 brain wave benchmark index interval to the Td brain wave benchmark index interval;

[0302] It should be noted here that:

[0303] The T1 brain wave benchmark index interval to the Td brain wave benchmark index interval are obtained as follows:

[0304] Obtain historical consultation records corresponding to the sample consultation scenario, select multiple historical users without psychological problems from the historical consultation records, obtain the T1 brainwave monitoring index value corresponding to the T1 scene feature point of each historical user, obtain multiple T1 brainwave monitoring index values, average the obtained multiple T1 brainwave monitoring index values ​​to obtain the T1 average brainwave monitoring index value, calculate the standard deviation of the obtained multiple T1 brainwave monitoring index values ​​to obtain the T1 brainwave monitoring standard deviation, calculate the difference between the T1 average brainwave monitoring index value and the T1 brainwave monitoring standard deviation to obtain the lower limit of the T1 brainwave benchmark index interval, calculate the sum of the T1 average brainwave monitoring index value and the T1 brainwave monitoring standard deviation to obtain the upper limit of the T1 brainwave benchmark index interval, and define the value between the lower limit of the T1 brainwave benchmark index interval and the upper limit of the T1 brainwave benchmark index interval as the T1 brainwave benchmark index interval;

[0305] Among them, the lower limit of the T1 brain wave benchmark index interval is less than the upper limit of the T1 brain wave benchmark index interval;

[0306] Repeat the process of obtaining the T1 brain wave benchmark index interval to obtain the T2 brain wave benchmark index interval to the Td brain wave benchmark index interval respectively.

[0307] If the T1 EEG monitoring index value is within the T1 EEG benchmark index interval, the T1 scene feature point is marked as a normal scene feature point; if the T1 EEG monitoring index value is not within the T1 EEG benchmark index interval, the T1 scene feature point is marked as an abnormal scene feature point;

[0308] Similarly, if the Ta brainwave monitoring index value is within the Ta brainwave benchmark index interval, then the Ta scene feature point is marked as a normal scene feature point; if the Ta brainwave monitoring index value is not within the Ta brainwave benchmark index interval, then the Ta scene feature point is marked as an abnormal scene feature point;

[0309] It should be noted here that:

[0310] In this application, the normal scene feature points include the boundaries of the brain wave benchmark index interval.

[0311] Count the number of normal scene feature points and abnormal scene feature points respectively to obtain the number of normal feature points and the number of abnormal feature points. If the number of normal feature points is greater than the number of abnormal feature points, the sample consultation scene is judged to be a normal interaction scene. If the number of normal feature points is less than or equal to the number of abnormal feature points, the sample consultation scene is judged to be an abnormal interaction scene.

[0312] The brainwaves of sample model users in each psychological counseling scenario are monitored separately. Based on the monitoring, the abnormality of scene interaction in the psychological counseling scenario is judged. The number of abnormal interaction scenes is counted to obtain the abnormal number value of the interaction scene.

[0313] Calculate the ratio of the number of interaction scene anomalies to p to obtain the scene interaction anomaly degree corresponding to the sample model user;

[0314] Repeat the process of obtaining the scene interaction abnormality corresponding to the sample model user, obtain the scene interaction abnormality corresponding to each model user, and obtain the interaction behavior monitoring data;

[0315] The consultation feedback module provides consultation result feedback to model users based on interactive vocabulary analysis data and interactive behavior monitoring data;

[0316] The details are as follows:

[0317] Obtaining interactive vocabulary analysis data, obtaining multiple vocabulary negativity indexes corresponding to model users based on the interactive vocabulary analysis data, and averaging the obtained multiple vocabulary negativity indexes to obtain an average vocabulary negativity index;

[0318] Obtain interactive behavior monitoring data, and obtain the scene interaction abnormality corresponding to the model user based on the interactive behavior monitoring data;

[0319] Obtaining the preset interval of vocabulary negativity index and the preset interval of scene interaction abnormality;

[0320] It should be noted here that:

[0321] Get the preset interval of the vocabulary negativity index as follows:

[0322] Select several historical model patients with normal mental states, obtain the average vocabulary negativity index corresponding to each historical model patient, and calculate the average to obtain the average vocabulary negativity index, calculate the standard deviation of the obtained multiple average vocabulary negativity indices to obtain the standard deviation of the vocabulary negativity index, calculate the difference between the average vocabulary negativity index and the standard deviation of the vocabulary negativity index to obtain the upper limit of the preset interval of the vocabulary negativity index, if the model user does not output negative words, the average vocabulary negativity index is 0, so the lower limit of the preset interval of the vocabulary negativity index is 0, and the value between the upper limit of the preset interval of the vocabulary negativity index and the upper limit of the preset interval of the vocabulary negativity index is defined as the preset interval of the vocabulary negativity index;

[0323] Get the preset interval of scene interaction abnormality, as follows:

[0324] Select several historical model patients with normal mental states, obtain the scene interaction abnormality corresponding to each historical model patient, and calculate the average to obtain the scene interaction abnormality average, calculate the standard deviation of the obtained multiple scene interaction abnormalities to obtain the scene interaction abnormality standard deviation, calculate the difference between the scene interaction abnormality average and the scene interaction abnormality standard deviation to obtain the upper limit of the scene interaction abnormality preset interval. If the model user does not have an abnormal interaction scene, the scene interaction abnormality is 0, so the lower limit of the scene interaction abnormality is 0, and the value between the upper limit of the vocabulary negativity index preset interval and the upper limit of the vocabulary negativity index preset interval is defined as the vocabulary negativity index preset interval;

[0325] If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then it is judged that the model user's psychological state is normal;

[0326] If the average vocabulary negativity index is not within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then the model user's psychological state is judged to be abnormal;

[0327] If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is not within the preset scene interaction abnormality range, it is judged that the model user's psychological state is abnormal.

[0328] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A psychological consultation method based on artificial intelligence, characterized in that: include: Step S1: Obtain several psychological consultation records, conduct question-answering concentration assessment on the model user in each psychological consultation record, screen the psychological consultations based on the assessment results, and obtain psychological consultation screening data; Step S2: performing a vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtaining a vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and obtaining interactive vocabulary analysis data; Step S3: Based on the psychological consultation screening data, scenario-based interactive psychological consultation is conducted on the model users corresponding to the valid psychological consultation records, and brainwave monitoring is performed on the model users during the consultation process. Based on the monitoring results, the scenario interaction abnormality corresponding to each model user is obtained to obtain interactive behavior monitoring data; Step S4: Provide consultation result feedback to the model user based on the interactive vocabulary analysis data and interactive behavior monitoring data.

2. The artificial intelligence-based psychological consultation method according to claim 1, characterized in that: The step S1 further includes the following specific steps: Step S11: obtaining a plurality of different batches of psychological consultation records, and arbitrarily selecting a sample psychological consultation record from the obtained plurality of psychological consultation records; Step S12: monitoring physiological indicators of the model user corresponding to the sample psychological consultation record, and obtaining the consultation focus time ratio corresponding to the sample psychological consultation based on the monitoring results; Step S13: Obtain the consultation focus time ratio corresponding to each psychological consultation record to obtain multiple consultation focus time ratios; Step S14: Obtaining a reference interval for the consultation focus time ratio. If the consultation focus time ratio is within the reference interval, marking the corresponding psychological consultation record as a valid psychological consultation record; if the consultation focus time ratio is not within the reference interval, marking the corresponding psychological consultation record as an invalid psychological consultation record. Step S15: defining the obtained multiple valid psychological counseling records as psychological counseling screening data.

3. The artificial intelligence-based psychological consultation method according to claim 2, characterized in that: The step S12 further includes the following specific steps: The question and answer content involved in the sample psychological consultation record is obtained, and the obtained question and answer content is marked as J1 question and answer content to Ja question and answer content according to the interaction time sequence; Get the user response time from J1 Q&A content to Ja Q&A content respectively, and get the response time from J1 user to Ja user; Physiological indicators of model users interacting with J1 question-and-answer content are monitored, and J1 focus duration is obtained based on the monitoring results; The corresponding consultation focus time ratio Zxb of the sample psychological consultation is obtained by calculating the J1 focus time Zzj1 to Ja focus time Zzja and the J1 user reply time Zcjq to Ja user reply time Zcja. The specific formula is as follows:

4. The artificial intelligence-based psychological consultation method according to claim 3, characterized in that: The step S12 further includes the following specific steps: Conduct three-dimensional modeling of the spatial environment where the model user and the psychological counseling model output terminal are located to obtain a three-dimensional model of the terminal environment; Create a coordinate system in the terminal environment 3D model to obtain the scene space coordinate system; In the scene space coordinate system, mark the center point of the model user's left pupil as the first pupil marker point, and mark the center point of the model user's right pupil as the second pupil marker point; Acquire a gaze ray starting from the first pupil marking point to obtain a first pupil ray, and acquire a gaze ray starting from the second pupil marking point to obtain a second pupil ray; Marking a period in which the first pupil ray intersects the first model interaction plane as a first gaze period, acquiring the area of ​​the left pupil of the target user in the first target period to obtain a real-time area value of the left pupil, obtaining a left pupil focus area value, and marking the first target period in which the real-time area value of the left pupil is greater than or equal to the left pupil focus area value as the first pupil focus period; Performing focus analysis on the second pupil ray to obtain the second pupil focus period; The common time period of the first pupil focus period and the second pupil focus period is obtained to obtain the model user focus period, and the duration of the model user focus period is obtained to obtain the J1 focus duration; The concentration duration corresponding to the J2 question and answer content to the Ja question and answer content is obtained respectively, and the J2 concentration duration to the Ja concentration duration is obtained.

5. The artificial intelligence-based psychological consultation method according to claim 1, characterized in that: The step S2 further includes the following specific steps: Step S21: Acquire psychological consultation screening data, acquire each valid psychological consultation record according to the psychological consultation screening data, and arbitrarily select a characteristic psychological consultation record from the multiple valid psychological consultation records acquired; Step S22: performing a lexical depth analysis on the characteristic psychological counseling record to obtain a lexical negativity index corresponding to the characteristic psychological counseling record; Step S23: Obtain the vocabulary negativity index corresponding to each valid psychological counseling record to obtain interactive vocabulary analysis data.

6. The artificial intelligence-based psychological consultation method according to claim 5, characterized in that: The step S22 further includes the following specific steps: Step S221: acquiring the user output text in the characteristic psychological consultation record, acquiring negative words in the user output text, and re-labeling the acquired multiple negative words as C1 negative words to Cb negative words; Step S222: Obtain the sentiment intensity corresponding to the negative words C1 to Cb, and obtain the sentiment intensity value of the words C1 to Cb; Step S223: Analyze the frequency of occurrence of C1 negative words in the user output text, and obtain the C1 word frequency index value according to the analysis result; Step S224: Acquire the word frequency index values ​​corresponding to the negative words C2 to Cb to obtain the word frequency index values ​​C2 to Cb; Step S225: The C1 vocabulary emotion intensity value Qqd1 to the Cb vocabulary emotion intensity value Qqda and the C1 word frequency index value Cpz1 to the Cb word frequency index value Cpza are calculated to obtain the vocabulary negativity index Cxz corresponding to the characteristic psychological counseling record. The formula is as follows:

7. The artificial intelligence-based psychological consultation method according to claim 1, characterized in that: The step S3 further includes the following specific steps: Step S31: Acquire model users corresponding to valid psychological consultation records, and select a sample model user from the acquired multiple model patients; Step S32: setting p different types of psychological consultation scenarios and selecting a sample consultation scenario; Step S33: performing brain wave monitoring on the sample model user in the sample consultation scenario, and determining abnormality of scene interaction in the sample consultation scenario; Step S34: performing brainwave monitoring on the sample model users in each psychological counseling scenario, determining the abnormality of scene interaction in the psychological counseling scenario, and counting the number of abnormal interaction scenarios to obtain the abnormal number value of the interaction scenarios; Step S35: Calculate the ratio of the number of interaction scene anomalies to p to obtain the scene interaction anomaly degree corresponding to the sample model user. Obtain the scene interaction anomaly degree for each model user to obtain interaction behavior monitoring data. The step S33 further includes the following specific steps: Step S331: Mark scene feature points T1 to Td in the sample consultation scene; Step S332: performing real-time brainwave monitoring on the sample model user at the T1 scene feature point to obtain a T1 brainwave monitoring index value; Step S333: obtaining T2 brain wave monitoring index values ​​to Ta brain wave monitoring index values; Step S334: obtaining the brain wave reference index interval corresponding to the T1 scene feature point to the Td scene feature point, and obtaining the T1 brain wave reference index interval to the Td brain wave reference index interval; Step S335: If the T1 EEG monitoring index value is within the T1 EEG benchmark index interval, the T1 scene feature point is marked as a normal scene feature point; if it is not within the T1 EEG benchmark index interval, it is marked as an abnormal scene feature point; Step S336: Similarly, mark Ta scene feature points; Step S337: Count the number of marked scene feature points to obtain the number of normal feature points and the number of abnormal feature points. If the number of normal feature points is greater than the number of abnormal feature points, the sample consultation scene is judged to be a normal interaction scene; otherwise, the sample consultation scene is judged to be an abnormal interaction scene.

8. The artificial intelligence-based psychological consultation method according to claim 7, characterized in that: The step S332 further includes the following specific steps: The T1 scene feature point is used as the intermediate time node to mark a T1 scene monitoring cycle, which is divided into multiple monitoring time points. A feature monitoring time point is arbitrarily selected; Perform real-time brainwave monitoring on the sample model user at the characteristic monitoring time point, obtain the alpha wave power and beta wave power corresponding to the sample model user at the characteristic monitoring time point, calculate the ratio of the alpha wave power to the beta wave power, and obtain the brainwave monitoring index value corresponding to the characteristic monitoring time point; Acquiring the brainwave monitoring index value to obtain the brainwave monitoring index value; The monitoring time point mark is used as the horizontal axis, and the brain wave monitoring index value is marked as the vertical axis to obtain multiple brain wave monitoring coordinate points, and the obtained multiple brain wave monitoring coordinate points are connected to obtain a brain wave waveform graph; Acquire brain wave waveform segments in the same monotonic state in the brain wave waveform graph to obtain multiple monotonic identical waveform segments, and obtain the peak value, trough value, and peak-to-trough interval duration between the peak value and the trough value of each monotonic identical waveform segment; The peak value Bfz, trough value Bgz and peak-to-trough interval time Jfg corresponding to the same monotonous waveform segment are calculated by calculating the waveform change slope Bbx; Obtain the waveform change slope corresponding to each monotonically identical waveform segment respectively, and calculate the average to obtain the waveform average change slope; Calculating the difference between the peak value and the trough value corresponding to each monotonic identical waveform segment, taking the absolute value of the obtained difference to obtain multiple monotonic waveform amplitudes, and averaging the obtained multiple monotonic waveform amplitudes to obtain the waveform average amplitude; The ratio of the average waveform amplitude to the average waveform change slope is calculated to obtain the T1 EEG monitoring index value.

9. The artificial intelligence-based psychological consultation method according to claim 1, characterized in that: The step S4 further includes the following specific steps: Obtaining interactive vocabulary analysis data, obtaining multiple vocabulary negativity indexes corresponding to model users based on the interactive vocabulary analysis data, and averaging the obtained multiple vocabulary negativity indexes to obtain an average vocabulary negativity index; Obtain interactive behavior monitoring data, and obtain the scene interaction abnormality corresponding to the model user based on the interactive behavior monitoring data; Obtaining the preset interval of vocabulary negativity index and the preset interval of scene interaction abnormality; If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then it is judged that the model user's psychological state is normal; If the average vocabulary negativity index is not within the preset vocabulary negativity index range, and the scene interaction abnormality is within the preset scene interaction abnormality range, then the model user's psychological state is judged to be abnormal; If the average vocabulary negativity index is within the preset vocabulary negativity index range, and the scene interaction abnormality is not within the preset scene interaction abnormality range, it is judged that the model user's psychological state is abnormal.

10. An artificial intelligence-based psychological consultation system, applicable to an artificial intelligence-based psychological consultation method according to any one of claims 1 to 9, characterized in that: The psychological consultation system includes: Question and answer status module: obtains several psychological consultation records, evaluates the question and answer concentration of the model user in each psychological consultation record, and screens the psychological consultations based on the evaluation results to obtain psychological consultation screening data; Interactive vocabulary module: Perform vocabulary negative feature analysis on each valid psychological counseling record in the psychological counseling screening data, obtain the vocabulary negativity index corresponding to each valid psychological counseling record based on the analysis results, and obtain interactive vocabulary analysis data; Interactive behavior module: Based on the psychological consultation screening data, scenario-based interactive psychological consultation is conducted on the model users corresponding to the valid psychological consultation records. The model users' brain waves are monitored during the consultation process. Based on the monitoring results, the scenario interaction abnormality corresponding to each model user is obtained to obtain the interactive behavior monitoring data; Consultation feedback module: Provides consultation result feedback to model users based on interactive vocabulary analysis data and interactive behavior monitoring data.

Images