CN108932511B - Shopping decision method based on brain-computer interaction - Google Patents
Shopping decision method based on brain-computer interaction Download PDFInfo
- Publication number
- CN108932511B CN108932511B CN201811170120.0A CN201811170120A CN108932511B CN 108932511 B CN108932511 B CN 108932511B CN 201811170120 A CN201811170120 A CN 201811170120A CN 108932511 B CN108932511 B CN 108932511B
- Authority
- CN
- China
- Prior art keywords
- user
- electroencephalogram
- commodities
- concentration
- data
- 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.)
- Active
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
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Artificial Intelligence (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Signal Processing (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a shopping decision method based on computer-computer interaction, which is applied to a commodity recommendation function of an e-commerce platform and is used for judging the commodity likeness of a user based on electroencephalogram data of an article selected by the user on the e-commerce platform, and comprises the following steps: firstly, acquiring electroencephalogram signal data of a user through electroencephalogram equipment, and recording and storing the electroencephalogram signal data; preprocessing by adopting a band-pass filter and an independent component analysis method, and extracting to obtain sample entropy; then obtaining a feature vector based on the sample entropy, identifying different emotions by adopting a naive Bayes method, and setting the excitation level of each different emotion; then, extracting features by adopting an average energy ratio method, and carrying out normalization processing to obtain a concentration degree numerical value; finally, matching degree when the user browses the corresponding commodities is obtained by combining the excitement degree grade and the concentration degree value, a threshold value Th is set, the matching degree and the threshold value Th are compared, and the corresponding commodities are recommended to the user according to the comparison result; the invention improves the shopping experience and shopping efficiency of the user and is convenient to select.
Description
Technical Field
The invention belongs to the field of electroencephalogram signal processing and electronic commerce technology intersection, and particularly relates to a shopping decision method based on brain-computer interaction.
Background
In the age of rapid development of the mobile internet, online shopping becomes the first choice for shopping of many people. Although the online shopping brings convenience to people, the massive commodities are more difficult to decide when people select the commodities, and the favorite commodities are difficult to determine due to the fact that a great deal of time is possibly spent. The classification and identification of human emotions through physiological signals of the human body provides a new approach for solving the above problems.
The brain-computer interface is a pathway established between the brain and external devices to communicate information. The brain-computer interface recognizes the thinking of people by collecting and extracting brain-electrical signals generated by the brain, and accordingly generates control signals to fulfill the aim of information transmission and control between the brain and external equipment. A standard brain-computer interface system can accurately and quickly acquire and identify brain electrical signals of a human brain under various thought activities. Researches on neuroscience, psychology and cognitive science show that a lot of mental activities and cognitive behaviors of people can be reflected through electroencephalogram characteristics. Therefore, the emotion recognition based on the corresponding physiological signals can obtain more objective and real results, and is also suitable for practical application.
According to the method, the electroencephalogram data of the user are collected, processing and analysis are carried out, and the matching degree is obtained by utilizing emotion recognition and concentration degree characteristic extraction, so that a scheme capable of actively detecting the preference of the user is obtained. The scheme can know the preference of the user, and is convenient for the user to better browse the commodities wanted by the user.
Disclosure of Invention
The invention mainly aims to provide a shopping decision method based on brain-computer interaction, which is characterized in that electroencephalogram signal data of a user are collected through electroencephalogram equipment, after processing and analysis, emotion recognition and concentration characteristic extraction are utilized to extract matching degree, commodities are recommended to the user according to the matching degree, the shopping experience of the user is improved, the shopping efficiency is improved, and shopping selection is facilitated; the specific technical scheme is as follows:
a shopping decision method based on computer-computer interaction is characterized in that the likeness of a user to a commodity is judged based on electroencephalogram data of an article selected by the user on a computer-to-computer platform, the method is applied to a commodity recommendation function of the computer-to-computer platform, and the electroencephalogram data are collected through electroencephalogram equipment, and the method comprises the following steps:
s1, acquiring electroencephalogram data of a user when the user browses a certain commodity on an e-commerce platform through electroencephalogram equipment, and recording and storing the data;
s2, preprocessing the electroencephalogram signal data by adopting a band-pass filter and an independent component analysis method, and extracting to obtain sample entropy;
s3, obtaining feature vectors formed by different emotions of a user when the user browses commodities by adopting a sample entropy method based on the sample entropy, identifying different emotions corresponding to the feature vectors by adopting a naive Bayes method, and setting an excitation level rho of each different emotion;
s4, extracting features of the feature vectors by adopting an average energy ratio method, and normalizing the feature vectors to obtain concentration numerical values of the user on browsed commodities;
s5, obtaining the matching degree P when the user browses the corresponding commodities according to a specified algorithm by combining the excitement level rho and the concentration degree value, setting a threshold Th, and comparing the matching degree P with the threshold Th; if the matching degree P is smaller than the threshold Th, removing the commodity corresponding to the matching degree P, otherwise executing the step S6;
and S6, setting a sorting rule to sort the commodities, and recommending the commodity with the highest matching degree P to the user.
Further, before step S1, the method further includes the steps of: and the user opens the e-commerce platform and wears the electroencephalogram equipment on the head.
Furthermore, the EEG signal data comprises 5 frequency band data such as a delta frequency band (1-4 Hz), a theta frequency band (4-8 Hz), an alpha frequency band (8-13 Hz), a beta frequency band (13-30 Hz) and a gamma frequency band (36-44 Hz), and the frequency of the band-pass filter is [1Hz,44Hz ].
Further, the feature vector is composed of the sample entropy composed of electrodes in the electroencephalogram signal data with significant difference; the emotions include four levels of excitement, like, calmness, and dislike, and the excitement level p includes four levels of 3, 2, 1, and 0.
Further, the average energy ratio method is that the high-frequency energy in the electroencephalogram signal data is divided by the low-frequency energy, and the formula is used for dividing the high-frequency energy by the low-frequency energy
And formula
And (c) calculating, wherein,
is the energy value of the Beta rhythm;
is the energy value of Theta rhythm; a. the β/θ (t) is the concentration value for the t-th time; n is a radical of c Is the window length;
is the average of the concentration over the selected window length.
Further, after the feature vector normalization processing in step S4, the expression formula of the concentration value can be obtained:
wherein, AG i Is the current concentration value; ag (i) is the average of the concentration values at time i; ag (min) and ag (max) represent the minimum and maximum values of the concentration value.
Further, the algorithm combining the excitement level ρ and the concentration value in step S5 may be formulated by
Calculating, wherein rho is the excitement level; n is the electroencephalogram data set length; AG i Is a concentration value; th is a set threshold.
Further, the ordering rule includes: sorting according to different styles of the same commodity and sorting according to different brands of the same commodity; and sequencing all the commodities browsed by the user.
According to the shopping decision method based on brain-computer interaction, when a user shops and browses commodities on the internet, electroencephalogram data of different types of commodities browsed by the user are acquired through electroencephalogram equipment, the acquired electroencephalogram data are sequentially subjected to filtering and independent component analysis processing, then a sample entropy method is used for obtaining feature vectors, different emotions corresponding to the different feature vectors are identified by the aid of a naive Bayes method for the feature vectors, and the excitement level of the emotion is set; then, obtaining a concentration degree value of the user by adopting an average energy ratio method and normalization processing based on sample entropy; finally, matching degree is obtained by combining the excitement level and the concentration degree value, the matching degree is compared with a set threshold value, and the commodity love degree of the user is judged according to the comparison result, so that similar commodities can be pushed to the user on the basis of the comparison result for the user to select; compared with the prior art, the invention has the following beneficial effects: according to objective physiological data, the method can evaluate the user experience more finely and accurately, and has positive significance for promoting the development of the e-commerce industry; according to the invention, the electroencephalogram data of the user are collected in real time, the commodities browsed by the user at present can be analyzed in real time, and the timeliness is ensured; the method provided by the invention can actively detect the preference of the user, can facilitate the user to better browse the commodity wanted by the user, and has certain practicability.
Drawings
FIG. 1 is a block diagram illustrating a flow chart of a shopping decision method based on brain-computer interaction according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, 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.
Example one
Referring to fig. 1, in an embodiment of the present invention, a shopping decision method based on computer-computer interaction is provided, where the method determines a user's likeness to a commodity based on electroencephalogram data of an item selected by the user on an e-commerce platform, and is applied to a commodity recommendation function of the e-commerce platform, and the electroencephalogram data is collected by an electroencephalogram device, and the method includes the steps of:
s1, acquiring electroencephalogram signal data of a user browsing a certain commodity on the e-commerce platform through electroencephalogram equipment, and recording and storing the data;
preferably, the non-invasive electroencephalogram equipment is adopted to collect electroencephalogram data of a user browsing commodities when the user inputs the commodities to be purchased on an e-commerce platform, compared with the invasive electroencephalogram equipment, the non-invasive electroencephalogram equipment can be directly worn by the user and the head of the user to collect the electroencephalogram of the scalp, and is non-invasive, convenient to collect the electroencephalogram, high in safety, low in cost and easy to accept by people.
S2, preprocessing the electroencephalogram signal data by adopting a band-pass filter and an independent component analysis method, and extracting to obtain sample entropy;
the invention discloses a method for preprocessing acquired original electroencephalogram signal data, which comprises two methods of band-pass filtering and independent component analysis, wherein a frequency band which is easy to generate interference is filtered out by the band-pass filtering method, and artifacts which are difficult to remove by filtering are removed by the independent component analysis method on the basis of the band-pass filtering, and the specific process comprises the following steps:
firstly, the EEG signal can be divided into 5 main frequency bands of delta frequency band (1 Hz-4 Hz), theta frequency band (4 Hz-8 Hz), alpha frequency band (8 Hz-13 Hz), beta frequency band (13 Hz-30 Hz) and gamma frequency band (36 Hz-44 Hz), and a band-pass filter with the frequency band of 1 Hz-44 Hz is adopted according to the signal frequency to be analyzed for filtering unnecessary signals in the EEG signal data.
Then, removing artifacts which are not easy to be removed by filtering by adopting an independent component analysis method, finding out an interference signal and separating the interference signal from the electroencephalogram signal, wherein the basic process is as follows:
let X be an n-dimensional observation signal vector, and X ∈ R n×1 In X, there are n members, each member being a random variable X ═ X 1 ,x 2 ,x 3 ,...,x n-1 ,x n ) T Wherein x is i Is a random variable; the n random variables in X are mutually dependent, and under a certain assumption, X can be expressed again by linear combination of the n random variables which are mutually independent, namely (X) 1 ,x 2 ,x 3 ,...,x n-1 ,x n ) T ＝A(s 1 ,s 2 ,s 3 ,...,s n-1 ,s n ) T In the formula, s i Is a random variable, and is independent of each other, A is a mixed matrix and full rank, let independent signal source S be S ═ S (S) 1 ,s 2 ,s 3 ,...,s n-1 ,s n ) T When X is AS, and S is A -1 X, let the solution mixing matrix W ═ A -1 Then S ═ WX can be obtained; the values of A, W and S can be estimated under the condition that only X is known through independent component analysis, wherein the solving process of W is as follows:
the probability density function of the ith source signal in the electroencephalogram signal data is assumed to be P i (s), the n source signals at the jth time instant are denoted as a vector s j Then at time j the vector s j Has a combined density of
Then according to x j ＝A·s j And knowledge of the probability density of the complex function is available
Thus, a likelihood function is obtained:
wherein W ═ W 1 ,w 2 ,w 3 ,...,w n-1 ,w n ) T 。
Further taking logarithm to obtain an objective function
And then, solving the partial derivative of W by using knowledge of the maximum likelihood estimation:
thus, in a given training set (x) 1 ,x 2 ,x 3 ,...,x n-1 ,x n ) For a training example x t From the gradient descent formula, one can obtain:
where α is the learning rate.
Finally, W is solved according to the above, thereby separating the independent components to obtain a preprocessed data set.
S3, obtaining feature vectors composed of different emotions of the user when browsing the commodities based on the sample entropy, identifying different emotions corresponding to the feature vectors by adopting a naive Bayes method, and setting an excitation level rho of each different emotion;
the invention adopts sample entropy for feature extraction in emotion recognition, wherein the calculation process of the sample entropy is as follows:
(1) setting original data as u (1), u (2),.. u (N) and N points;
(2) and construct a set of m-dimensional vectors: x m (1),X m (2),...X m (N-m) wherein X m (i)＝[u(i),u(i+1),...,u(i+m-1)](i is 1 to N-m); these vectors represent the values of m u in succession from the point i;
(3) defining the distance between any two m-dimensional vectors as the largest one of the two corresponding elements, i.e. d [ X ] m (i),X m (j)]＝max{|u(i+k)-u(j+k)|,0≤k≤m-1；i,j＝1～N-m,j≠i}；
(4) Setting tolerance threshold r of matching process, and counting d [ X ] for each value of i ≦ N-m m (i),X m (j)]Number N smaller than r × σ m (i) Wherein σ represents the standard deviation of the sequence; the ratio of this number to the total number of distances N-m-1 is then calculated and recorded as
This process is called X m (i) The process of template matching is carried out,
represents any one of X m (j) Probability of matching with a template;
(5) finding X m (i) Average value for all i
(6) Then, the user can use the device to perform the operation,adding the dimension number to 1, namely repeating the steps (2) to (5) for the m + 1-dimensional vector to obtain
The available sample entropy is
When N is a finite value, the above formula is represented as
And when calculating the sample entropy SampEn, selecting three parameters of m, r and N, wherein the specific process of selection is as follows:
wherein m is a window length selected when the sample entropy is calculated, also called an embedding dimension, and m can be 1 or 2 in an actual situation, preferably, m is 2 selected by the method; r is the similarity tolerance between the modes, namely an effective threshold value, is the radius of a hypercube with the dimension of m, and generally, the effective statistical characteristic can be estimated when r is between 0.1 sigma and 0.25 sigma; here, in the present invention, r is preferably 0.2 σ; n is the data length, and particularly, in order to obtain effective statistical characteristics and small pseudo errors, the number of data points is set within 100-5000.
Then, calculating and identifying the corresponding emotion of different electroencephalogram signal data by adopting a naive Bayes algorithm, wherein the specific process comprises the following steps:
suppose there is one marked data set x (i) ,y (i) ]Wherein y is (i) ∈[C 1 ,C 2 ,...,C b ]That is, the data set has b categories; x is a radical of a fluorine atom (i) ＝[x 1 ,x 2 ,...,x n] I.e. there are n input features in total; predicting the value of y for a new sample x, namely classifying x; in particular, in the present invention, the description may be made using a statistical language, that is, when it is observed that the input sample is x, the class y to which it belongs is C k Is represented as P (C) k | x); wherein, C k ∈[C 1 ,C 2 ,...,C b ]In this case, it is only necessary to determine the probabilities of all the b classes, and then to select the C with the highest probability k That is, the category to which x belongs, specifically, the following may be obtained by performing a transformation using bayesian theorem:
for a defined data set, C k P (x) are all fixed values; thus, P (C) can be obtained k |x)∝P(C k )P(x|C k ) Wherein, oc means proportional; then according to the joint probability formula, P (C) can be obtained k )P(x|C k )＝P(C k X); and because x has n eigenvectors, i.e., x ═ x 1 ,x 2 ,...,x n ]Obtaining P (C) k )P(x|C k )＝P(C k ,x)＝P(C k ,x 1 ,x 2 ,...,x n ) (ii) a According to the definition of chain rule and conditional probability, the formula can be further deduced:
P(C k ,x 1 ,x 2 ,...,x n )＝P(x 1 ,x 2 ,...,x n ,C k )
P(x 1 ,x 2 ,...,x n ,C k )＝P(x 1 |x 2 ,...,x n ,C k )P(x 1 ,x 2 ,...,x n ,C k )
P(x 1 ,x 2 ,...,x n ,C k )＝P(x 1 |x 2 ,...,x n ,C k )P(x 2 |x 3 ,...,x n ,C k )…P(x n |C k )P(C k )
based on the formula, P (x) can be obtained according to the principle of condition independence i |x i+1 ,...,x n ,C k )＝P(x i |C k ) (ii) a Thus, the formula P (x) 1 ,x 2 ,...,x n ,C k )＝P(x 1 |x 2 ,...,x n ,C k )P(x 2 |x 3 ,...,x n ,C k )…P(x n |C k )P(C k ) Simplified to P (x) 1 ,x 2 ,...,x n ,C k )＝P(x 1 |C k )P(x 2 |C k )…P(x n |C k )P(C k ) From this, the final naive bayes classification formula is:
wherein, n is a successive multiplication symbol; p (C) k ) Representing the probability of occurrence of each category, and finally taking the C with the highest probability k I.e. the category to which x belongs.
S4, extracting features of the feature vectors by adopting an average energy ratio method, and normalizing the feature vectors to obtain concentration numerical values of the user on browsed commodities;
concentration is the concentration degree of continuous attention, and currently, the related fields generally consider that the concentration degree is related to Alpha rhythm, Beta rhythm and Theta rhythm in electroencephalogram signal data; under the condition that normal people concentrate on attention, the high-frequency energy of the electroencephalogram signal near the forehead can be increased, and the low-frequency energy of the electroencephalogram signal can be reduced, so that the high-frequency energy in the electroencephalogram signal data is divided by the low-frequency energy to measure, specifically, the average energy ratio method is adopted for calculation, and the calculation can be specifically carried out by a formula
And formula
And (3) calculating, wherein,
is the energy value of the Beta rhythm;
is the energy value of Theta rhythm; a. the β/θ (t) is the concentration number for the t-th time, N c Is the window length;
is the average of concentration within the selected window length; and then, after the feature vector normalization processing, obtaining an expression formula of the concentration numerical value:
in the formula, AG i Is the current concentration value; ag (i) is the average of the concentration values at the i-th time; ag (min) and ag (max) represent the minimum and maximum values of the concentration value, thereby obtaining a specific value of the concentration value.
S5, obtaining the matching degree P when the user browses the corresponding commodities according to a specified algorithm by combining the excitement level rho and the concentration degree value, setting a threshold Th, and comparing the matching degree P with the threshold Th; if the matching degree P is smaller than the threshold Th, removing the commodity corresponding to the matching degree P, otherwise executing the step S6; and step S6, setting a sort rule to sort the commodities, and recommending the commodity with the highest matching degree P to the user.
Specifically, the invention combines the excitement level and concentration value obtained in the steps through a formula
Obtaining a matching degree P, wherein rho is the excitement level; n is the electroencephalogram data set length; AG i Is a concentration value; th is a set threshold.
In a specific case, after the matching degree P is compared with the threshold Th, the invention can sort according to different styles of the same commodity; sorting according to different brands of the same commodity; and three sorting rules for sorting all the commodities browsed by the user are used for recommending the corresponding commodities to the user so as to improve the shopping experience of the user.
In order to clearly reflect the specific situation of a user in the process of browsing commodities, the emotion corresponding to the electroencephalogram signals is divided into four classes of excitation, liking, calmness and dislike, and the excitation grade rho is set to be four grades of 3, 2, 1 and 0.
Example two
The shopping decision method based on brain-computer interaction of the present invention will be explained below according to a specific embodiment; firstly, a user logs in a Taobao account and wears electroencephalogram equipment, preferably, a special hat Emotiv Insight with electrodes, which is researched by neuroscience of san Francisco, Calif., is taken as an example for explanation, wherein five electrodes of AF3, AF4, T7, T8 and Pz are arranged on the Emotiv Insight; after the electroencephalogram equipment is worn by a user, the original data of the current electroencephalogram of the user can be extracted in real time, the original electroencephalogram data are transmitted in a wireless Bluetooth mode, and a receiving end receives data by using an official standard USB Dongle; inputting a commodity to be purchased by a user; assuming that a commodity input by a user is trousers, acquiring original electroencephalogram data of five electrode channels, namely AF3, AF4, T7, T8 and Pz of the user through electroencephalogram equipment Emotiv Instrument, and recording an electroencephalogram data set D of the trousers browsed by the user currently, wherein the trousers are Harmony trousers, ninth trousers, casual trousers and sport trousers in the style of the trousers; the brand of trousers is illustrated by Jiumowang, Hailangjia, Senma, Baishidun and seven wolfs.
Then, preprocessing the acquired original electroencephalogram signals, firstly filtering frequency bands which are easy to generate interference by adopting a band-pass filtering method, and removing artifacts which are difficult to remove by filtering by adopting an independent component analysis method; and separating out independent components by an independent component analysis method to obtain a preprocessed data set D'.
Then, after artifact removal and filtering processing are carried out on the electroencephalogram data, five frequency band characteristic data of alpha, beta, gamma, delta and theta are decomposed from five electrode channels by using a sample entropy method, and a characteristic data set D' is obtained.
Then, for the obtained characteristic data set D', the probability of each emotion is calculated by using a naive Bayes method, and if P (excited) > P (like) > P (calm) > P (aversion), the emotion of the current commodity browsed is excited, namely, the excitement grade rho is 3; similarly, the emotions of other commodities can be identified, and if the emotions are like, calm and dislike, the corresponding excitement grades rho are 2, 1 and 0 in sequence; substituting the characteristic data set D' acquired by the sample entropy into the concentration degree value solving formula
And formula
Among them: wherein, the sampling rate of the electroencephalogram equipment is 128HZ, the length of the selected window is 320, and the time is about 2.5 seconds; then adopting a normalization processing formula
After the treatment, the concentration degree value ranges from 0 to 100.
Finally, the obtained data of the excitement level and the concentration degree value are substituted into a matching degree formula
In the method, the matching degree P of the current commodity can be calculated; at this time, if sorting is carried out according to different styles of the same commodities, P (casual pants) > P (sports pants) > P (ninth pants) > P (Harmony pants) can be obtained, and the style of the pants which is the favorite of the user is judged to be casual pants; if sorting is carried out according to different brands of the same commodity, obtaining P (Baishidun) > P (Hailangjia) > P (Jiumowang) > P (Senma) > P (seven wolfs), and judging that the favorite trousers brand of the user is the Baishidun; if the commodities are sorted according to the browsed commodities of the user, P can be obtained 1 ＞P 2 ＞P 3 ＞P 4 ＞P 5 Then it is determined that the favorite article of the user is trousers 1.
According to the shopping decision method based on brain-computer interaction, when a user shops and browses commodities on the internet, electroencephalogram data of different types of commodities browsed by the user are collected through electroencephalogram equipment, the collected electroencephalogram data are subjected to filtering and independent component analysis processing in sequence, a sample entropy method is used for obtaining feature vectors, different emotions corresponding to different feature vectors are recognized through a naive Bayes method for the feature vectors, and the excitation level of the emotion is set; then, obtaining a concentration degree value of the user by adopting an average energy ratio method and normalization processing based on sample entropy; finally, matching degree is obtained by combining the excitement level and the concentration degree value, the matching degree is compared with a set threshold value, and the commodity love degree of the user is judged according to the comparison result, so that similar commodities can be pushed to the user on the basis of the comparison result for the user to select; compared with the prior art, the invention has the beneficial effects that: according to objective physiological data, the method can evaluate the user experience more finely and accurately, and has positive significance for promoting the development of the e-commerce industry; according to the invention, the electroencephalogram data of the user are collected in real time, so that the current browsed commodities of the user can be analyzed in real time, and the timeliness is ensured; the method provided by the invention can realize the detection of the preference and the preference of the user, is convenient for the user to better browse the commodities wanted by the user and has strong practicability.
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments described in the foregoing detailed description, or equivalent arrangements may be substituted for those skilled in the art. All equivalent structures made by using the contents of the specification and the attached drawings of the invention can be directly or indirectly applied to other related technical fields, and are also within the protection scope of the patent of the invention.
Claims (4)
1. A shopping decision method based on computer-computer interaction is characterized in that the method judges the like degree of a user to commodities based on electroencephalogram data of articles selected by the user on an e-commerce platform, and is applied to a commodity recommendation function of the e-commerce platform, and the electroencephalogram data are collected through electroencephalogram equipment, and the method comprises the following steps:
s1, acquiring electroencephalogram signal data of a user browsing a certain commodity on the e-commerce platform through electroencephalogram equipment, and recording and storing the data;
s2, preprocessing the electroencephalogram signal data by adopting a band-pass filter and an independent component analysis method, and extracting to obtain sample entropy;
s3, obtaining feature vectors formed by different emotions of a user when the user browses commodities by adopting a sample entropy method based on the sample entropy, identifying different emotions corresponding to the feature vectors by adopting a naive Bayes method, and setting an excitation level rho of each different emotion;
s4, extracting the features of the feature vectors by adopting an average energy ratio method, and normalizing the feature vectors to obtain the concentration degree value of the user on the browsed commodities;
s5, obtaining the matching degree P when the user browses corresponding commodities according to a specified algorithm by combining the excitement level rho and the concentration degree value, setting a threshold Th, and comparing the matching degree P with the threshold Th; if the matching degree P is smaller than the threshold Th, removing the commodity corresponding to the matching degree P, otherwise executing the step S6;
s6, setting a sorting rule to sort the commodities and recommending the commodity with the highest matching degree P to a user;
the EEG signal data comprises 5 frequency band data of a delta frequency band (1-4 Hz), a theta frequency band (4-8 Hz), an alpha frequency band (8-13 Hz), a beta frequency band (13-30 Hz) and a gamma frequency band (36-44 Hz), and the frequency of the band-pass filter is [1Hz,44Hz ];
the feature vector is composed of the sample entropy composed of electrodes in the electroencephalogram signal data with significant differences; the emotion comprises four levels of excitement, like, calmness and aversion, and the excitement level rho comprises four levels of 3, 2, 1 and 0;
the average energy ratio method is that the high-frequency energy in the electroencephalogram signal data is divided by the low-frequency energy, and the average energy ratio method is represented by a formula
And formula
And (c) calculating, wherein,
is the energy value of the Beta rhythm;
is the energy value of Theta rhythm; a. the β/θ (t) is the concentration number for the t-th time; n is a radical of c Is the window length;
is the average of concentration within the selected window length;
the expression formula of the concentration value can be obtained after the feature vector normalization processing in step S4:
wherein, AG i Is the current concentration value; ag (i) is the average of the concentration values at the i-th time; ag (min) and ag (max) represent the minimum and maximum values of the concentration value.
2. The shopping decision method based on brain-computer interaction as claimed in claim 1, further comprising, before step S1, the steps of: and the user opens the e-commerce platform and wears the electroencephalogram equipment on the head.
3. The shopping decision method based on brain-computer interaction as claimed in claim 1, wherein the algorithm combining the excitement level ρ and the concentration value in step S5 is formulated by formula
Calculating, wherein rho is the excitement level; n is the electroencephalogram data set length; AG i Is the current concentration value; th is a set threshold.
4. The brain-computer interaction based shopping decision method according to claim 1, wherein the ranking rule comprises: sorting according to different styles of the same commodity and sorting according to different brands of the same commodity; and sequencing all the commodities browsed by the user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811170120.0A CN108932511B (en) | 2018-10-09 | 2018-10-09 | Shopping decision method based on brain-computer interaction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811170120.0A CN108932511B (en) | 2018-10-09 | 2018-10-09 | Shopping decision method based on brain-computer interaction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108932511A CN108932511A (en) | 2018-12-04 |
| CN108932511B true CN108932511B (en) | 2022-09-13 |
Family
ID=64444056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811170120.0A Active CN108932511B (en) | 2018-10-09 | 2018-10-09 | Shopping decision method based on brain-computer interaction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108932511B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109961018B (en) * | 2019-02-27 | 2021-03-05 | 易念科技（深圳）有限公司 | Electroencephalogram signal analysis method and system and terminal equipment |
| CN110377160A (en) * | 2019-07-31 | 2019-10-25 | 深圳市宏智力科技有限公司 | A kind of gift delivery system |
| CN113421141A (en) * | 2020-07-29 | 2021-09-21 | 阿里巴巴集团控股有限公司 | Shopping processing method and device based on brain-computer and brain nerve signals and electronic equipment |
| CN112116422B (en) * | 2020-09-10 | 2022-11-04 | 湖南工商大学 | Online shopping system and method based on brain-computer interface |
| CN113053023B (en) * | 2021-03-16 | 2022-08-05 | 广东便捷神科技股份有限公司 | Unmanned vending machine based on artificial intelligence technique |
| CN115587873A (en) * | 2022-11-01 | 2023-01-10 | 广州易海创腾信息科技有限公司 | Artificial intelligence based commodity recommendation system and method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105249963A (en) * | 2015-11-16 | 2016-01-20 | 陕西师范大学 | N400 evoked potential lie detection method based on sample entropy |
| CN108073284A (en) * | 2017-12-15 | 2018-05-25 | 南京信息工程大学 | Purchase system based on brain wave identification mood |
| CN108345676A (en) * | 2018-02-11 | 2018-07-31 | 广东欧珀移动通信有限公司 | Information-pushing method and Related product |
| CN108446635A (en) * | 2018-03-19 | 2018-08-24 | 西北大学 | It is a kind of to obtain Collaborative Filtering Recommendation System and method using EEG signals auxiliary preference |
| CN108537704A (en) * | 2018-04-17 | 2018-09-14 | 深圳市心流科技有限公司 | Classroom evaluating method, device and computer readable storage medium |
-
2018
- 2018-10-09 CN CN201811170120.0A patent/CN108932511B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105249963A (en) * | 2015-11-16 | 2016-01-20 | 陕西师范大学 | N400 evoked potential lie detection method based on sample entropy |
| CN108073284A (en) * | 2017-12-15 | 2018-05-25 | 南京信息工程大学 | Purchase system based on brain wave identification mood |
| CN108345676A (en) * | 2018-02-11 | 2018-07-31 | 广东欧珀移动通信有限公司 | Information-pushing method and Related product |
| CN108446635A (en) * | 2018-03-19 | 2018-08-24 | 西北大学 | It is a kind of to obtain Collaborative Filtering Recommendation System and method using EEG signals auxiliary preference |
| CN108537704A (en) * | 2018-04-17 | 2018-09-14 | 深圳市心流科技有限公司 | Classroom evaluating method, device and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108932511A (en) | 2018-12-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108932511B (en) | Shopping decision method based on brain-computer interaction | |
| Eldele et al. | Time-series representation learning via temporal and contextual contrasting | |
| Yin et al. | Cross-subject EEG feature selection for emotion recognition using transfer recursive feature elimination | |
| Shanir et al. | Automatic seizure detection based on morphological features using one-dimensional local binary pattern on long-term EEG | |
| Kächele et al. | Methods for person-centered continuous pain intensity assessment from bio-physiological channels | |
| US8478542B2 (en) | Non-parametric modeling apparatus and method for classification, especially of activity state | |
| JP2011065504A (en) | Preference prediction server for generating prediction model concerning preference relation of user and method therefor | |
| Agarwal et al. | Classification of alcoholic and non-alcoholic EEG signals based on sliding-SSA and independent component analysis | |
| Goshvarpour et al. | Evaluation of novel entropy-based complex wavelet sub-bands measures of PPG in an emotion recognition system | |
| Soni et al. | Graphical representation learning-based approach for automatic classification of electroencephalogram signals in depression | |
| Ramzan et al. | Learning-based classification of valence emotion from electroencephalography | |
| Al Machot et al. | Improving subject-independent human emotion recognition using electrodermal activity sensors for active and assisted living | |
| Baygin et al. | Automated mental arithmetic performance detection using quantum pattern-and triangle pooling techniques with EEG signals | |
| Gannouni et al. | Electroencephalography based emotion detection using ensemble classification and asymmetric brain activity | |
| Mohdiwale et al. | Investigating Feature Ranking Methods for Sub‐Band and Relative Power Features in Motor Imagery Task Classification | |
| Álvarez Casado et al. | Meditation detection using sensors from wearable devices | |
| Wankhade et al. | IKKN predictor: An EEG signal based emotion recognition for HCI | |
| Hoffmann et al. | Application of the evidence framework to brain-computer interfaces | |
| Can | Stressed or just running? Differentiation of mental stress and physical activityby using machine learning | |
| Hajian et al. | Quantification of depression disorder using EEG signal | |
| Rahman et al. | Affective state recognition through EEG signals feature level fusion and ensemble classifier | |
| Thejaswini et al. | EEG Classification Using Modified KNN Algorithm | |
| Ilyas et al. | An efficient emotion recognition frameworks for affective computing | |
| Wu et al. | Robust EEG-based emotion recognition using multi-feature joint sparse representation | |
| de Carvalho et al. | A combined Fourier analysis and support vector machine for EEG classification. |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |