A data analysis method and system based on Bayesian network
Through Bayesian network analysis method, the problem of excessive user answering questions and incomplete time points in psychological data analysis is solved, and the accurate analysis of user psychological changes and identification of question-asking process is achieved, providing a clear causal relationship Represents the framework.
Patent Information
- Application Number
- CN202411225544.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-03
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-09-03
AI Technical Summary
In the prior art, in psychological data analysis, users need to answer a large number of questions, which leads to users being unwilling to answer and being unable to accurately reflect psychological changes, especially in the analysis of time points, which cannot fully capture the user's psychological situation.
Using Bayesian network-based method, by obtaining psychological data at multiple time points, constructing psychological images, forming a collection of time psychological images, extracting psychological characteristics and clustering, and combining with classifiers to identify psychological analysis types, realizing the user's Accurate analysis of psychological changes.
It provides a clear causal representation framework, which can better understand the psychological changes of users, find the psychological changes of users through cluster analysis, and realizes accurate analysis of users' psychological problems and identify the process of questioning.
Smart Images

Figure CN118736352B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer technology, and in particular to a data analysis method and system based on a Bayesian network. Background Art
[0002] At present, when analyzing psychological data, users are often required to answer all the psychological questions in the entire question bank, so that the psychological analysis of the user can be performed based on these questions. However, if a clear user is required for psychological analysis, too many psychological questions will cause the user to not want to answer them later, and the number of stored questions is large. In addition, the user's psychology may change over time, and only analyzing the user at a time point cannot accurately analyze the user's psychological situation. Summary of the invention
[0003] The purpose of the present invention is to provide a psychological data analysis method and system based on Bayesian network to solve the above problems existing in the prior art.
[0004] In a first aspect, an embodiment of the present invention provides a psychological data analysis method based on a Bayesian network, comprising:
[0005] Obtaining psychological data at multiple time points; the psychological data represents the types of answers selected by the user when asking questions for psychological consultation;
[0006] According to the psychological data, a psychological image is obtained through a Bayesian network; a plurality of psychological images are obtained corresponding to a plurality of time points; the psychological image is an image that uses pixel values to represent the relationship between the psychological data;
[0007] The three mental images of consecutive time points constitute a time mental image set; n mental images correspond to n-2 time mental image sets;
[0008] According to the set of time psychological images, psychological characteristics are obtained through a time psychological change network; n-2 time psychological images correspond to obtaining n-2 psychological characteristics;
[0009] Clustering is performed based on n-2 psychological characteristics to obtain cluster center characteristics;
[0010] According to the cluster center characteristics, the psychological analysis type is obtained through a classifier.
[0011] Optionally, obtaining psychological characteristics according to the temporal psychological image set through a temporal psychological change network includes:
[0012] Judging psychological changes based on two psychological images in the temporal psychological image set to obtain two-dimensional change features;
[0013] Construct a three-dimensional time mental image by arranging the mental images in the time mental image set in the order from far to near in time;
[0014] According to the time-psychology three-dimensional image, a three-dimensional change feature is obtained through a time-psychology change network;
[0015] The three-dimensional change feature and the two-dimensional change feature are averaged to obtain the psychological feature.
[0016] Optionally, judging the psychological change based on the two psychological images in the temporal psychological image set to obtain the two-dimensional change feature includes:
[0017] A first mental image and a third mental image are obtained; the first mental image is an image in the time mental image set whose corresponding time point is less than the time points corresponding to other mental images; the third mental image is an image in the time mental image set whose corresponding time point is greater than the time points corresponding to other mental images;
[0018] Inputting the first psychological map into a psychological judgment network to obtain a first psychological judgment value and a first psychological judgment feature;
[0019] Inputting the third psychological map into a psychological judgment network to obtain a third psychological judgment value and a third psychological judgment feature;
[0020] The third psychological judgment feature and the first psychological judgment feature are input into a two-dimensional change network to obtain a two-dimensional change feature.
[0021] Optionally, the temporal psychological three-dimensional image is convolved in sequence in both the width and height directions of the image with a three-dimensional convolution kernel and a step number of 1 through a temporal psychological change network to obtain a three-dimensional change feature.
[0022] Optionally, obtaining a psychological image through a Bayesian network based on the psychological data includes:
[0023] Inputting the psychological data into the Bayesian network to obtain a psychological question output; the psychological question output is the psychological question to be asked next time corresponding to the current answer;
[0024] According to the output of psychological questions, a probability distribution table is obtained through Bayesian estimation;
[0025] According to the probability distribution table, a probabilistic psychological problem is obtained; the probabilistic psychological problem indicates a psychological problem with a probability greater than that of other psychological problems to be selected;
[0026] Obtaining the number of psychological questions asked; the number of psychological questions asked represents the total number of psychological questions asked from the initial psychological question to the current answer;
[0027] A binary image is constructed with the number of psychological problems as width and height;
[0028] The position corresponding to the probabilistic psychological question is set to 1, and the values of other positions are set to 0 to obtain a psychological image; the value of 1 in the psychological image represents the next psychological question that is the only one in the number of psychological questions asked.
[0029] Optionally, the Bayesian network construction method includes:
[0030] Obtain training psychological data and annotated psychological question data; the training psychological data represents answers to multiple psychological questions answered by multiple historical users; the annotated psychological question data is a question to be selected by the user that better reflects the type of psychological analysis;
[0031] Inputting the training psychological data into the Bayesian network to obtain the training psychological question probability; the psychological question probability indicates the probability that answering the previous psychological question can identify the next psychological question;
[0032] According to the training psychological problem probability and the labeled psychological problem data, the Bayesian network is backward trained through the loss function to obtain a trained Bayesian network.
[0033] Optionally, obtaining the psychological analysis type through a classifier according to the cluster center feature includes:
[0034] Inputting the cluster center feature into a fully connected neural network, extracting the features of the cluster center feature, and obtaining a cluster analysis feature;
[0035] The cluster analysis features are input into a classifier for classification to obtain a psychological analysis type.
[0036] Optional training methods for psychological judgment networks include:
[0037] Obtaining a psychological judgment training set; the psychological judgment training set includes a plurality of training psychological images and a plurality of corresponding annotated psychological analysis types;
[0038] Inputting the training psychological image into a psychological judgment network to obtain an output psychological analysis type;
[0039] The output psychological analysis type and the annotated psychological analysis type are calculated and back-propagated to obtain a trained psychological judgment network.
[0040] Optional training methods for two-dimensional transform networks include:
[0041] Obtaining a two-dimensional variation training set; the training set includes a plurality of training psychological judgment features and a plurality of corresponding annotated psychological analysis types;
[0042] Inputting the psychological judgment feature into a two-dimensional variation network to obtain an output psychological analysis type;
[0043] The output psychological analysis type and the labeled psychological analysis type are calculated and propagated back through the loss to obtain a trained two-dimensional variation network.
[0044] In a second aspect, an embodiment of the present invention provides a psychological data analysis system based on a Bayesian network, comprising:
[0045] Acquisition module: obtains psychological data at multiple time points; the psychological data represents the types of answers selected by the user when asking questions for psychological consultation;
[0046] Bayesian module: Based on the psychological data, a psychological image is obtained through a Bayesian network; multiple psychological images are obtained corresponding to multiple time points; the psychological image is an image that uses pixel values to represent the relationship between psychological data;
[0047] Mental image module: three mental images at consecutive time points constitute a temporal mental image set; n mental images correspond to n-2 temporal mental image sets;
[0048] Psychological change detection module: according to the temporal psychological image set, psychological characteristics are obtained through the temporal psychological change network; n-2 temporal psychological images correspond to n-2 psychological characteristics;
[0049] Clustering module: clustering based on n-2 psychological characteristics to obtain cluster center features;
[0050] Classification module: According to the cluster center characteristics, the psychological analysis type is obtained through a classifier.
[0051] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects:
[0052] The embodiment of the present invention also provides a psychological data analysis method and system based on a Bayesian network, the method comprising: obtaining psychological data at multiple time points. The psychological data represents the type of answer selected by the user for psychological consultation. Based on the psychological data, a psychological image is obtained through a Bayesian network. Multiple psychological images are obtained corresponding to multiple time points. The psychological image is an image that uses pixel values to represent the relationship between psychological data. Three psychological images at consecutive time points constitute a time psychological image set. n psychological images correspond to n-2 time psychological image sets. Based on the time psychological image set, psychological features are obtained through a time psychological change network. n-2 time psychological images correspond to n-2 psychological features. Clustering is performed based on the n-2 psychological features to obtain cluster center features. Based on the cluster center features, a psychological analysis type is obtained through a classifier.
[0053] In the present invention, a Bayesian network is used to provide a convenient framework structure for people to express causal relationships, and different users can choose different questions based on different answers they have chosen at previous time points, so that uncertainty reasoning becomes more logically clear and understandable. In addition, the feature changes of two images at separate continuous time points and the change features of three images at continuous time points are used to find the user's psychological changes through clustering, so as to perform a temporal analysis of the user's psychological problems. It is possible to better obtain the questioning process of different questions and the corresponding difficult questions based on the question selection at each step. Thus, the technical effect of accurately analyzing the user's psychology is better achieved. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 It is a flow chart of a psychological data analysis method based on Bayesian network provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0055] The present invention will be described in detail below in conjunction with the accompanying drawings.
[0056] Example 1
[0057] like Figure 1 As shown, an embodiment of the present invention provides a psychological data analysis method based on a Bayesian network, the method comprising:
[0058] S101: Obtaining psychological data at multiple time points. The psychological data represents the types of answers selected by the user when asking questions for psychological consultation.
[0059] The total number of questions and answer types is fixed, and the next question is different depending on the answer type. For example, question 1 is "Who do you usually communicate with at a party?" A. "With many people, including strangers" B. "Only with some friends." If the answer to question 1 is A, question 2 is asked, "When making a choice, your general choice is:" A. "Very careful" B. "A little impulsive." If the answer to question 1 is B, question 3 is asked, "At work, you hope your tasks are:" A. "Scheduled in advance" B. + "Not scheduled."
[0060] Among them, the question bank between the psychological data at multiple time points is the same.
[0061] S102: Based on the psychological data, a psychological image is obtained through a Bayesian network. Multiple psychological images are obtained corresponding to multiple time points. The psychological image is an image that uses pixel values to represent the relationship between psychological data.
[0062] The mental image is transformed into an image with a matrix having a fixed meaning, and different mental images are represented by different pixel values corresponding to positions.
[0063] S103: The three mental images at consecutive time points constitute a temporal mental image set. n mental images correspond to n-2 temporal mental image sets.
[0064] Among them, for example, the psychological data obtained on the 1st day, the 2nd day, the 3rd day and the 4th day constitute 4 psychological images, the 3 psychological images on the 1st day, the 2nd day and the 3rd day constitute a time psychological image set, and the 3 psychological images on the 2nd day, the 3rd day and the 4th day constitute a time psychological image set.
[0065] S104: Obtain psychological characteristics according to the temporal psychological image set through a temporal psychological change network. n-2 temporal psychological image sets correspond to obtaining n-2 psychological characteristics.
[0066] S105: Clustering is performed according to the n-2 psychological characteristics to obtain cluster center characteristics.
[0067] In this embodiment, clustering is performed using the k-means clustering algorithm.
[0068] S106: Obtaining a psychological analysis type through a classifier according to the cluster center features.
[0069] In this embodiment, the classifier uses a softmax classifier for classification.
[0070] Optionally, obtaining psychological characteristics according to the temporal psychological image set through a temporal psychological change network includes:
[0071] According to the two psychological images in the temporal psychological image set, the psychological change is judged to obtain a two-dimensional change feature.
[0072] Construct a three-dimensional time psychological image by arranging the mental images in the time psychological image set in the order from far to near in time.
[0073] According to the time-psychological three-dimensional image, a three-dimensional change feature is obtained through a time-psychological change network.
[0074] The three-dimensional change feature and the two-dimensional change feature are averaged to obtain the psychological feature.
[0075] In this embodiment, the data corresponding to the subscripts of the three-dimensional change feature and the two-dimensional change feature are averaged to form the psychological feature.
[0076] Optionally, judging the psychological change based on the two psychological images in the temporal psychological image set to obtain the two-dimensional change feature includes:
[0077] A first mental image and a third mental image are obtained. The first mental image is an image in the time mental image set whose corresponding time point is less than the time points corresponding to other mental images. The third mental image is an image in the time mental image set whose corresponding time point is greater than the time points corresponding to other mental images.
[0078] Among them, for example, three mental images on the first day, the second day, and the third day constitute a time mental image set, and the first mental image is a mental image obtained on the first day, and the third mental image is a mental image obtained on the third day.
[0079] The first psychological map is input into a psychological judgment network to obtain a first psychological judgment value and a first psychological judgment feature.
[0080] The first psychological judgment value is the psychological analysis type output by the first psychological map through the psychological judgment network, and the first psychological judgment feature is the feature vector before the final classification.
[0081] The third psychological map is input into the psychological judgment network to obtain a third psychological judgment value and a third psychological judgment feature.
[0082] The third psychological judgment value is the psychological analysis type output by the third psychological map through the psychological judgment network, and the third psychological judgment feature is the feature vector before the final classification.
[0083] The third psychological judgment feature and the first psychological judgment feature are input into a two-dimensional change network to obtain a two-dimensional change feature.
[0084] In this embodiment, the two-dimensional variation network is a fully connected neural network (Fully Connected Neural Network, FCNN), and the number of output neurons is the same as the number of psychological analysis categories, which is 4 in this embodiment.
[0085] Optionally, the temporal psychological three-dimensional image is convolved in sequence in both the width and height directions of the image with a three-dimensional convolution kernel and a step number of 1 through a temporal psychological change network to obtain a three-dimensional change feature.
[0086] Optionally, obtaining a psychological image based on the psychological data through a Bayesian network includes: (
[0087] The psychological data is input into the Bayesian network to obtain a psychological question output, which is the psychological question to be asked next time corresponding to the current answer.
[0088] According to the output of psychological questions, a probability distribution table is obtained through Bayesian estimation.
[0089] The probability distribution table is a table constructed to represent the probability of the next question corresponding to the selected answer.
[0090] According to the probability distribution table, a probabilistic psychological problem is obtained. The probabilistic psychological problem represents a psychological problem whose probability is greater than the probability corresponding to other psychological problems to be selected.
[0091] Among them, the probabilistic psychological question is the psychological question corresponding to the current answer with the highest probability.
[0092] The number of psychological questions asked is obtained. The number of psychological questions asked represents the total number of psychological questions asked from the initial psychological question to the current answer corresponding to the psychological question.
[0093] Among them, if the current question is the answer to the fourth question asked, and the fifth psychological question is to be found, then the number of psychological questions asked is 4. In the 16x16 image, the position 4,7 means that the next question corresponding to the answer selected for the fourth psychological question is question 7.
[0094] Construct a binary image with the number of psychological problems as width and height.
[0095] Set the position corresponding to the probabilistic psychological question to 1 and the values of other positions to 0 to obtain the psychological image.
[0096] Through the above method, since the final number of psychological questions asked is the same as the total number of psychological questions, the total number of psychological questions is directly used to construct the same width and height. A Bayesian network is established based on the number of questions and the relationship between the questions. For example, how to select the next question in the first layer.
[0097] Optionally, the Bayesian network construction method includes:
[0098] The training psychological data and the annotated psychological question data are obtained. The training psychological data represent answers to multiple psychological questions answered by multiple historical users. The annotated psychological question data are questions to be selected by the user that better reflect the type of psychological analysis.
[0099] Among them, if the training psychological data is question 1 "Who do you usually communicate with at a party?" A. "Communicate with many people, including strangers" B. "Only communicate with some friends." The choice of answer A can be question 21 or question 3, which means that it is better to ask question 2 when the answer is A. Question 2 is "When making a choice, you usually choose:" A. "Very careful" B. "A little impulsive." If the answer to question 1 is B, ask question 3 "At work, you hope your tasks are:" A. "Pre-arranged" B. + "Unscheduled".
[0100] The training psychological data is input into the Bayesian network to obtain the training psychological question probability, which indicates the probability that answering the previous psychological question can identify the next psychological question.
[0101] The Bayesian network is a network constructed according to multiple psychological problems in the database.
[0102] According to the training psychological problem probability and the labeled psychological problem data, the Bayesian network is backward trained through the loss function to obtain a trained Bayesian network.
[0103] Among them, in this embodiment, the loss function is a cross entropy loss function.
[0104] Through the above method, the probability of the previous question is used to determine the next question.
[0105] Optionally, obtaining the psychological analysis type through a classifier according to the cluster center feature includes:
[0106] The cluster center features are input into a fully connected neural network, the features of the cluster center features are extracted, and cluster analysis features are obtained.
[0107] Among them, in this embodiment, the fully connected neural network (Fully Connected Neural Network, FCNN) is a network that is different from the two-dimensional variation network in terms of the number of neurons and structure.
[0108] The cluster analysis features are input into a classifier for classification to obtain a psychological analysis type.
[0109] In this embodiment, the classifier is a softmax function.
[0110] Optional training methods for psychological judgment networks include:
[0111] A psychological judgment training set is obtained, wherein the psychological judgment training set includes a plurality of training psychological images and a plurality of corresponding annotated psychological analysis types.
[0112] The training psychological image is input into the psychological judgment network to obtain an output psychological analysis type.
[0113] In this embodiment, the psychological judgment network is a convolutional neural network (CNN), and the number of output heads is the same as the number of psychological analysis categories, which is 4 in this embodiment.
[0114] The output psychological analysis type and the annotated psychological analysis type are calculated and back-propagated to obtain a trained psychological judgment network.
[0115] Among them, in this embodiment, the loss function for calculating the loss is a cross entropy loss function.
[0116] Optional training methods for two-dimensional transform networks include:
[0117] A two-dimensional variation training set is obtained, wherein the two-dimensional variation training set includes a plurality of training psychological judgment features and a plurality of corresponding annotated psychological analysis types.
[0118] The psychological judgment feature is input into a two-dimensional variation network to obtain an output psychological analysis type.
[0119] In this embodiment, the two-dimensional variation network is a fully connected neural network (Fully Connected Neural Network, FCNN), and the number of output neurons is the same as the number of psychological analysis categories, which is 4 in this embodiment.
[0120] The output psychological analysis type and the labeled psychological analysis type are calculated and propagated back through the loss to obtain a trained two-dimensional variation network.
[0121] Among them, in this embodiment, the loss function for calculating the loss is a cross entropy loss function.
[0122] Through the above method, the two-dimensional change network is used as an auxiliary condition to better judge the psychological type after the change.
[0123] Example 2
[0124] Based on the above-mentioned psychological data analysis method based on Bayesian network, an embodiment of the present invention further provides a psychological data analysis system based on Bayesian network, the system comprising an acquisition module, a Bayesian module, a psychological image module, a psychological change detection module, a clustering module and a classification module:
[0125] The acquisition module is used to obtain psychological data at multiple time points. The psychological data represents the types of answers selected by the user when asking questions for psychological consultation.
[0126] The Bayesian module is used to obtain a psychological image based on the psychological data through a Bayesian network. Multiple psychological images are obtained corresponding to multiple time points. The psychological image is an image that uses pixel values to represent the relationship between psychological data.
[0127] The mental image module is used to form a temporal mental image set from three mental images at consecutive time points. n mental images correspond to n-2 temporal mental image sets.
[0128] The psychological change detection module is used to obtain psychological features according to the temporal psychological image set through the temporal psychological change network. n-2 temporal psychological images correspond to n-2 psychological features.
[0129] The clustering module is used to cluster according to n-2 psychological characteristics and obtain cluster center features.
[0130] The classification module is used to obtain the psychological analysis type through a classifier according to the cluster center characteristics.
[0131] The algorithm and display provided herein are not inherently related to any particular computer, virtual system or other device. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious that the structure required for constructing such systems. In addition, the present invention is not directed to any specific programming language either. It should be understood that various programming languages can be utilized to realize the content of the present invention described herein, and the description of the above specific languages is for disclosing the best mode of the present invention.
[0132] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.
[0133] Similarly, it should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the intention that the claimed invention requires more features than those explicitly recited in each claim. More specifically, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present invention.
[0134] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and further may be divided into a plurality of submodules or subunits or subcomponents. All features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device so disclosed may be combined in any combination, except that at least some of such features and / or processes or units are mutually exclusive. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature that provides the same, equivalent or similar purpose.
[0135] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.
[0136] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the apparatus according to an embodiment of the present invention. The present invention may also be implemented as a device or device program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.
[0137] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.
Claims
1. A psychological data analysis method based on Bayesian network, characterized in that: include: Obtaining psychological data at multiple time points; the psychological data represents the types of answers selected by the user when asking questions for psychological consultation; According to the psychological data, a psychological image is obtained through a Bayesian network; a plurality of psychological images are obtained corresponding to a plurality of time points; the psychological image is an image that uses pixel values to represent the relationship between the psychological data; The three mental images of consecutive time points constitute a time mental image set; n mental images correspond to n-2 time mental image sets; According to the time psychological image set, psychological characteristics are obtained through a time psychological change network; n-2 time mental images correspond to n-2 mental features; Clustering is performed based on n-2 psychological characteristics to obtain cluster center characteristics; According to the cluster center characteristics, a psychological analysis type is obtained through a classifier; The method of obtaining a psychological image based on the psychological data through a Bayesian network includes: Inputting the psychological data into the Bayesian network to obtain a psychological question output; the psychological question output is the psychological question to be asked next time corresponding to the current answer; According to the output of psychological questions, a probability distribution table is obtained through Bayesian estimation; According to the probability distribution table, a probabilistic psychological problem is obtained; the probabilistic psychological problem indicates a psychological problem with a probability greater than that of other psychological problems to be selected; Obtaining the number of psychological questions asked; the number of psychological questions asked represents the total number of psychological questions asked from the initial psychological question to the current answer; A binary image is constructed with the number of psychological problems as width and height; The position corresponding to the probabilistic psychological question is set to 1, and the values of other positions are set to 0 to obtain a psychological image; the value of 1 in the psychological image represents the next psychological question that is the only one in the number of psychological questions asked.
2. A Bayesian network-based psychological data analysis method according to claim 1, characterized in that: The psychological characteristics are obtained according to the time psychological image set through the time psychological change network, including: Judging psychological changes based on two psychological images in the temporal psychological image set to obtain two-dimensional change features; Construct a three-dimensional time mental image by arranging the mental images in the time mental image set in the order from far to near in time; According to the time-psychology three-dimensional image, a three-dimensional change feature is obtained through a time-psychology change network; The three-dimensional change feature and the two-dimensional change feature are averaged to obtain the psychological feature.
3. A Bayesian network-based psychological data analysis method according to claim 2, characterized in that: The step of judging the psychological change based on the two psychological images in the temporal psychological image set to obtain the two-dimensional change feature includes: A first mental image and a third mental image are obtained; the first mental image is an image in the time mental image set whose corresponding time point is less than the time points corresponding to other mental images; the third mental image is an image in the time mental image set whose corresponding time point is greater than the time points corresponding to other mental images; Inputting the first psychological map into a psychological judgment network to obtain a first psychological judgment value and a first psychological judgment feature; Inputting the third psychological map into a psychological judgment network to obtain a third psychological judgment value and a third psychological judgment feature; The third psychological judgment feature and the first psychological judgment feature are input into a two-dimensional change network to obtain a two-dimensional change feature.
4. The psychological data analysis method based on Bayesian network according to claim 1, characterized in that: Through the temporal psychological change network, the temporal psychological three-dimensional image is convolved in sequence in the width and height directions of the image with a three-dimensional convolution kernel with a step number of 1 to obtain a three-dimensional change feature.
5. The psychological data analysis method based on Bayesian network according to claim 1, characterized in that: The Bayesian network construction method comprises: Obtain training psychological data and annotated psychological question data; the training psychological data represents answers to multiple psychological questions answered by multiple historical users; the annotated psychological question data is a question to be selected by the user that better reflects the type of psychological analysis; Inputting the training psychological data into the Bayesian network to obtain the training psychological question probability; the psychological question probability indicates the probability that answering the previous psychological question can identify the next psychological question; According to the training psychological problem probability and the labeled psychological problem data, the Bayesian network is backward trained through the loss function to obtain a trained Bayesian network.
6. The psychological data analysis method based on Bayesian network according to claim 1, characterized in that: The psychological analysis type is obtained by a classifier according to the cluster center characteristics, including: Inputting the cluster center feature into a fully connected neural network, extracting the features of the cluster center feature, and obtaining a cluster analysis feature; The cluster analysis features are input into a classifier for classification to obtain a psychological analysis type.
7. The psychological data analysis method based on Bayesian network according to claim 1, characterized in that: The training method of the psychological judgment network includes: Obtaining a psychological judgment training set; the psychological judgment training set includes a plurality of training psychological images and a plurality of corresponding annotated psychological analysis types; Inputting the training psychological image into a psychological judgment network to obtain an output psychological analysis type; The output psychological analysis type and the annotated psychological analysis type are calculated and back-propagated to obtain a trained psychological judgment network.
8. The psychological data analysis method based on Bayesian network according to claim 1, characterized in that: The training method of the two-dimensional change network includes: Obtaining a two-dimensional variation training set; the training set includes a plurality of training psychological judgment features and a plurality of corresponding annotated psychological analysis types; Inputting the psychological judgment feature into a two-dimensional variation network to obtain an output psychological analysis type; The output psychological analysis type and the labeled psychological analysis type are calculated and propagated back through the loss to obtain a trained two-dimensional variation network.
9. A psychological data analysis system based on Bayesian network, characterized in that: include: Acquisition module: obtains psychological data at multiple time points; the psychological data represents the types of answers selected by the user when asking questions for psychological consultation; Bayesian module: Based on the psychological data, a psychological image is obtained through a Bayesian network; multiple psychological images are obtained corresponding to multiple time points; the psychological image is an image that uses pixel values to represent the relationship between psychological data; Mental image module: three mental images at consecutive time points constitute a temporal mental image set; n mental images correspond to n-2 temporal mental image sets; Psychological change detection module: obtaining psychological characteristics through a temporal psychological change network according to the temporal psychological image set; n-2 time mental images correspond to n-2 mental features; Clustering module: clustering based on n-2 psychological characteristics to obtain cluster center features; Classification module: according to the cluster center characteristics, the psychological analysis type is obtained through a classifier; The method of obtaining a psychological image based on the psychological data through a Bayesian network includes: Inputting the psychological data into the Bayesian network to obtain a psychological question output; the psychological question output is the psychological question to be asked next time corresponding to the current answer; According to the output of psychological questions, a probability distribution table is obtained through Bayesian estimation; According to the probability distribution table, a probabilistic psychological problem is obtained; the probabilistic psychological problem indicates a psychological problem with a probability greater than that of other psychological problems to be selected; Obtaining the number of psychological questions asked; the number of psychological questions asked represents the total number of psychological questions asked from the initial psychological question to the current answer; A binary image is constructed with the number of psychological problems as width and height; The position corresponding to the probabilistic psychological question is set to 1, and the values of other positions are set to 0 to obtain a psychological image; the value of 1 in the psychological image represents the next psychological question that is the only one in the number of psychological questions asked.
Citation Information
Patent Citations
Systems, methods, kits, and apparatuses for digital product network systems and biology-based value chain networks
CA3177585A1
Predicting psychometric profiles from behavioral data using machine-learning while maintaining user anonymity
CN109451757A
Automatic observing technology for metal abnormal change based on social network
CN109903851A
Image fusion method based on Bayesian theory
CN110570386A
Art psychology analyzer using voice, text, and picture data and art psychology analysis method performed by it
KR102403702B1
