Potential cargo anchor prediction method and device based on neuroimaging psychological topic analysis

Abstract

The invention discloses a potential cargo anchor prediction method and device based on neural imaging psychological topic analysis, relates to the technical field of cargo anchor potential prediction, and solves the problems of relatively high subjectivity, long evaluation period and high trial and error cost of cargo anchor potential evaluation in the prior art. The method comprises the following steps: obtaining functional magnetic resonance imaging signals and behavior data of tested audiences, information of goods in a goods carrying video and market performance data of a goods carrying anchor, and analyzing the preprocessed functional magnetic resonance imaging signals to obtain an average brain activation image, individual psychological theme score data are generated and calculated to obtain psychological theme score data of each anchor carrying goods, a data set is constructed and trained to obtain a market performance prediction model and apply the model, and total consumer feedback of the market is predicted through data of laboratory small sample testees. The efficiency, accuracy and scientificity of potential cargo anchor prediction are improved, and the trial and error cost is remarkably reduced.

Description

Technical Field

[0001] This application relates to the technical field of potential prediction of live streaming salespersons, and in particular to a method and device for predicting potential live streaming salespersons based on neuroimaging psychological theme analysis. Background Art

[0002] As one of the core forms of digital marketing, e-commerce live streaming has shown great development potential in China, and the market scale has continued to grow. In this field, the role of live streaming salespersons is crucial, and they have a key impact on product sales and brand image improvement. However, how to accurately predict the market performance of live streaming salespersons and screen out potential live streaming salespersons is still a challenging task.

[0003] Existing technologies for predicting the market performance of live streaming salespersons are difficult to meet the needs of the e-commerce industry for efficient and accurate screening of live streaming salespersons, and cannot truly reflect the actual reactions of consumers to live streaming salespersons during e-commerce live streaming, resulting in limited reference value of prediction conclusions. There are mainly two existing methods for predicting potential live streaming salespersons. One is to identify the key personal traits of successful live streaming salespersons through interviews and questionnaires, and use them to evaluate the live streaming salespersons to be predicted, such as appearance attractiveness and behavior performance. This method has the following defects: (1) Subjective bias: relying on subjective reports of evaluators, it is easily affected by subjective cognition, resulting in insufficient data reliability; (2) Single data dimension: only relying on explicit behavior scoring data, it is difficult to comprehensively reflect consumers' true preferences for live streaming salespersons. The other is to use historical and real-time data to predict live streaming sales, for example, by analyzing the historical live streaming data and real-time behavior data of live streaming salespersons, constructing a machine learning model to predict sales performance. This method has the following defects: (1) Insufficient timeliness: it is necessary to accumulate market data through a period of trial broadcasts, and it is impossible to quickly evaluate the potential at the initial stage of live streaming salespersons, and the prediction has a lag; (2) High trial-and-error cost: brand parties need to invest a large amount of resources in pre-tests to accumulate enough data for model training. Summary of the Invention

[0004] The purpose of this application is to overcome the problems of strong subjectivity, long evaluation cycle, and high trial-and-error cost in the evaluation of the potential of live streaming salespersons in the prior art, and provide a method and device for predicting potential live streaming salespersons based on neuroimaging psychological theme analysis.

[0005] In a first aspect, a method for predicting potential live streaming salespersons based on neuroimaging psychological theme analysis is provided, including:

[0006] Obtaining functional magnetic resonance imaging signals and behavioral data during the process of multiple test audiences watching static faces of live streaming salespersons and dynamic e-commerce live streaming video segments, obtaining product information of the products sold in the corresponding dynamic e-commerce live streaming video segments, and obtaining market performance data of the corresponding live streaming salespersons;

[0007] Preprocess the market performance data and functional magnetic resonance imaging signals;

[0008] Perform a group analysis on the preprocessed functional magnetic resonance imaging signals to obtain the average brain activation image of the test audience during the viewing of the dynamic live streaming video clip;

[0009] Map the brain activation image to the brain region activation statistical chart related to the psychological theme in the neuroscience meta-analysis database to generate the individual psychological theme score data of a single test audience for a single live streaming salesperson;

[0010] Perform an inter-group average calculation on the individual psychological theme score data of all test audiences and output the psychological theme score data of each live streaming salesperson;

[0011] Make the psychological theme score data, behavioral data, live streaming sales product information, and market performance data into a data set;

[0012] Use the data set to train a neural network model to obtain a market performance prediction model;

[0013] Use the market performance prediction model to predict the market performance of the live streaming salesperson.

[0014] In some possible implementation manners, it further includes: performing an interactive operation on the prediction result for visual display, where the interactive operation includes rotation and scaling of the brain activation image, screening and comparison of the scoring dimensions, and export and sharing of the prediction report, and the prediction result is visually displayed in the form of a multi-dimensional chart.

[0015] In some possible implementation manners, the behavioral data includes scores for the facial attractiveness of the live streaming salesperson, scores for the willingness to watch the live stream, scores for the willingness to purchase the live streaming sales product, and the duration of watching the live streaming video clip.

[0016] In some possible implementation manners, the market performance data includes the basic information of the live streaming salesperson, the fan portrait, the data performance of the live stream where the intercepted live streaming video clip is located, the live streaming sales product information, and the commercial success evaluation index.

[0017] In some possible implementation manners, the preprocessing includes data format conversion, extraction of brain images, head motion correction, temporal slice correction, spatial registration, spatial normalization, spatial smoothing, and filtering.

[0018] In some possible implementation manners, training a neural network model using the dataset to obtain a market performance prediction model includes: using the psychological theme scoring data and behavioral data as independent variables, the market performance data as the dependent variable, introducing the information of the goods sold in the live broadcast as a covariate control, and fitting the model using the least squares method or the ridge regression algorithm to optimize the weight parameters. The calculation formula of the market performance prediction model is:

[0019]

[0020] where Y is the market performance data, X （psy） is the psychological theme scoring data, X （beh） is the behavioral data of the test audience, C （prod） is the information of the goods sold in the live broadcast, , and are the weight parameters, is the random error term;

[0021] Dividing the dataset into a training set, a validation set, and a test set to train, validate, and test the neural network model to obtain a market performance prediction model.

[0022] In some possible implementation manners, the predicted data of the market performance potential of the live broadcast host output by the market performance prediction model includes the predicted values and corresponding confidence intervals of the average stay duration per person in the live broadcast room, the sales amount, and the sales conversion rate, as well as the psychological theme score.

[0023] In a second aspect, a potential live broadcast host prediction device based on neuroimaging psychological theme analysis is provided, including:

[0024] A data acquisition module, configured to acquire functional magnetic resonance imaging signals and behavioral data of multiple test audiences during the process of watching static faces and dynamic live broadcast video segments of live broadcast hosts, acquire the information of the goods sold in the corresponding dynamic live broadcast video segments, and acquire the market performance data of the corresponding live broadcast hosts;

[0025] A preprocessing module, configured to preprocess the market performance data and the functional magnetic resonance imaging signals;

[0026] An analysis module, configured to perform a group analysis on the preprocessed functional magnetic resonance imaging signals to obtain an average brain activation image of the test audience during the process of watching the dynamic live broadcast video segments;

[0027] A psychological theme mapping module, configured to map the brain activation image to a brain region activation statistical chart related to psychological themes in a neuroscience meta-analysis database to generate individual psychological theme scoring data of a single test audience for a single live broadcast host;

[0028] A calculation module for performing between-group averaging calculations on the individual psychological theme scoring data of all participating audiences and outputting the psychological theme scoring data of each live-streaming salesperson;

[0029] A dataset construction module for making the psychological theme scoring data, behavior data, live-streaming sales product information, and market performance data into a dataset;

[0030] A model construction and training module for training a neural network model using the dataset to obtain a market performance prediction model;

[0031] A prediction module for predicting the market performance of live-streaming salespersons using the market performance prediction model.

[0032] In a third aspect, a computer-readable storage medium is provided. The computer-readable medium stores program code for a device to execute, and the program code includes steps for executing the method in any implementation manner in the first aspect as described above.

[0033] In a fourth aspect, an electronic device is provided. The electronic device includes a processor, a memory, and a program or instruction stored on the memory and executable on the processor. When the program or instruction is executed by the processor, the method in any implementation manner in the first aspect as described above is implemented.

[0034] The present application has the following beneficial effects:

[0035] 1. Break through the limitations of traditional prediction technologies: By integrating consumers' subconscious neural responses to live-streaming salespersons with explicit behavior data, a multi-dimensional prediction model is constructed, effectively overcoming the defects of existing technologies that rely on subjective reports or lagging market data, and significantly improving the objectivity and accuracy of prediction results.

[0036] 2. Realize early potential assessment and cost control: The overall market feedback can be predicted based on small-sample neural data in the laboratory without relying on a large amount of data accumulated from trial broadcasts, solving the problems of insufficient timeliness and high trial-and-error costs in existing technologies, and helping to quickly screen high-potential candidates at the initial stage of recruiting live-streaming salespersons.

[0037] 3. Provide interpretable decision-making support: Quantify the ability of live-streaming salespersons to mobilize consumers' psychological responses through psychological theme scoring (such as the activation intensity of brain maps related to emotion, language, and memory), provide interpretable neural evidence for enterprises, and assist in optimizing the training strategies of live-streaming salespersons. Description of the Drawings

[0038] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application.

[0039] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0040] Figure 1 is a flowchart of a potential live-streaming salesperson prediction method based on neuroimaging psychological theme analysis in Embodiment 1 of the present application;

[0041] Figure 2 is a structural block diagram of a potential live-streaming salesperson prediction device based on neuroimaging psychological theme analysis in Embodiment 2 of the present application;

[0042] Figure 3 is an internal structural schematic diagram of an electronic device in Embodiment 4 of the present application.

[0043] Reference numerals:

[0044] 100, data acquisition module; 200, preprocessing module; 300, analysis module; 400, psychological theme mapping module; 500, calculation module; 600, dataset construction module; 700, model construction and training module; 800, prediction module. Detailed implementation manners

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the protection scope of the present invention.

[0046] Embodiment 1

[0047] As Figure 1 shown, a potential live-streaming salesperson prediction method based on neuroimaging psychological theme analysis involved in Embodiment 1 of the present application includes:

[0048] S100. Obtain functional magnetic resonance imaging signals and behavioral data of multiple test audiences during the process of watching static faces of live-streaming salespersons and dynamic live-streaming sales video segments, obtain product information in the corresponding dynamic live-streaming sales video segments, and obtain market performance data of the corresponding live-streaming salespersons;

[0049] The static face images of each live commerce host include four face images with a resolution of 400*400 pixels. These images are cropped from different live commerce broadcasts or short videos of the host, ensuring differences in elements such as the background and actions. In the images, the host is preferably looking directly at the camera, without obvious eye line deviation, and not wearing accessories that cover the face such as glasses. No exaggerated special effects or filters are used in the images to ensure the natural expression of the host and reflect their true image.

[0050] The dynamic live commerce video of each host is a 90-second video clip of live commerce. These live commerce video clips are recorded and cropped from the host's actual live broadcasts according to the following unified standards to control irrelevant variables as much as possible: In the live commerce video clips, the host's face should be kept as much as possible in the center of the screen, and the front appearance of other auxiliary personnel in the live team should be avoided. The live commerce video clips should preferably not contain distracting text or advertising logos. In the 90-second live commerce video clips, the content should be concentrated as much as possible on the host's explanation of a single product, and the product brand should be hidden to exclude the potential influence of brand effects on consumers. The video resolution should be kept as high-definition as possible to ensure the presentation effect of the experimental stimuli.

[0051] The behavioral data of the test audience includes scores for the attractiveness of the host's face, scores for the willingness to watch the live broadcast, scores for the willingness to purchase the promoted product, and the duration of watching the live commerce video clips, etc.

[0052] The market performance data of the host includes the host's basic information (name or stage name, gender, number of fans, etc.), fan portraits, data performance of the live broadcast where the intercepted live video clip is located (live broadcast duration, product explanation duration, minute traffic acquisition ability, average stay duration, etc.), promoted product information of the live broadcast (product category, brand, specifications, price, etc.), and commercial success evaluation indicators of the live broadcast (product conversion rate, sales volume, sales amount, etc.).

[0053] S200. Preprocess the market performance data and functional magnetic resonance imaging signals;

[0054] Preprocessing the market performance data of the host includes: removing duplicate data, handling missing values and outliers, unifying data formats, correcting data errors, etc., to ensure data integrity and accuracy.

[0055] For some data with a large degree of variation among different hosts (such as the number of fans, product price), data standardization (such as taking logarithms) can be performed to eliminate the influence of different magnitudes, and then the importance of each data dimension can be evaluated more equally.

[0056] Preprocessing the functional magnetic resonance imaging signals of the test audience includes:

[0057] Data format conversion: Convert the format of the original magnetic resonance images, preserving the three-dimensional spatial coordinates and time series information to be compatible with subsequent data analysis tools.

[0058] Extraction of brain images: Adopt a brain tissue extraction algorithm based on threshold segmentation to remove non-brain tissues such as the skull and scalp, generating a mask image containing only the brain parenchyma to eliminate the interference of non-brain tissue signals on the analysis of functional images, reduce the computational amount of invalid data, and improve the accuracy and efficiency of subsequent brain region activation detection.

[0059] Head motion correction: Calculate the head translation and rotation parameters at each time point through rigid transformation, align all time series images to the reference position of the first frame, and eliminate abnormal data with excessive displacement or rotation to eliminate the spatial misregistration artifacts caused by the head movement of the test audience and provide a stable spatio-temporal reference for subsequent analysis.

[0060] Time slice correction: For the time sequence deviation caused by multi-slice interleaved acquisition, synchronize the signal time points of each slice to the same time axis to eliminate the time misregistration between slices and avoid activation detection errors caused by the acquisition delay between slices.

[0061] Spatial registration: Perform rigid body registration on the functional image and the high-resolution structural image (T1-weighted) to optimize the spatial anatomical correspondence of the functional image.

[0062] Spatial normalization: Use a non-linear transformation algorithm to register the individual structural image to the standard space and apply the same transformation parameters to the functional image to achieve spatial alignment of brain regions across test audiences and provide a stable spatial reference system for subsequent psychological theme scoring analysis.

[0063] Spatial smoothing: Perform convolution processing and smoothing on the functional image to improve the signal-to-noise ratio and meet the assumptions of the random field theory for statistical analysis, suppress the random noise of single voxels, and enhance the statistical test power of neural activation clusters.

[0064] Filtering: Apply frequency domain band-pass filtering to suppress high-frequency physiological noises (such as heartbeat and respiration) and ultra-low frequency drift signals, and retain the frequency bands related to the blood oxygenation-dependent (BOLD) response to improve the reliability of neural signal extraction and recognition.

[0065] S300. Perform group analysis on the preprocessed functional magnetic resonance imaging signals. Among them, taking the live streamers with goods as units, the test audiences watching the same live streamer with goods are grouped as one group. If the same test audience watches different live streamers with goods, they will be assigned to different groups. Group analysis refers to analyzing the preprocessed functional magnetic resonance imaging signals of the test audiences within the group to obtain the average brain activation images of the test audiences during the period of watching the dynamic live stream video clips with goods.

[0066] S400. Map the brain activation image to the statistical chart of brain region activations related to psychological themes in the neuroscience meta-analysis database to generate individual psychological theme score data for a single viewer of a single live-streaming product promoter;

[0067] Specifically, based on the neuroscience meta-analysis database Neurosynth, load the statistical chart of brain region activations related to psychological themes (such as emotion, language, memory, etc.), perform spatial matching between the brain activation image of the viewer and each psychological theme template, and calculate the weighted value of the brain region activation intensity according to the formula:

[0068]

[0069] Among them, is the template voxel weight (determined by the threshold of the statistical chart of brain region activations in the meta-analysis), is the blood oxygenation level-dependent (BOLD) signal intensity value of each voxel of the viewer.

[0070] For the continuous neural signals of the viewer watching the dynamic live-streaming video of the product promoter, use the rolling window method to divide the playing duration of the live-streaming segment into multiple time windows, extract the neural signals in this time period for each window, generate a dynamic psychological theme score sequence, and track the dynamic psychological reactions and neural signals during the viewing process.

[0071] S500. Perform between-group averaging calculation on the individual psychological theme score data of all viewers and output the psychological theme score data of each live-streaming product promoter;

[0072] For each live-streaming product promoter, summarize the individual score data of all viewers in each psychological theme dimension (such as emotion, language, and memory, etc.) to form a multi-dimensional score matrix.

[0073] Perform arithmetic averaging on the individual scores of each psychological theme dimension. The calculation formula is:

[0074]

[0075] Among them, is the total number of viewers, represents the th viewer's score in this psychological theme.

[0076] S600. Make the psychological theme score data, behavioral data, product information of the promoted products, and market performance data into a data set. Among them, use the psychological theme score data, behavioral data, and product information of the promoted products as model input data, and use the market performance data as output data;

[0077] S700. Train a neural network model using the dataset to obtain a market performance prediction model;

[0078] Specifically, using the psychological theme score data and behavioral data as independent variables, and the market performance data as the dependent variable, introducing the information of the goods sold in the live broadcast (such as product category, brand, specification, price, etc.) as a covariate control, and using the least squares method or ridge regression algorithm to fit the model and optimize the weight parameters.

[0079] The calculation formula of the market performance prediction model is:

[0080]

[0081] Among them, Y is the market performance data, Y is the dependent variable, including L≥1 observation indicators (such as click-through rate, conversion rate, etc.), and Y represents the output result of the neural prediction model; X （psy） is the psychological theme score data, X （psy） is the independent variable, including K≥1 neurocognitive dimension scores, and X （psy） represents the potential influence of the live-streaming host on the psychological state of the audience; X （beh） is the behavioral data of the test audience, X （beh） is the independent variable, including M≥1 real-time behavior observation indicators), and X （beh） represents the obvious influence of the live-streaming host on the behavior of the audience; C （prod） is the information of the goods sold in the live broadcast, C （prod） is the covariate, including N≥1 product attribute characteristics, and C （prod） controls the inherent influence of the product itself on the market performance; and and are the weight parameters; is the random error term.

[0082] Since the correlations between multiple psychological themes and market performance indicators are analyzed simultaneously, in order to control false positives, multiple comparison correction (such as FDR correction) is used to ensure that only those themes that are significantly related to the predicted market performance of the live-streaming host are included in the further analysis.

[0083] The specific method for model validation is: calculate the Pearson correlation coefficient, root mean square error (RMSE), and permutation test significance between the predicted value and the true value on the test set.

[0084] S800. Use the market performance prediction model to predict the market performance of the live-streaming host.

[0085] Specifically, obtain the functional magnetic resonance imaging signals and behavioral data of the test audience when watching the static face pictures and dynamic live streaming videos of the live streamers to be evaluated. Calculate the psychological theme scores based on the neural signals, and input the psychological theme scores and behavioral data into the prediction model of the market performance of the live streamers to obtain the predicted data of the market performance potential of the live streamers. The predicted data of the market performance potential of the live streamers includes the predicted estimates and corresponding confidence intervals of the average stay time per person in the live room, sales volume, and sales conversion rate, as well as the psychological theme scores.

[0086] In a further embodiment, it further includes: a visual display for interactive operation of the prediction results, where the interactive operations include rotation and scaling of the brain activation images, screening and comparison of the scoring dimensions, and export and sharing of the prediction reports. The prediction results are visually displayed in the form of a multi-dimensional chart, including:

[0087] Combination of bar chart and line chart: Compare the predicted market performance indicators with the industry benchmark levels;

[0088] Psychological theme ability radar chart: Display the scoring distribution of the live streamers in each psychological theme dimension, and quantify the ability of the live streamers to mobilize consumers' psychological reactions;

[0089] Dynamic neural activation heat map: Generate a curve graph of the psychological theme scores changing with time based on the neural signals segmented by time windows, and perform temporal correlation annotation with the key events of the corresponding video segments (such as: product explanation, promotional terms, etc.).

[0090] Embodiment 2

[0091] As Figure 2 shown, a potential live streamer prediction device based on neuroimaging psychological theme analysis according to Embodiment 2 of the present application includes:

[0092] A data acquisition module 100, configured to acquire the functional magnetic resonance imaging signals and behavioral data of multiple test audiences during the process of watching the static faces of the live streamers and the dynamic live streaming video segments, acquire the information of the products sold in the corresponding dynamic live streaming video segments, and acquire the market performance data of the corresponding live streamers;

[0093] A preprocessing module 200, configured to preprocess the market performance data and the functional magnetic resonance imaging signals;

[0094] An analysis module 300, configured to perform a group analysis on the preprocessed functional magnetic resonance imaging signals to obtain the average brain activation image of the test audience during the process of watching the dynamic live streaming video segments;

[0095] The psychological theme mapping module 400 is used to map the brain activation images to the statistical charts of brain region activations related to psychological themes in the neuroscience meta-analysis database, generating individual psychological theme score data of a single test audience for a single live-streaming salesperson;

[0096] The calculation module 500 is used to perform between-group averaging calculations on the individual psychological theme score data of all test audiences and output the psychological theme score data of each live-streaming salesperson;

[0097] The dataset construction module 600 is used to make the psychological theme score data, behavioral data, live-streaming sales product information, and market performance data into a dataset;

[0098] The model construction and training module 700 is used to train a neural network model using the dataset to obtain a market performance prediction model;

[0099] The prediction module 800 is used to predict the market performance of live-streaming salespersons using the market performance prediction model.

[0100] It should be noted that for other specific implementation manners of the potential live-streaming salesperson prediction device based on neuroimaging psychological theme analysis in this embodiment, reference can be made to the specific implementation manners of the potential live-streaming salesperson prediction method based on neuroimaging psychological theme analysis above. To avoid redundancy, it will not be elaborated here.

[0101] Embodiment 3

[0102] A computer-readable storage medium involved in Embodiment 3 of the present application, where the computer-readable medium stores program codes for device execution, and the program codes include steps for executing the method in any one of the implementation manners in Embodiment 1 of the present application;

[0103] Among them, the computer-readable storage medium can be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM); the computer-readable storage medium can store program codes, and when the program stored in the computer-readable storage medium is executed by a processor, the processor is used to execute the steps of the method in any one of the implementation manners in Embodiment 1 of the present application.

[0104] Embodiment 4

[0105] As Figure 3 shown, an electronic device involved in Embodiment 4 of the present application, where the electronic device includes a processor, a memory, and a program or instruction stored on the memory and executable on the processor, and when the program or instruction is executed by the processor, it implements the method in any one of the implementation manners in Embodiment 1 of the present application;

[0106] Among them, the processor can be a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), a graphics processing unit (GPU), or one or more integrated circuits, which are used to execute relevant programs to implement the method in any one of the implementation manners in Embodiment 1 of the present application.

[0107] The processor can also be an integrated circuit electronic device with signal processing capabilities. During the implementation process, each step of the method in any one of the implementation manners in Embodiment 1 of the present application can be completed by the integrated logic circuit in the hardware of the processor or the instructions in the form of software.

[0108] The above-mentioned processor can also be a general-purpose processor, a digital signal processor, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as being completed by the hardware decoding processor, or completed by a combination of the hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the functions required to be executed by the units included in the data processing device of the embodiments of the present application, or execute the method in any one of the implementation manners in Embodiment 1 of the present application.

[0109] The above is only a preferred specific implementation manner of the present application; however, the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application, according to the technical solution of the present application and its improved concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present application.

Claims

1. A potential live-streaming product promoter prediction method based on neuroimaging psychological theme analysis, characterized in that, Including: Obtaining functional magnetic resonance imaging signals and behavioral data of multiple test audiences during the process of watching the static faces of live commerce hosts and dynamic live commerce live video clips, obtaining the information of the live commerce products in the corresponding dynamic live commerce live video clips, and obtaining the market performance data of the corresponding live commerce hosts; Preprocessing the market performance data and the functional magnetic resonance imaging signals; Performing group analysis on the preprocessed functional magnetic resonance imaging signals to obtain the average brain activation images of the test audiences during the process of watching the dynamic live commerce live video clips; Mapping the brain activation images to the brain region activation statistical charts related to psychological themes in the neuroscience meta-analysis database to generate the individual psychological theme score data of a single test audience for a single live commerce host; Performing between-group averaging calculation on the individual psychological theme score data of all test audiences and outputting the psychological theme score data of each live commerce host; Making the psychological theme score data, behavioral data, live commerce product information, and market performance data into a data set; Training a neural network model using the data set to obtain a market performance prediction model; Predicting the market performance of live commerce hosts using the market performance prediction model.

2. The potential live-streaming salesperson prediction method based on neuroimaging psychological theme analysis according to claim 1, wherein Also including: Performing an interactive operation-based visual display on the prediction results, where the interactive operations include the rotation and scaling of the brain activation images, the screening and comparison of the scoring dimensions, and the export and sharing of the prediction reports, and the prediction results are visually displayed in the form of a multi-dimensional chart.

3. The potential live-streaming salesperson prediction method based on neuroimaging psychological theme analysis according to claim 1, characterized in that, The behavioral data includes the scores for the facial attractiveness of the live commerce host, the scores for the willingness to watch the live broadcast, the scores for the willingness to purchase the live commerce products, and the duration of watching the live commerce live video clips.

4. The potential live-streaming product-selling anchor prediction method based on neuroimaging psychological theme analysis according to claim 1, wherein, The market performance data includes the basic information of the live commerce host, the fan portrait, the data performance of the live broadcast where the intercepted live commerce live video clip is located, the live commerce product information, and the commercial success evaluation indicators.

5. The potential live-streaming product promotion anchor prediction method based on neuroimaging psychological theme analysis according to claim 1, wherein, The preprocessing includes data format conversion, brain image extraction, head motion correction, time slice correction, spatial registration, spatial normalization, spatial smoothing, and filtering.

6. The potential live-streaming product-selling anchor prediction method based on neuroimaging psychological theme analysis according to claim 1, wherein Training a neural network model using the data set to obtain a market performance prediction model, including: using the psychological theme score data and behavioral data as independent variables, the market performance data as the dependent variable, introducing the information of the live commerce products in the live broadcast as a covariate control, and fitting the model using the least squares method or the ridge regression algorithm to optimize the weight parameters. The calculation formula of the market performance prediction model is: ； Among them, Y is market performance data, X （psy） is psychological theme scoring data, X （beh） is the behavioral data of the tested audience, C （prod） is the information of the goods sold during the live broadcast, 、 and are weight parameters, is the random error term; Dividing the data set into a training set, a validation set, and a test set to train, validate, and test the neural network model to obtain a market performance prediction model.

7. The potential live-streaming product-selling anchor prediction method based on neuroimaging psychological theme analysis according to claim 1, wherein The predicted data of the market performance potential of the live commerce hosts output by the market performance prediction model includes the predicted estimates of the average residence time per person in the live broadcast room, the sales volume, the sales conversion rate, and the corresponding confidence intervals, as well as the psychological theme scores.

8. A potential live-streaming salesperson prediction device based on neuroimaging psychological theme analysis, characterized in that, Including: A data acquisition module for obtaining functional magnetic resonance imaging signals and behavioral data of multiple test audiences during the process of watching the static faces of live commerce hosts and dynamic live commerce live video clips, obtaining the information of the live commerce products in the corresponding dynamic live commerce live video clips, and obtaining the market performance data of the corresponding live commerce hosts; A preprocessing module for preprocessing the market performance data and functional magnetic resonance imaging signals; An analysis module for performing group analysis on the preprocessed functional magnetic resonance imaging signals to obtain an average brain activation image of the test audience during the viewing of dynamic live streaming video clips of product promotion; A psychological theme mapping module for mapping the brain activation image to a statistical chart of brain region activations related to psychological themes in a neuroscience meta-analysis database to generate individual psychological theme score data of a single test audience for a single product promotion anchor; A calculation module for performing between-group averaging calculation on the individual psychological theme score data of all test audiences and outputting the psychological theme score data of each product promotion anchor; A data set construction module for making a data set from the psychological theme score data, behavioral data, product promotion information, and market performance data; A model construction and training module for training a neural network model using the data set to obtain a market performance prediction model; A prediction module for predicting the market performance of a product promotion anchor using the market performance prediction model.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program code for execution by a device, and the program code includes steps for executing the method according to any one of claims 1-7.

10. An electronic device, characterized in that, The electronic device includes a processor, a memory, and a program or instruction stored on the memory and executable on the processor. When the program or instruction is executed by the processor, the method according to any one of claims 1-7 is implemented.

Images