CN114004253A - Information flow estimation method and system based on inter-brain interaction - Google Patents
Information flow estimation method and system based on inter-brain interaction Download PDFInfo
- Publication number
- CN114004253A CN114004253A CN202111222218.8A CN202111222218A CN114004253A CN 114004253 A CN114004253 A CN 114004253A CN 202111222218 A CN202111222218 A CN 202111222218A CN 114004253 A CN114004253 A CN 114004253A
- Authority
- CN
- China
- Prior art keywords
- group
- training
- frequency band
- electroencephalogram signal
- recommender
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2503/00—Evaluating a particular growth phase or type of persons or animals
- A61B2503/12—Healthy persons not otherwise provided for, e.g. subjects of a marketing survey
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Abstract
The invention provides an information flow estimation method and system based on interactivity between brains, comprising the following steps: acquiring an original electroencephalogram signal of a training recommender and an original electroencephalogram signal of a training recommender; carrying out real-time preprocessing and wavelet transformation on the original electroencephalogram signal to obtain a processed electroencephalogram signal; acquiring a frequency band electroencephalogram signal of a preset frequency band from the processed electroencephalogram signal; performing group correlation calculation on the frequency band electroencephalograms corresponding to every two training recommenders to obtain a neural similarity index, and calculating the frequency band electroencephalograms of the training recommenders and the frequency band electroencephalograms of the training recommenders to obtain a PDC value; training a neural network model according to the nerve similarity index and the PDC value of the interphalangeal connection degree to create a multi-brain connectivity model; and a multi-brain connectivity model is adopted to estimate the brain information transfer direction between two recommenders or between a recommender and a recommender. The neural similarity of the population can be effectively monitored, and the estimation accuracy is further improved.
Description
Technical Field
The invention relates to the field of brain signal learning, in particular to an information flow estimation method and system based on interactivity between brains.
Background
The technical development promotes live broadcast delivery, video delivery and other commercial forms to advance in a crossing manner, and the competition is more and more intense. The e-commerce industry has raised higher requirements for accurately knowing audience requirements, preferences, experiences and feedback. In live broadcast recommendation and video recommendation, the interest and the like of consumers are dynamically changed, which is called information flow recommendation and cannot be simply measured. How to accurately capture the preference of an online consumer and enable the consumer to receive the recommendation of a main broadcast is crucial to the development of a live broadcast platform of an e-commerce and becomes a difficult problem which needs to be broken through urgently in the field of live broadcast of the e-commerce.
The information flow recommendation refers to the recommendation of related product information through a computer system and a communication network, namely live broadcast delivery, video recommendation and the like in daily life. In information flow recommendation, information interaction exists between a recommending person and a recommended person, and the interaction drives dynamic change of brain nerve activity. However, the difficulty of the current mainstream research technology, method and implementation is extremely difficult to accurately capture the dynamic change of nerves in the brain, and more studies are focused on the analysis and recognition of the internal activities of the brain at an individual level, so that the interphalangeal nerve similarity calculation at a population level is neglected. On the other hand, at the present stage, measurement of inter-brain similarity of population is still mainly carried out in traditional modes such as questionnaires, self-reports, behavioral experiments and the like, and an objective and procedural population neural similarity calculation method is lacked, so that a method for recommending, predicting and optimizing information flow based on inter-brain interactivity and individual inter-brain EEG correlation is lacked.
Disclosure of Invention
Aiming at the problems in the prior art, an information flow estimation method and system based on the interactivity between brains are provided.
The specific technical scheme is as follows:
an information flow estimation method based on brain interactivity comprises the following steps:
step S1, dividing the experimental subject group into a training recommender group and a training recommender group according to the task type, dividing a training recommender and a training recommender into a group of experimental groups, and acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training recommender, wherein the training recommender group and the training recommender group have the same stimulus source;
step S2, performing real-time preprocessing and wavelet transformation on the first original electroencephalogram signal to obtain a first processed electroencephalogram signal;
performing real-time preprocessing and wavelet transformation on the second original electroencephalogram signal to obtain a second processed electroencephalogram signal;
step S3, acquiring data of a first preset frequency band from the first processed electroencephalogram signal, and recording the acquired data of the preset frequency band as the electroencephalogram signal of the first frequency band;
acquiring data of a second preset frequency band from the second processed electroencephalogram signal, and recording the acquired data of the preset frequency band as the electroencephalogram signal of the second frequency band;
step S4, performing inter-group correlation calculation on every two second frequency band electroencephalogram signals to obtain a neural similarity index group of the training recommended group, wherein the neural similarity index group comprises a neural similarity index between every two second frequency band electroencephalogram signals;
calculating the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups according to the bias orientation coherent analysis so as to calculate and obtain a PDC value group of the interphalangeal connection degree between the training recommender group and the training recommender group, wherein the PDC value group comprises the PDC value of the interphalamic connection degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups;
step S5, training the neural network model according to the PDC value group of the neural similarity index group and the interphalangeal connection degree to create and obtain a multi-brain connectivity model;
step S6, estimating the transfer direction of the brain information between at least two recommenders to be estimated in the recommenders to be estimated group by adopting a multi-brain connectivity model, and evaluating the stimulus sources of the recommenders to be estimated group according to the estimation result;
estimating the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated by adopting a multi-brain connectivity model, and evaluating the recommender group to be estimated according to the estimation result.
Preferably, the information flow prediction method based on the inter-brain interaction further includes, after step S2:
step A1, acquiring data of a first preset channel from the first processed electroencephalogram signal, and recording the acquired data of the preset channel as the first channel electroencephalogram signal;
acquiring data of a second preset channel from the second processed electroencephalogram signal, and recording the acquired data of the preset channel as a second channel electroencephalogram signal;
step A2, performing connectivity estimation on all second channel electroencephalograms to obtain first estimation data;
and performing connectivity estimation on the first channel electroencephalogram signal and the second channel electroencephalogram signal to obtain second estimation data.
Preferably, the information flow prediction method based on the brain interactivity further includes the following steps in step S5:
judging whether the neural network model completes training or not by adopting the first estimation data and the second estimation data;
if yes, a multi-brain connectivity model is created;
if not, the process returns to step S1.
Preferably, the information flow prediction method based on the inter-brain interaction includes the following steps in step S1: and synchronously acquiring a first original electroencephalogram signal of a training recommender and a second original electroencephalogram signal of a training recommender in the experimental group by adopting a multi-channel electroencephalogram measuring device.
Preferably, the information flow prediction method based on the inter-brain interaction includes the following steps in step S3:
acquiring an energy value of each first preset time period of a first preset frequency band from the first processed electroencephalogram signal, and taking all energy values in the first preset frequency band as the first frequency band electroencephalogram signal;
and acquiring the energy value of each second preset time period of the second preset frequency band from the second processing electroencephalogram signal, and taking all the energy values in the second preset frequency band as the electroencephalogram signal of the second frequency band.
Preferably, the information flow prediction method based on the inter-brain interaction, wherein the obtaining of the neural similarity index set in step S4 specifically includes the following steps:
pairwise matching the second frequency band electroencephalogram signals to obtain second frequency band electroencephalogram signal sets, and performing inter-group correlation calculation on each group of second frequency band electroencephalogram signal sets by adopting a first preset calculation tool to obtain the neural similarity index of each group of second frequency band electroencephalogram signal sets.
Preferably, the information flow prediction method based on brain interactivity includes, in step S4, obtaining the PDC value group, specifically including the following steps:
inputting the first frequency band electroencephalogram signal and the second frequency band electroencephalogram signal in each experimental group into the following formula to obtain a PDC value group:
wherein, PDCxy(f) When 1, αx(f) And alphay(f) The consistency is achieved;
Axy(f)a Fourier transform for representing coefficients of the multivariate autoregressive model;
αx(f) column x for axy (f);
x is used for representing a first frequency band electroencephalogram signal;
y is used for representing a second frequency band electroencephalogram signal;
PDCxy(f) guidance of electroencephalogram representing recommender to electroencephalogram of recommenderThe degree of action.
Preferably, the information flow prediction method based on the brain interactivity comprises a neural network model and a multi-brain connectivity model, wherein the neural network model comprises a first neural network and a second neural network, and the multi-brain connectivity model comprises a first multi-brain connectivity model and a second multi-brain connectivity model;
step S5 specifically includes the following steps:
step S51, performing data conversion on the neural similarity indexes in the neural similarity index group by adopting a preset conversion algorithm to obtain first conversion data;
performing data conversion on the PDC values in the PDC value set by adopting a preset conversion algorithm to obtain second conversion data;
step S52, dividing the first transformed data into a first test set and a first training set;
dividing the second conversion data into a second test set and a second training set;
step S53, obtaining data feature dimensions of the first conversion data and the second conversion data;
step S54, when the data characteristic dimension is smaller than the preset characteristic dimension threshold value, adopting a first training set and a second training set to train a first neural network to obtain a first training neural network model, and adopting a first testing set and a second testing set to verify the first training neural network model to create and obtain a first multi-brain connectivity model;
and when the data characteristic dimension is larger than or equal to a preset characteristic dimension threshold value, training the second neural network by adopting the first training set and the second training set to obtain a second training neural network model, and verifying the second training neural network model by adopting the first testing set and the second testing set to create and obtain a first multi-brain connectivity model.
Preferably, the information flow prediction method based on the brain interactivity is characterized in that the second neural network is an SVM-AdaBoost classification model, and the SVM-AdaBoost classification model comprises a plurality of SVM classifiers.
An information flow prediction apparatus based on inter-brain interactivity, comprising:
the acquisition unit is used for dividing the experimental subject group into a training recommender group and a training referee group according to the task type, dividing a training recommender and a training referee into a group of experimental groups, and acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training referee, wherein the training recommender group and the training referee group have the same stimulus source;
the processing unit is connected with the acquisition unit and is used for carrying out real-time preprocessing and wavelet transformation on the first original electroencephalogram signal to obtain a first processed electroencephalogram signal;
the system is also used for carrying out real-time preprocessing and wavelet transformation on the second original electroencephalogram signal to obtain a second processed electroencephalogram signal;
the frequency band data acquisition unit is connected with the processing unit and used for acquiring data of a first preset frequency band from the first processed electroencephalogram signal and recording the acquired data of the preset frequency band as the first frequency band electroencephalogram signal;
the electroencephalogram processing module is also used for acquiring data of a second preset frequency band from the second processed electroencephalogram signal and recording the acquired data of the preset frequency band as the electroencephalogram signal of the second frequency band;
the calculating unit is connected with the frequency band data acquiring unit and is used for performing inter-group correlation calculation on every two second frequency band electroencephalogram signals to obtain a neural similarity index group of a training recommended group, wherein the neural similarity index group comprises a neural similarity index between every two second frequency band electroencephalogram signals;
calculating the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups according to the bias orientation coherent analysis to obtain a PDC value group of the interphalangeal connection degree between the training recommender group and the training recommender group, wherein the PDC value group comprises the PDC value of the interphalamic connection degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups;
the creating unit is connected with the calculating unit and used for training the neural network model according to the PDC value group of the neural similarity index group and the interphalangeal connection degree so as to create and obtain a multi-brain connectivity model;
the estimation unit is used for estimating the transfer direction of the brain information between at least two recommenders to be estimated in the recommenders to be estimated group by adopting a multi-brain connectivity model and evaluating the stimulus sources of the recommenders to be estimated group according to the estimation result;
estimating the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated by adopting a multi-brain connectivity model, and evaluating the recommender group to be estimated according to the estimation result.
The technical scheme has the following advantages or beneficial effects: the data of the group neural similarity in information flow recommendation is obtained in real time through real-time monitoring, so that the E-commerce live broadcast platform is helped to make scientific decisions. Meanwhile, the method has the characteristics of non-invasiveness, safety, high efficiency and low cost, can be used for popularizing the fields of live broadcast recommendation, movie and television series recommendation, advertisement recommendation effect evaluation and the like, and has wide market application prospects.
Drawings
Embodiments of the present invention will now be described more fully hereinafter with reference to the accompanying drawings. The drawings are, however, to be regarded as illustrative and explanatory only and are not restrictive of the scope of the invention.
Fig. 1 is a flowchart of an embodiment of an information flow estimation method based on inter-brain interactivity according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
The invention is further described with reference to the following drawings and specific examples, which are not intended to be limiting.
The invention comprises an information flow estimation method based on the interactivity between brains, as shown in figure 1, comprising the following steps:
step S1, dividing the experimental subject group into a training recommender group and a training recommender group according to the task type, dividing a training recommender and a training recommender into a group of experimental groups, and acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training recommender, wherein the training recommender group and the training recommender group have the same stimulus source;
step S2, performing real-time preprocessing and wavelet transformation on the first original electroencephalogram signal to obtain a first processed electroencephalogram signal;
performing real-time preprocessing and wavelet transformation on the second original electroencephalogram signal to obtain a second processed electroencephalogram signal;
step S3, acquiring data of a first preset frequency band from the first processed electroencephalogram signal, and recording the acquired data of the preset frequency band as the electroencephalogram signal of the first frequency band;
acquiring data of a second preset frequency band from the second processed electroencephalogram signal, and recording the acquired data of the preset frequency band as the electroencephalogram signal of the second frequency band;
step S4, performing inter-subject correlation (ISC) calculation on every two second frequency band electroencephalogram signals to obtain a neural similarity index group of a training referee group, wherein the neural similarity index group comprises a neural similarity index (E) between every two second frequency band electroencephalogram signals;
calculating the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each experimental group according to bias-directed coherent analysis (PDC) so as to calculate a PDC value group of the interphalangeal connection degree between the training recommender group and the training recommender group, wherein the PDC value group comprises the PDC value of the interphalamic connection degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each experimental group;
step S5, training the neural network model according to the PDC value group of the neural similarity index group and the interphalangeal connection degree to create and obtain a multi-brain connectivity model;
step S6, estimating the transfer direction of the brain information between at least two recommenders to be estimated in the recommenders to be estimated group by adopting a multi-brain connectivity model, and evaluating the stimulus sources of the recommenders to be estimated group according to the estimation result;
estimating the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated by adopting a multi-brain connectivity model, and evaluating the recommender group to be estimated according to the estimation result.
In the embodiment, the neural similarity of the groups can be effectively monitored by monitoring the electroencephalogram signals of a plurality of people simultaneously and carrying out real-time dynamic analysis on the electroencephalogram signals according to the correlation of the neural activity between brains.
In the above embodiment, the multi-brain connectivity model obtained by creation may be used to estimate a brain information transmission direction between at least two to-be-estimated recommenders of a group of recommenders to be estimated, and evaluate a stimulus source of the group of recommenders to be estimated according to an estimation result, so as to determine whether the stimulus source meets a user requirement, for example, when the stimulus source is a sales advertisement, it may be determined whether the sales advertisement can stimulate a desire to purchase of the recommenders according to the evaluation result;
the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated can be estimated through the created multi-brain connectivity model, and the recommender group to be estimated is evaluated according to the estimation result, so as to determine whether the recommender to be estimated meets the user requirement, for example, whether the recommender can stimulate the purchase desire of the recommender can be determined according to the evaluation result.
In the embodiment, the data of the group neural similarity in information flow recommendation is obtained in real time through real-time monitoring, so that a live E-commerce platform is helped to make scientific decisions. Meanwhile, the method has the characteristics of non-invasiveness, safety, high efficiency and low cost, can be used for popularizing the fields of live broadcast recommendation, movie and television series recommendation, advertisement recommendation effect evaluation and the like, and has wide market application prospects.
In the above embodiment, the task type may be to sell a certain product and to buy the product, and the task type of the training recommender population at this time is to "sell a certain product", and therefore the task type of the training recommender population at this time is to "buy the product".
In the above embodiment, the stimulus sources of the training recommender population and the training referee population are the same, and the stimulus sources may be emotion inducing environments: recommended scenarios such as products, apps, or movies;
for example, the emotional inducing environment of the training recommender population and the training referee population may be a recommendation scenario of a product, that is, the stimulus sources of the training recommender population and the training referee population may be advertisements of the product, but the task types of the training recommender population and the training referee population are different, for example, the task type of the training recommender population is "sell the product" and the task type of the training referee population is "buy the product", so the training recommender population may watch the advertisements of the product with an idea of how to sell the product, and the training referee population may watch the advertisements of the product with an idea of whether to buy the product.
In step S1, the training recommenders and the training recommenders are classified into a set of experimental groups in a one-to-one manner.
In the above embodiment, after step S2, the method further includes:
step A1, acquiring data of a first preset channel from the first processed electroencephalogram signal, and recording the acquired data of the preset channel as the first channel electroencephalogram signal;
acquiring data of a second preset channel from the second processed electroencephalogram signal, and recording the acquired data of the preset channel as a second channel electroencephalogram signal;
step A2, performing connectivity estimation on all second channel electroencephalograms to obtain first estimation data;
and performing connectivity estimation on the first channel electroencephalogram signal and the second channel electroencephalogram signal to obtain second estimation data.
In the above embodiment, the first preset channel and the second preset channel may be the same channel, that is, the first channel electroencephalogram signal and the second channel electroencephalogram signal of the same preset channel are obtained from the first processed electroencephalogram signal and the second processed electroencephalogram signal respectively, then connectivity estimation is performed on all the second channel electroencephalogram signals to obtain the first estimation data, and connectivity estimation can be performed on the first channel electroencephalogram signal and the second channel electroencephalogram signal to obtain the second estimation data, so that whether the neural network model completes training or not is determined subsequently.
In the above embodiment, the step S5 further includes the following steps:
judging whether the neural network model completes training or not by adopting the first estimation data and the second estimation data;
if yes, the multi-brain connectivity model is created and obtained;
if not, the process returns to step S1.
In the above embodiment, whether the neural network model completes training is determined by the first estimation data and the second estimation data, so as to improve the confidence of the created multi-brain connectivity model.
For example, determining whether the neural network model completes training using the first estimation data and the second estimation data may include the following steps:
step A3, judging whether the connectivity degree between every two second frequency band electroencephalogram signals in the neural network model exceeds a first preset threshold value by adopting first estimation data;
if yes, interpupillary interaction exists between every two second frequency band electroencephalogram signals;
if not, the interactivity between every two second frequency band electroencephalograms is poor, all processes need to be checked, and the step S1 is returned;
step A4, judging whether the connectivity degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each experimental group in the neural network model exceeds a second preset threshold value by adopting second estimation data;
if yes, representing that interphalangeal interaction exists between the first frequency band electroencephalogram signal and the second frequency band electroencephalogram signal in each experimental group, and judging that the neural network model completes training to create and obtain the multi-brain connectivity model;
if not, the interactivity between the first frequency band electroencephalogram signal and the second frequency band electroencephalogram signal in each experimental group is poor, all the processes need to be checked, and the step S1 is returned.
It should be noted that the connectivity degree includes the degree of flow direction and consistency of the brain electricity.
The sequence of the step A3 and the sequence of the step A4 can be changed, and a multi-brain connectivity model can be created only when the judgment results of the interphalamic interaction between every two second frequency band electroencephalograms and the interphalamic interaction between the first frequency band electroencephalograms and the second frequency band electroencephalograms in each experimental group are obtained.
In the above embodiment, step S1 specifically includes the following steps: and synchronously acquiring a first original electroencephalogram signal of a training recommender and a second original electroencephalogram signal of a training recommender in the experimental group by adopting a multi-channel electroencephalogram measuring device.
In the embodiment, the training recommenders and the training recommenders in a group of experimental groups are synchronously acquired to synchronously acquire the first original electroencephalogram signal and the second original electroencephalogram signal, so that the synchronous real-time acquisition of the original electroencephalogram signals by multiple persons is realized, the subsequent synchronous real-time monitoring and dynamic analysis of the electroencephalogram signals by the multiple persons are facilitated, the study on the similarity of group nerves in a process and a whole flow is further realized, and the problems of post-sampling, strong subjectivity, social period deviation and the like of the traditional measuring means are effectively avoided.
The multi-channel electroencephalogram measuring equipment can adopt 64-lead electrode caps, saline water or gel electrodes are used, and the impedance of each electrode point is lower than 5k omega, so that the electrode caps are well contacted with the scalp, and electroencephalogram signals can be conveniently acquired.
In the above embodiment, the step of preprocessing the acquired first raw brain electrical signal in step S2 includes, but is not limited to: amplifying the acquired first original electroencephalogram signal, analyzing section interception, noise reduction, artifact removal and band-pass filtering, wherein the artifact removal comprises but is not limited to: and the interference of electrooculogram, myoelectricity, electrocardio and power frequency is removed.
Similarly, the step of preprocessing the acquired second original electroencephalogram signal in step S2 is the same as the step of preprocessing the first original electroencephalogram signal, and will not be elaborated here.
And performing wavelet transformation on the preprocessed first original electroencephalogram signal to obtain a first processed electroencephalogram signal, an
And performing wavelet transformation on the preprocessed second original electroencephalogram signal to obtain a second processed electroencephalogram signal.
In the above embodiment, step S3 specifically includes the following steps:
acquiring an energy value of each first preset time period of a first preset frequency band from the first processed electroencephalogram signal, and taking all energy values in the first preset frequency band as the first frequency band electroencephalogram signal;
and acquiring the energy value of each second preset time period of the second preset frequency band from the second processing electroencephalogram signal, and taking all the energy values in the second preset frequency band as the electroencephalogram signal of the second frequency band.
In the above embodiment, the first preset frequency band and the second preset frequency band may be the same frequency band;
for example, at least one of "theta", "alpha", "beta", "gamma", etc. may be selected as the first preset frequency band.
In the above embodiment, the first preset time period and the second preset time period may be the same time period;
for example, the first preset time period and the second preset time period may be set to 1 second each.
In the above embodiment, all energy values in the first preset frequency band are used as first frequency band electroencephalogram signals, and all energy values in the second preset frequency band are used as second frequency band electroencephalogram signals, so that subsequent group interphalangeal neural similarity calculation depending on the electroencephalogram frequency band energy of a specific brain area is realized, and the accuracy of the neural similarity index and the PDC value is improved.
In the above embodiment, the obtaining of the neural similarity index set in step S4 specifically includes the following steps:
pairwise matching the second frequency band electroencephalogram signals to obtain second frequency band electroencephalogram signal sets, and performing inter-group correlation calculation on each group of second frequency band electroencephalogram signal sets by adopting a first preset calculation tool to obtain the neural similarity index of each group of second frequency band electroencephalogram signal sets.
In the above embodiment, the first predetermined computing tool may employ a NeuroRA toolkit based on python.
In the above embodiment, the obtaining of the PDC value group in step S4 specifically includes the following steps:
inputting the first frequency band electroencephalogram signal and the second frequency band electroencephalogram signal in each experimental group into the following formula to obtain a PDC value group:
wherein when the PDCxy(f) When 1, αx(f) And
the consistency is achieved;
Axy(f)is a fourier transform of the multiple autoregressive model coefficients;
αx(f) for the representation of Axy(f)The x-th column of (1);
x is used to represent the recommended human brain electrical signal;
y is used to represent the consumer brain electrical signal;
PDCxy(f) the brain-electrical signal guide system is used for expressing the degree of the guide effect of the recommended human brain-electrical signal on the brain-electrical signal of the recommended person;
in the above embodiment, the neural network model includes a first neural network and a second neural network, and the multiple brain connectivity model includes a first multiple brain connectivity model and a second multiple brain connectivity model;
step S5 specifically includes the following steps:
step S51, performing data conversion on the neural similarity indexes in the neural similarity index group by adopting a preset conversion algorithm to obtain first conversion data;
performing data conversion on the PDC values in the PDC value set by adopting a preset conversion algorithm to obtain second conversion data;
in the embodiment, the neural similarity index and the PDC value can be subjected to data conversion through a preset conversion algorithm, so that the data characteristics of the converted data are smoother when displayed, and the input requirement of a neural network model is met.
Preferably, the preset transformation algorithm may use a Box-Cox algorithm.
Step S52, dividing the first transformed data into a first test set and a first training set;
dividing the second conversion data into a second test set and a second training set;
in the above embodiment, because the two sets of data sets are separated from each other in the same data set, the data feature quantities of the two sets of data sets are ensured to be the same, the test set used for evaluation and the training set used for model training are ensured not to affect each other, and finally, the test set can be used for analyzing and judging the accuracy of the model.
The division ratio of the test set and the training set can be set in a user-defined mode according to requirements, for example, the total number of the data in the data set is N, the N data are averagely divided into K subsets, and each subset is provided with N/K data. At each verification, a plurality of the K subsets which are divided are used as a test set, then the rest subsets are used as a training set, and then model evaluation is carried out by using the test set.
The division mode of the second conversion data is consistent with that of the first conversion data;
for example, 30% of the first transformed data may be used as the first test set, and 70% of the first transformed data may be used as the first training set;
then 30% of the second transformed data is required as the second test set and 70% of the second transformed data is required as the second training set.
Step S53, obtaining data feature dimensions of the first conversion data and the second conversion data;
step S54, when the data characteristic dimension is smaller than the preset characteristic dimension threshold value, adopting a first training set and a second training set to train a first neural network to obtain a first training neural network model, and adopting a first testing set and a second testing set to verify the first training neural network model to create and obtain a first multi-brain connectivity model;
and when the data characteristic dimension is larger than or equal to a preset characteristic dimension threshold value, training the second neural network by adopting the first training set and the second training set to obtain a second training neural network model, and verifying the second training neural network model by adopting the first testing set and the second testing set to create and obtain a first multi-brain connectivity model.
In the above embodiments, the first neural network and the second neural network are different, for example, the first neural network may be an SVM classifier and the second neural network may be an SVM-AdaBoost classification model.
For example, in the aspect of training the SVM classifier, a grid search method may be used to determine SVM parameters for a training set, and then a radial basis function is selected according to the SVM parameters, for example, when the SVM parameters are determined to be 1, then an RBF gaussian radial basis kernel function may be used as the radial basis function in the SVM classifier.
In the above embodiment, the second neural network is an SVM-AdaBoost classification model, and the SVM-AdaBoost classification model includes a plurality of SVM classifiers.
In the above embodiments, in order to improve the utility of the second multi-brain connectivity model, a loss ratio may be set in the second multi-brain connectivity model. The second neural network is an SVM-AdaBoost classification model, the SVM-AdaBoost classification model is a strong classifier model formed by weighting a plurality of weak classifiers, the SVM-AdaBoost classification model comprises a plurality of SVM classifiers, the SVM classifier is a strong classifier, and the SVM classifier is taken as a weak classifier and added into the AdaBoost model, so that a good prediction effect can be achieved.
Specifically, in the process of training the second neural network, the data features of the training set are weighted through a loss proportion, weight initialization is carried out according to the importance of each data feature in the model, each data feature corresponds to one SVM classifier, and therefore weighting processing of each SVM classifier in the second neural network is achieved; and adding a weighted SVM in the AdaBoost lifting frame to pass through the lifting frame, so that the diversity of training set subsets can be obtained from the initial training set, and then generating a base classifier through the subsets, thereby calculating the error rate of SVM classification. The base classifier is composed of a plurality of SVM features, so that a corresponding base classifier can be generated after K rounds of training numbers are given. Although the recognition rate of each base classifier is not very high, the combined classifier has a very high recognition rate, so that the algorithm recognition rate is improved. And weighting the promoted classifier to generate a final result classifier.
In the above embodiment, in step S6, for the estimation of the brain information transfer direction between at least two recommenders to be estimated of the recommenders to be estimated group of recommenders, the multi-brain connectivity model may take the coincidence or non-coincidence as an output value;
for example, when the neural similarity index is greater than or equal to the first preset index threshold, it is determined that the neural similarity of the population is strong, and the multi-brain connectivity model at this time may be output as "consistent", that is, the evaluation result is: the group of recommenders to be estimated has resonance to the stimulus;
when the neural similarity index is greater than or equal to a second preset index threshold and smaller than a first preset index threshold, determining that the group neural similarity approaches, determining that the group neural similarity is strong, and outputting the consistency of the multi-brain connectivity model at the moment, namely the evaluation result is: the group of recommenders to be estimated has resonance to the stimulus;
when the neural similarity index is smaller than a second preset index threshold value, the neural similarity of the population is determined to be insufficient, the multi-brain connectivity model at the moment can be output as inconsistent, namely the evaluation result is as follows: the group of recommenders to be estimated does not resonate the stimulus.
The first preset index threshold is greater than the second preset index threshold, and the first preset index threshold and the second preset index threshold can be set in a self-defined manner, for example, here, the first preset index threshold can be set to 0.6, and the second preset index threshold can be set to 0.4.
In the above embodiment, for the estimation of the inter-brain information transfer direction between the recommender group to be estimated and the recommender group to be estimated, the multi-brain connectivity model may take acceptance or not as an output value;
for example, when the PDC value is close to the first preset PDC threshold, the recommender to be estimated in a group of experimental groups is considered to have a significant guiding effect on the recommender to be estimated, and the multi-brain connectivity model may be output as "receive", that is, the evaluation result is: the influence of the recommender to be estimated on the recommender to be estimated is high;
when the PDC value is close to the second preset PDC threshold value, it is determined that the interphalangeal neural similarity between the training recommenders and the training recommenders in a group of experimental groups is insufficient, and the multi-brain connectivity model at this time may be output as "unreceived", that is, the evaluation result is: the influence of the recommender to be estimated on the recommender to be estimated is low.
The first predetermined PDC threshold is greater than the second predetermined PDC threshold, and the first predetermined PDC threshold and the second predetermined PDC threshold may be custom set, for example, where the first predetermined PDC threshold may be set to 1 and the second predetermined PDC threshold may be set to 0, such that a PDC value greater than or equal to 0.5 may be considered to be close to the first predetermined PDC threshold and a PDC value less than 0.5 may be considered to be close to the second predetermined PDC threshold.
In the above embodiment, the recommended scene stimulation material may be presented by the inducing device, and the recommended scene stimulation material may be used as a visual stimulation source, and may be processed and stored.
In the embodiment, a comprehensive calculation strategy is formed from the interbrain consistency of the same group and the neural similarity between the recommender and the referee in the information flow recommendation scene, and the method has the characteristics of real-time monitoring and quantification, and can scientifically and comprehensively analyze the neural similarity between the recommender and the consumer in the recommendation situation. At present, electroencephalogram identification is mostly researched based on amplitude characteristics, and the inter-brain phase synchronization characteristics and the inter-brain signal activity directionality of a frequency space are ignored. The phase synchronization characteristic detects the correlation between signal pairs through the instantaneous phase relationship between signals, and can effectively identify the neural activity synchronism of people. The signal activity directionality can be used for determining the directional influence between signals in the multivariate data set, provides the directionality (frequency space) of the activity between brains on the basis of analyzing the synchronism, and is helpful for understanding the role relationship of brain areas among different individuals. The neural similarity of the recommenders of every two of the recommended population can be effectively identified by the correlation coefficient of the neural similarity values, and the neural similarity between the recommenders and the recommenders can be effectively identified by the correlation coefficient of the PDC values.
In addition, the invention innovatively provides a method for calculating the neural similarity index between a recommender and a referee by relying on the electroencephalogram frequency band energy of a specific brain area and an SVM-AdaBoost classification model, and the neural similarity between the recommender and a consumer (the recommender) and between all the consumers (the referee) under the information flow recommendation situation can be accurately predicted by the SVM-AdaBoost model.
Also included is an information flow prediction device based on brain interactivity, which includes:
the acquisition unit is used for dividing the experimental subject group into a training recommender group and a training referee group according to the task type, dividing a training recommender and a training referee into a group of experimental groups, and acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training referee, wherein the training recommender group and the training referee group have the same stimulus source;
the processing unit is connected with the acquisition unit and is used for carrying out real-time preprocessing and wavelet transformation on the first original electroencephalogram signal to obtain a first processed electroencephalogram signal;
the system is also used for carrying out real-time preprocessing and wavelet transformation on the second original electroencephalogram signal to obtain a second processed electroencephalogram signal;
the frequency band data acquisition unit is connected with the processing unit and used for acquiring data of a first preset frequency band from the first processed electroencephalogram signal and recording the acquired data of the preset frequency band as the first frequency band electroencephalogram signal;
the electroencephalogram processing module is also used for acquiring data of a second preset frequency band from the second processed electroencephalogram signal and recording the acquired data of the preset frequency band as the electroencephalogram signal of the second frequency band;
the calculating unit is connected with the frequency band data acquiring unit and is used for performing inter-group correlation calculation on every two second frequency band electroencephalogram signals to obtain a neural similarity index group of a training recommended group, and the neural similarity index group comprises a neural similarity index between every two second frequency band electroencephalogram signals;
calculating the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups according to the bias-orientation coherent analysis to obtain a PDC value group of the interphalangeal connection degree between the training recommender group and the training recommender group, wherein the PDC value group comprises the PDC value of the interphalamic connection degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups;
the creating unit is connected with the calculating unit and used for training the neural network model according to the PDC value group of the neural similarity index group and the interphalangeal connection degree so as to create and obtain a multi-brain connectivity model;
the estimation unit is used for estimating the transfer direction of the brain information between at least two recommenders to be estimated in the recommenders to be estimated group by adopting a multi-brain connectivity model and evaluating the stimulus sources of the recommenders to be estimated group according to the estimation result;
estimating the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated by adopting a multi-brain connectivity model, and evaluating the recommender group to be estimated according to the estimation result.
In the above embodiment, the information flow prediction apparatus further includes a visualization module, configured to display the evaluation result.
The information flow prediction device has the functions of dynamic prediction and visualization, and can help a user to visually acquire data of group neural similarity in information flow recommendation, so that a live E-commerce platform is helped to make scientific decisions. Meanwhile, the method has the characteristics of non-invasiveness, safety, high efficiency and low cost, can be used for popularizing the fields of live broadcast recommendation, movie and television series recommendation, advertisement recommendation effect evaluation and the like, and has wide market application prospects.
It should be noted that, the embodiments of the information flow prediction apparatus of the present invention are the same as the embodiments of the information flow prediction method, and are not described herein again.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.
Claims (10)
1. An information flow estimation method based on brain interactivity is characterized by comprising the following steps:
step S1, dividing an experimental subject group into a training recommender group and a training recommender group according to task types, dividing a training recommender and a training recommender into a group of experimental groups, and acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training recommender, wherein the training recommender group and the training recommender group have the same stimulus source;
step S2, performing real-time preprocessing and wavelet transformation on the first original electroencephalogram signal to obtain a first processed electroencephalogram signal;
performing real-time preprocessing and wavelet transformation on the second original electroencephalogram signal to obtain a second processed electroencephalogram signal;
step S3, acquiring data of a first preset frequency band from the first processed electroencephalogram signal, and recording the acquired data of the preset frequency band as the electroencephalogram signal of the first frequency band;
acquiring data of a second preset frequency band from the second processed electroencephalogram signal, and recording the acquired data of the preset frequency band as the electroencephalogram signal of the second frequency band;
step S4, performing inter-group correlation calculation on every two second frequency band electroencephalogram signals to obtain a neural similarity index group of the training referee group, wherein the neural similarity index group comprises a neural similarity index between every two second frequency band electroencephalogram signals;
calculating the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups according to bias-orientation coherent analysis so as to calculate a PDC value group of the interphalangeal connection degree between the training recommender group and the training recommender group, wherein the PDC value group comprises the PDC value of the interphalamic connection degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups;
step S5, training a neural network model according to the PDC value group of the neural similarity index group and the interphalangeal connection degree to create a multi-brain connectivity model;
step S6, estimating the transfer direction of the brain information between at least two recommenders to be estimated in the recommenders to be estimated group by adopting the multi-brain connectivity model, and evaluating the stimulus sources of the recommenders to be estimated according to the estimation result;
estimating the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated by using the multi-brain connectivity model, and evaluating the recommender group to be estimated according to the estimation result.
2. The method for predicting information flow based on inter-brain interactivity according to claim 1, further comprising after step S2:
step A1, acquiring data of a first preset channel from the first processed electroencephalogram signal, and recording the acquired data of the preset channel as a first channel electroencephalogram signal;
acquiring data of a second preset channel from the second processed electroencephalogram signal, and recording the acquired data of the preset channel as a second channel electroencephalogram signal;
step A2, performing connectivity estimation on all the second channel electroencephalograms to obtain first estimation data;
and performing connectivity estimation on the first channel electroencephalogram signal and the second channel electroencephalogram signal to obtain second estimation data.
3. The method for predicting information flow based on inter-brain interactivity according to claim 2, further comprising the following steps in said step S5:
judging whether the neural network model completes training or not by adopting the first estimation data and the second estimation data;
if yes, the multi-brain connectivity model is created and obtained;
if not, the process returns to step S1.
4. The method for predicting information flow based on inter-brain interactivity according to claim 1, wherein the step S1 specifically includes the following steps: and synchronously acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training recommender in the experimental group by adopting a multi-channel electroencephalogram measuring device.
5. The method for predicting information flow based on inter-brain interactivity according to claim 1, wherein the step S3 specifically includes the following steps:
acquiring an energy value of each first preset time period of the first preset frequency band from the first processed electroencephalogram signal, and taking all energy values in the first preset frequency band as the first frequency band electroencephalogram signal;
and acquiring the energy value of each second preset time period of the second preset frequency band from the second processed electroencephalogram signal, and taking all the energy values in the second preset frequency band as the electroencephalogram signal of the second frequency band.
6. The method for predicting information flow based on inter-brain interaction according to claim 1, wherein the obtaining the set of neural similarity indicators in step S4 specifically includes the following steps:
pairwise matching the second frequency band electroencephalogram signals to obtain second frequency band electroencephalogram signal sets, and performing inter-group correlation calculation on each group of second frequency band electroencephalogram signal sets by adopting a first preset calculation tool to obtain the neural similarity index of each group of second frequency band electroencephalogram signal sets.
7. The method for predicting information flow based on interactivity between brains according to claim 1, wherein the obtaining of the PDC value set in step S4 specifically includes the following steps:
inputting the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each experimental group into the following formula to obtain the PDC value group:
wherein, PDCxy(f) When 1, αx(f) And alphay(f) The consistency is achieved;
Axy(f) a Fourier transform for representing coefficients of the multivariate autoregressive model;
αx(f) for the representation of Axy(f) The x-th column of (1);
x is used for representing a first frequency band electroencephalogram signal;
y is used for representing a second frequency band electroencephalogram signal;
PDCxy(f) used for representing the guiding action degree of the electroencephalogram signal of the recommender to the electroencephalogram signal of the recommender.
8. The inter-brain interactivity based information flow prediction method of claim 1, wherein the neural network model comprises a first neural network and a second neural network, the multi-brain connectivity model comprises a first multi-brain connectivity model and a second multi-brain connectivity model;
the step S5 specifically includes the following steps:
step S51, performing data conversion on the neural similarity indexes in the neural similarity index group by adopting a preset conversion algorithm to obtain first conversion data;
performing data conversion on the PDC values in the PDC value set by adopting the preset conversion algorithm to obtain second conversion data;
step S52, dividing the first conversion data into a first test set and a first training set;
dividing the second transformed data into a second test set and a second training set;
step S53, obtaining data feature dimensions of the first converted data and the second converted data;
step S54, when the data characteristic dimension is smaller than a preset characteristic dimension threshold, adopting the first training set and the second training set to train the first neural network so as to obtain a first training neural network model, and adopting the first test set and the second test set to verify the first training neural network model so as to create and obtain the first multi-brain connectivity model;
and when the data characteristic dimension is larger than or equal to a preset characteristic dimension threshold value, training the second neural network by adopting the first training set and the second training set to obtain a second training neural network model, and verifying the second training neural network model by adopting the first testing set and the second testing set to create and obtain the first multi-brain connectivity model.
9. The inter-brain interaction based information flow prediction method of claim 1, wherein the second neural network is an SVM-AdaBoost classification model, the SVM-AdaBoost classification model comprising a plurality of SVM classifiers.
10. An information flow prediction apparatus based on inter-brain interactivity, comprising:
the acquisition unit is used for dividing an experimental subject group into a training recommender group and a training referee group according to task types, dividing a training recommender and a training referee into a group of experimental groups, and acquiring a first original electroencephalogram signal of the training recommender and a second original electroencephalogram signal of the training referee, wherein the training recommender group and the training referee group have the same stimulus source;
the processing unit is connected with the acquisition unit and is used for carrying out real-time preprocessing and wavelet transformation on the first original electroencephalogram signal to obtain a first processed electroencephalogram signal;
the second original electroencephalogram signal is also used for carrying out real-time preprocessing and wavelet transformation on the second original electroencephalogram signal to obtain a second processed electroencephalogram signal;
the frequency band data acquisition unit is connected with the processing unit and used for acquiring data of a first preset frequency band from the first processed electroencephalogram signal and recording the acquired data of the preset frequency band as the first frequency band electroencephalogram signal;
the electroencephalogram processing module is also used for acquiring data of a second preset frequency band from the second processed electroencephalogram signal and recording the acquired data of the preset frequency band as the electroencephalogram signal of the second frequency band;
the calculation unit is connected with the frequency band data acquisition unit and used for performing inter-group correlation calculation on every two second frequency band electroencephalogram signals to obtain a neural similarity index group of the training referee group, wherein the neural similarity index group comprises a neural similarity index between every two second frequency band electroencephalogram signals;
calculating the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups according to bias-orientation coherent analysis to obtain PDC value groups of the interphalangeal connection degree between the training recommender group and the training recommender group, wherein the PDC value groups comprise the PDC values of the interphalamic connection degree between the first frequency band electroencephalogram signals and the second frequency band electroencephalogram signals in each group of experiment groups;
the creating unit is connected with the calculating unit and used for training a neural network model according to the nerve similarity index group and the PDC value group of the interphalangeal connection degree so as to create and obtain a multi-brain connectivity model;
the estimation unit is used for estimating the transfer direction of the brain information between at least two recommenders to be estimated in the recommenders to be estimated group by adopting the multi-brain connectivity model, and evaluating the stimulus sources of the recommenders to be estimated according to the estimation result;
estimating the inter-brain information transmission direction between the recommender group to be estimated and the recommender group to be estimated by using the multi-brain connectivity model, and evaluating the recommender group to be estimated according to the estimation result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111222218.8A CN114004253A (en) | 2021-10-20 | 2021-10-20 | Information flow estimation method and system based on inter-brain interaction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111222218.8A CN114004253A (en) | 2021-10-20 | 2021-10-20 | Information flow estimation method and system based on inter-brain interaction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114004253A true CN114004253A (en) | 2022-02-01 |
Family
ID=79923322
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111222218.8A Pending CN114004253A (en) | 2021-10-20 | 2021-10-20 | Information flow estimation method and system based on inter-brain interaction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114004253A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115841272A (en) * | 2022-11-24 | 2023-03-24 | 北京津发科技股份有限公司 | Team performance prediction method, system and storage medium |
| CN116383487A (en) * | 2023-03-16 | 2023-07-04 | 上海外国语大学 | Information cocoon room identification method based on user retest credibility and group brain consistency |
| CN116509417A (en) * | 2023-04-25 | 2023-08-01 | 上海外国语大学 | Consumer preference consistency prediction method based on nerve similarity |
-
2021
- 2021-10-20 CN CN202111222218.8A patent/CN114004253A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115841272A (en) * | 2022-11-24 | 2023-03-24 | 北京津发科技股份有限公司 | Team performance prediction method, system and storage medium |
| CN116383487A (en) * | 2023-03-16 | 2023-07-04 | 上海外国语大学 | Information cocoon room identification method based on user retest credibility and group brain consistency |
| CN116383487B (en) * | 2023-03-16 | 2023-10-13 | 上海外国语大学 | Information cocoon room identification method based on user retest credibility and group brain consistency |
| CN116509417A (en) * | 2023-04-25 | 2023-08-01 | 上海外国语大学 | Consumer preference consistency prediction method based on nerve similarity |
| CN116509417B (en) * | 2023-04-25 | 2024-03-12 | 上海外国语大学 | Consumer preference consistency prediction method based on nerve similarity |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kumar et al. | Classification of seizure and seizure-free EEG signals using local binary patterns | |
| CN110399857B (en) | Electroencephalogram emotion recognition method based on graph convolution neural network | |
| CN114004253A (en) | Information flow estimation method and system based on inter-brain interaction | |
| George et al. | Recognition of emotional states using EEG signals based on time-frequency analysis and SVM classifier. | |
| CN112932502B (en) | Electroencephalogram emotion recognition method combining mutual information channel selection and hybrid neural network | |
| Brodu et al. | Exploring two novel features for EEG-based brain–computer interfaces: Multifractal cumulants and predictive complexity | |
| CN111523601B (en) | Potential emotion recognition method based on knowledge guidance and generation of countermeasure learning | |
| CN109255366B (en) | Emotional state adjusting system for online learning | |
| CN112244873A (en) | Electroencephalogram time-space feature learning and emotion classification method based on hybrid neural network | |
| Amin et al. | Consumer behavior analysis using EEG signals for neuromarketing application | |
| CN110151203A (en) | Fatigue driving recognition methods based on multistage avalanche type convolution Recursive Networks EEG analysis | |
| An et al. | Electroencephalogram emotion recognition based on 3D feature fusion and convolutional autoencoder | |
| Chen et al. | A novel ensemble deep learning approach for sleep-wake detection using heart rate variability and acceleration | |
| Ye et al. | Identification of mental fatigue levels in a language understanding task based on multi-domain EEG features and an ensemble convolutional neural network | |
| Ullah et al. | Neuromarketing solutions based on EEG signal analysis using machine learning | |
| Xiong et al. | Seizure detection algorithm based on fusion of spatio-temporal network constructed with dispersion index | |
| do Amaral et al. | EEG signal classification in usability experiments | |
| CN114366103A (en) | Attention assessment method and device and electronic equipment | |
| Thurman et al. | Diagnostic spatial frequencies and human efficiency for discriminating actions | |
| Kucukler et al. | EEG dataset for energy data visualizations | |
| CN113269084B (en) | Movie and television play market prediction method and system based on audience group emotional nerve similarity | |
| Guo et al. | Emotion recognition based on multi-modal electrophysiology multi-head attention Contrastive Learning | |
| Kim et al. | A study on user recognition using 2D ECG image based on ensemble networks for intelligent vehicles | |
| Bano | LSTM-Based Model For Human Brain Decisions Using EEG Signals Analysis | |
| Zuo et al. | Driver Distraction Detection Based on EEG Feature Fusion Using Random Forest |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |