Prediction analysis method and device for emission spectrum curve

Abstract

The embodiment of the invention relates to a prediction analysis method and device for an emission spectrum curve. The method comprises the following steps: constructing an emission intensity prediction model for predicting emission intensity; training an emission intensity prediction model based on a first data set constructed through data acquisition; after training is finished, task types and task configuration parameters are extracted from prediction analysis tasks input by the user; analyzing three types of curve characteristics (the maximum peak site, the full width at half maximum and the wavelength range at half maximum) of an emission spectrum prediction curve of a specified molecule under the condition that the temperature is not changed but the illumination intensities are different if the type is the first type; if the type is the second type, analyzing three types of curve characteristics of the specified molecules under the condition that the illumination intensity is not changed but the temperatures are different; if the type is the third type, analyzing three types of curve characteristics of the two specified molecules under the condition that the temperature and the illumination intensity are not changed; the prediction analysis efficiency can be improved, and the data analysis convenience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method and device for predicting and analyzing emission spectral curves. Background Art

[0002] For organic photovoltaic (OPV) materials applied to flexible light-emitting devices, it is very necessary to predict and analyze the emission spectral curves of each material molecule. The emission spectral curve mentioned here is a two-dimensional curve with the emission intensity (or simply intensity) as the ordinate and the wavelength as the abscissa. Currently, most of the prediction and analysis methods for organic molecular emission spectral curves are implemented based on quantum chemical calculation means (such as density functional theory calculation, etc.). However, from practical experience, this conventional prediction and analysis method has problems such as high computational complexity, long processing time, and low processing efficiency. Summary of the Invention

[0003] The object of the present invention is to provide a method, device, electronic device, and computer-readable storage medium for predicting and analyzing emission spectral curves in view of the defects of the prior art. The present invention customizes an emission intensity prediction model for predicting the emission intensity according to the molecular structure, temperature, light intensity, and wavelength input into the model and trains it; and after the training is completed, it parses the prediction and analysis tasks (composed of task types and task configuration parameters) input by the user; if the task type is the first type, it predicts multiple emission spectral curves of a specified molecule based on the prediction model and task configuration parameters under the condition of the same temperature but different light intensities and analyzes three types of curve characteristics (maximum peak position, full width at half maximum, and half-peak wavelength range) of each prediction curve to obtain a first analysis data table and feedback it to the user; if it is the second type, it predicts multiple emission spectral curves of a specified molecule based on the prediction model and task configuration parameters under the condition of the same light intensity but different temperatures and analyzes three types of curve characteristics of each prediction curve to obtain a second analysis data table and feedback it to the user; if it is the third type, it predicts the emission spectral curves of two specified molecules respectively based on the prediction model and task configuration parameters under the condition that both the temperature and the light intensity remain unchanged and analyzes three types of curve characteristics of the two prediction curves to obtain a third analysis data table and feedback it to the user. On the one hand, the present invention can achieve the technical purpose of reducing the processing complexity of the prediction and analysis task, shortening the processing time, and improving the processing efficiency through a customized emission intensity prediction model; on the other hand, it can achieve the purpose of improving the convenience of data analysis through three types of customized analysis means.

[0004] To achieve the above object, in the first aspect of the embodiments of the present invention, a method for predicting and analyzing emission spectral curves is provided, and the method includes:

[0005] Construct an emission intensity prediction model for predicting emission intensity; the emission intensity prediction model is used to perform emission intensity prediction processing based on the molecular structure M, temperature x, light intensity y, and wavelength z input into the model and output the corresponding predicted emission intensity u;

[0006] Construct a first data set for model training through data collection; and perform model training on the emission intensity prediction model according to the first data set;

[0007] After the model training is completed, receive the prediction analysis task input by the user and extract the corresponding task type and task configuration parameters therefrom; the prediction analysis task includes the task type and the task configuration parameters; the task type includes the first, second, and third types; when the task type is the first type, the corresponding task configuration parameters include the first molecular structure, the first set temperature, the first light intensity sequence, and the first wavelength sequence; when the task type is the second type, the corresponding task configuration parameters include the second molecular structure, the first set light intensity, the first temperature sequence, and the second wavelength sequence; when the task type is the third type, the corresponding task configuration parameters include the third molecular structure, the fourth molecular structure, the second set temperature, the second set light intensity, and the third wavelength sequence;

[0008] If the task type is the first type, under the set conditions where multiple temperatures are kept consistent but the light intensities are different, based on the emission intensity prediction model and the task configuration parameters, predict the multiple emission spectral curves of a specified molecule and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding first analysis data table;

[0009] If the task type is the second type, under the set conditions where multiple light intensities are kept consistent but the temperatures are different, based on the emission intensity prediction model and the task configuration parameters, predict the multiple emission spectral curves of a specified molecule and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding second analysis data table;

[0010] If the task type is the third type, under the set conditions where the temperature and light intensity are both kept consistent, based on the emission intensity prediction model and the task configuration parameters, respectively predict the emission spectral curves of two specified molecules and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of the two predicted curves to obtain the corresponding third analysis data table;

[0011] Feed back the obtained first, second, or third analysis data table to the current user.

[0012] Preferably, the molecular structure M includes multiple atoms mi , where 1 ≤ atomic index i ≤ N M , N M is the total number of atoms in the molecular structure M; each atom m i has atomic parameters composed of the corresponding atomic type and three-dimensional atomic coordinates;

[0013] The first data set includes a plurality of first data records; the first data records include a first training molecular structure, a first training temperature, a first training light intensity, a first training wavelength sequence, and a first emission intensity label sequence; the data structure type of the first training molecular structure is the same as the data structure type of the molecular structure M; the first training wavelength sequence is composed of a plurality of first training wavelengths; the first emission intensity label sequence is composed of a plurality of first emission intensity labels; the first emission intensity labels correspond one-to-one with the first training wavelengths;

[0014] When the task type is the first type, the first light intensity sequence of the task configuration parameters is composed of a plurality of first light intensities sorted in ascending order of intensity; when the task type is the second type, the first temperature sequence of the task configuration parameters is composed of a plurality of first temperatures sorted in ascending order of temperature; when the task type is the first, second, or third type, the corresponding first, second, or third wavelength sequence in the task configuration parameters is composed of a plurality of corresponding first, second, or third wavelengths sorted in ascending order of wavelength;

[0015] The data structure types of the first, second, third, and fourth molecular structures are all the same as the data structure type of the molecular structure M;

[0016] The first analysis data table is composed of a plurality of first analysis data records; the record fields of the first analysis data records are composed of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak site field, a full width at half maximum field, and a half-peak wavelength range field; the total number of records of the first analysis data records is the same as the total number of the first light intensities of the corresponding first light intensity sequence;

[0017] The second analysis data table is composed of a plurality of second analysis data records; the record fields of the second analysis data records are composed of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak site field, a full width at half maximum field, and a half-peak wavelength range field; the total number of records of the second analysis data records is the same as the total number of the first temperatures of the corresponding first temperature sequence;

[0018] The third analysis data table consists of two third analysis data records; the record fields of the third analysis data record consist of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a full width at half maximum field, and a half peak wavelength range field; the two third analysis data records correspond one-to-one to the third and fourth molecular structures.

[0019] Preferably, the first model input end of the emission intensity prediction model is used to receive the molecular structure M input to the model, the second model input end is used to receive the temperature x input to the model, the third model input end is used to receive the light intensity y input to the model, and the fourth model input end is used to receive the wavelength z input to the model; the model output end of the emission intensity prediction model is used to output the corresponding predicted emission intensity u;

[0020] The emission intensity prediction model includes an embedding encoding module, a molecular structure encoding module, a first Gaussian kernel module, a first vector mapping module, a second Gaussian kernel module, a second vector mapping module, a third Gaussian kernel module, a third vector mapping module, a feature fusion module, and an emission intensity prediction module; the molecular structure encoding module is implemented based on the Uni-Mol model or a type of graph neural network that satisfies SE(3) equivariance; the first, second, and third vector mapping modules and the emission intensity prediction module are all implemented based on the MLP model;

[0021] The input end of the embedding encoding module is connected to the first model input end, and the output end is connected to the input end of the molecular structure encoding module; the output end of the molecular structure encoding module is connected to the first input end of the feature fusion module; the input end of the first Gaussian kernel module is connected to the second model input end, and the output end is connected to the input end of the first vector mapping module; the output end of the first vector mapping module is connected to the second input end of the feature fusion module; the input end of the second Gaussian kernel module is connected to the third model input end, and the output end is connected to the input end of the second vector mapping module; the output end of the second vector mapping module is connected to the third input end of the feature fusion module; the input end of the third Gaussian kernel module is connected to the fourth model input end, and the output end is connected to the input end of the third vector mapping module; the output end of the third vector mapping module is connected to the fourth input end of the feature fusion module; the output end of the feature fusion module is connected to the input end of the emission intensity prediction module; the output end of the emission intensity prediction module is connected to the model output end;

[0022] The embedding encoding module is used to encode the molecular structure M input by the model according to the embedding encoding rules of the molecular structure encoding module and send the encoding result to the molecular structure encoding module. Specifically, when the molecular structure encoding module is implemented based on the Uni-Mol model, the embedding encoding module performs one-hot encoding of the atomic types of each atom m in the molecular structure M according to the atomic type one-hot encoding rules of the Uni-Mol model i to obtain the corresponding atomic encoding vector, and the N M atomic encoding vectors obtained form the corresponding atomic encoding tensor. Then, according to the pairwise feature initialization encoding rules of the Uni-Mol model, encoding is performed based on all three-dimensional atomic coordinates of the molecular structure M to obtain a pairwise encoding tensor with a tensor shape of N M ×N M . The obtained atomic encoding tensor and pairwise encoding tensor are sent to the molecular structure encoding module. When the molecular structure encoding module is implemented based on a class of graph neural networks that satisfy SE(3) equivariance, each atom m in the molecular structure M i serves as a corresponding first node, and the N M first nodes obtained form the corresponding first node set. A corresponding directed edge is made from each first node to any other node and denoted as the corresponding first edge, and the edge relationship of each first edge is set as the coordinate difference vector between the two corresponding atoms m i . The N M ×(N M -1) first edges obtained form the corresponding first edge set, and the obtained first node set and first edge set form a corresponding first graph structure and send it to the molecular structure encoding module;

[0023] When the molecular structure encoding module is based on the Uni-Mol model, it is used to perform atomic feature and pairwise feature extraction processing on the input atomic encoding tensor and pairwise encoding tensor to obtain the corresponding hidden feature tensor H and send it to the feature fusion module. The hidden feature tensor H consists of N M hidden feature vectors h i . The hidden feature vector h i corresponds one-to-one with the atom m i ;

[0024] When the molecular structure encoding module is implemented based on a class of graph neural networks that satisfy SE(3) equivariance, it is used to perform feature update processing on the node features and edge features of each first node and each first edge in the input first graph structure through the message passing mechanism of the graph neural network to obtain the corresponding hidden feature tensor H and send it to the feature fusion module;

[0025] The first Gaussian kernel module is used to preset a plurality of first Gaussian kernel functions within the module;

[0026] The first Gaussian kernel module is further used to perform corresponding kernel function calculations on the temperature x input to the model according to each of the first Gaussian kernel functions, and use the obtained numerical results as corresponding first result values; and form a corresponding temperature encoding vector from all the obtained first result values and send it to the first vector mapping module; the vector length of the temperature encoding vector is consistent with the total number of the first Gaussian kernel functions;

[0027] The first vector mapping module is used to map the hidden feature vector h i to a corresponding vector space as the target vector space; and perform a target vector space mapping on the temperature encoding vector to obtain a corresponding temperature mapping vector and send it to the feature fusion module; the temperature mapping vector is consistent with the hidden feature vector h i in terms of the feature dimension;

[0028] The second Gaussian kernel module is used to preset a plurality of second Gaussian kernel functions within the module;

[0029] The second Gaussian kernel module is further used to perform corresponding kernel function calculations on the light intensity y input to the model according to each of the second Gaussian kernel functions, and use the obtained numerical results as corresponding second result values; and form a corresponding light intensity encoding vector from all the obtained second result values and send it to the second vector mapping module; the vector length of the light intensity encoding vector is consistent with the total number of the second Gaussian kernel functions;

[0030] The second vector mapping module is used to map the hidden feature vector h i to a corresponding vector space as the target vector space; and perform a target vector space mapping on the light intensity encoding vector to obtain a corresponding light intensity mapping vector and send it to the feature fusion module; the light intensity mapping vector is consistent with the hidden feature vector h i in terms of the feature dimension;

[0031] The third Gaussian kernel module is used to preset a plurality of third Gaussian kernel functions within the module;

[0032] The third Gaussian kernel module is further used to perform corresponding kernel function calculations on the wavelength z input to the model according to each of the third Gaussian kernel functions, and use the obtained numerical results as corresponding third result values; and form a corresponding wavelength encoding vector from all the obtained third result values and send it to the third vector mapping module; the vector length of the wavelength encoding vector is consistent with the total number of the third Gaussian kernel functions;

[0033] The third vector mapping module is used to map the hidden feature vector h i to a corresponding vector space as the target vector space; and perform a target vector space mapping on the wavelength encoding vector to obtain a corresponding wavelength mapping vector and send it to the feature fusion module; the wavelength mapping vector and the hidden feature vector h i have the same feature dimension;

[0034] The feature fusion module is used to sequentially splice each of the hidden feature vectors h of the hidden feature tensor H i with the temperature mapping vector, the light intensity mapping vector and the wavelength mapping vector, and use the obtained spliced vector as a corresponding fused feature vector r i ; and form a corresponding fused feature tensor R from the obtained N M fused feature vectors r i and send it to the emission intensity prediction module;

[0035] The emission intensity prediction module is used to perform emission intensity regression calculation based on the input fused feature tensor R and output the calculation result as the corresponding predicted emission intensity u.

[0036] Preferably, a first data set for model training is constructed through data collection, specifically including:

[0037] Step 41, perform big data collection on the organic molecular structure of the organic photovoltaic material for the light-emitting device, its corresponding emission spectrum curve, and the temperature and light intensity information corresponding to the current curve through multiple data collection channels to obtain a plurality of first collection data;

[0038] Among them, the multiple data collection channels at least include a plurality of publicly available molecular databases in the field of organic photovoltaic materials, all publicly available literature and experimental databases in the field of organic photovoltaic materials;

[0039] The first collection data at least includes a first collection structure, a first collection curve data sequence, a first collection temperature, and a first collection light intensity; the first collection structure is the molecular structure of an organic molecule, and its data structure is the same as the data structure type of the molecular structure M; the first collection curve data sequence is the discretized sampling point data sequence corresponding to the emission spectrum curve of the current organic molecule, which is composed of a plurality of first sampling points, and each first sampling point is composed of a sampling point wavelength and its corresponding sampling point emission intensity; the first collection temperature and the first collection light intensity are respectively the ambient temperature and ambient light intensity corresponding to the current emission spectrum curve;

[0040] Step 42, perform a round of traversal on all the first acquisition data; and during this round of traversal, regard the currently traversed first acquisition data as the corresponding current acquisition data; and regard the first acquisition structure, the first acquisition temperature, and the first acquisition light intensity of the current acquisition data as the corresponding first training molecular structure, the first training temperature, and the first training light intensity; and regard the wavelengths of each sampling point of the current acquisition data as a corresponding first training wavelength, and the emission intensities of each sampling point as a corresponding first emission intensity label; and sort all the obtained first training wavelengths in the sorting order of the first sampling points to obtain the corresponding first training wavelength sequence; and sort all the obtained first emission intensity labels in the sorting order of the first sampling points to obtain the corresponding first emission intensity label sequence; and form a corresponding first data record from the first training molecular structure, the first training temperature, the first training light intensity, the first training wavelength sequence, and the first emission intensity label sequence corresponding to the current acquisition data; and at the end of this round of traversal, form the corresponding first data set from all the obtained first data records.

[0041] Preferably, the model training of the emission intensity prediction model based on the first data set specifically includes:

[0042] Step 51, divide the first data set into two sub-data sets denoted as the corresponding first training set and first evaluation set based on a preset first division ratio;

[0043] Among them, both the first training set and the first evaluation set are composed of multiple first data records; the ratio of the total number of records in the first training set and the first evaluation set satisfies the first division ratio;

[0044] Step 52, regard the first data record of the first training set as the corresponding current training record;

[0045] Step 53, regard the first training molecular structure, the first training temperature, and the first training light intensity of the current training record as the current molecular structure M, the temperature x, and the light intensity y;

[0046] Step 54: Statistically obtain the corresponding first total number N1 for the total number of the first training wavelengths in the current training record; perform a round of traversal on the N1 first training wavelengths in the current training record; during this round of traversal, use the currently traversed first training wavelength as the current wavelength z; input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output from this processing as a corresponding predicted intensity y pre,j , 1 ≤ index j ≤ N1; use the first emission intensity label corresponding to the currently traversed first training wavelength in the current training record as a corresponding label intensity y tag,j ; and from the predicted intensity y pre,j and the corresponding label intensity y tag,j form a corresponding first prediction-label pair (y pre,j , y tag,j );

[0047] Step 55: Substitute the obtained N1 first prediction-label pairs (y pre,j , y tag,j ) into the preset first model loss function L a ; and based on the preset first model optimizer, modulate the model parameters of the emission intensity prediction model in the direction of minimizing the first model loss function L a ;

[0048] Among them, the first model loss function L a is:

[0049]

[0050] The first model optimizer includes at least the Adam optimizer and the SGD optimizer;

[0051] Step 56: Identify whether the current training record is the last first data record in the first training set. If so, go to Step 57; if not, use the next first data record in the first training set as the new current training record and return to Step 53 to continue training;

[0052] Step 57: Perform a round of traversal on all the first data records in the first evaluation set; during this round of traversal, regard the currently traversed first data record as the corresponding current evaluation record; regard the first training molecular structure, the first training temperature, and the first training light intensity of the current evaluation record as the current molecular structure M, temperature x, and light intensity y respectively; regard each of the first training wavelengths of the current evaluation record as the corresponding current wavelength z in turn, and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and regard the predicted emission intensity u output by the current processing as a corresponding first predicted intensity; regard the first emission intensity label corresponding to each of the first predicted intensities in the current evaluation record as the corresponding first label intensity; and form a corresponding second prediction-label pair from each of the first predicted intensities and their corresponding first label intensities in the current evaluation record.

[0053] Step 58: After the end of this round of traversal of all the first data records in the first evaluation set, count the total number of the obtained second prediction-label pairs to obtain the corresponding second total number N2; and regard the first predicted intensity and the first label intensity of each of the second prediction-label pairs as a corresponding group of predicted intensity y pre,k and label intensity y tag,k , where 1 ≤ index k ≤ N2; and regard the obtained N2 groups of predicted intensity y pre,k and the label intensity y tag,k and substitute them into the preset first model evaluation function S a for calculation to obtain the corresponding first evaluation value.

[0054] Among them, the first model evaluation function S a is:

[0055]

[0056] Step 59: Identify whether the first evaluation value meets the preset first evaluation value range; if not, return to Step 52 to continue training; if it meets, stop training and confirm that the model training is completed.

[0057] Preferably, predicting multiple emission spectral curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters, and performing data analysis on the maximum peak site, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding first analysis data table specifically includes:

[0058] Extract the corresponding first molecular structure, first set temperature, first light intensity sequence, and first wavelength sequence from the current task configuration parameters; and use the first molecular structure and the first set temperature as the current molecular structure M and temperature x; and count the total number of the first light intensities in the first light intensity sequence to obtain the corresponding total number N R1 ;

[0059] And sequentially use each of the first light intensities in the first light intensity sequence as the corresponding current light intensity y; and perform a round of traversal on all the first wavelengths in the first wavelength sequence; and during this round of traversal, use the currently traversed first wavelength as the corresponding current wavelength z; and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelength and curve point emission intensity to form a corresponding curve point; and at the end of this round of traversal, sort all the curve points corresponding to the current light intensity y in ascending order of the curve point wavelength to form a corresponding curve data sequence;

[0060] And perform a round of traversal on the N R1 first light intensities in the first light intensity sequence; and during this round of traversal, use the currently traversed first light intensity as the corresponding current light intensity; and use the first molecular structure, the first set temperature, and the curve data sequence corresponding to the current light intensity as the corresponding current molecular structure, current set temperature, and current curve data sequence; and perform maximum peak site, full width at half maximum, and half-peak wavelength range analysis based on the current curve data sequence to obtain the corresponding current maximum peak site, current full width at half maximum, and current half-peak wavelength range; and use the current molecular structure, the current set temperature, the current light intensity, the current curve data sequence, the current maximum peak site, the current full width at half maximum, and the current half-peak wavelength range as the corresponding molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak site field, full width at half maximum field, and half-peak wavelength range field to form a corresponding first analysis data record;

[0061] And after the end of this round of traversal of the N R1 first light intensities in the first light intensity sequence, use the obtained N R1 first analysis data records to form the corresponding first analysis data table.

[0062] Preferably, based on the emission intensity prediction model and the task configuration parameters, multiple emission spectral curves of a specified molecule are predicted, and data analysis is performed on the maximum peak positions, full widths at half maximum, and half-peak wavelength ranges of each predicted curve to obtain a corresponding second analysis data table, specifically including:

[0063] Extract the corresponding second molecular structure, the first set illumination intensity, the first temperature sequence, and the second wavelength sequence from the current task configuration parameters; and use the first molecular structure and the first set illumination intensity as the current molecular structure M and the illumination intensity y; and count the total number of the first temperatures in the first temperature sequence to obtain a corresponding total number N R2 ;

[0064] And sequentially use each of the first temperatures in the first temperature sequence as the corresponding current temperature x; and perform a round of traversal on all the second wavelengths in the second wavelength sequence; and during this round of traversal, use the currently traversed second wavelength as the corresponding current wavelength z; and input the current molecular structure M, the temperature x, the illumination intensity y, and the wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a set of corresponding curve point wavelengths and curve point emission intensities to form a corresponding curve point; and at the end of this round of traversal, sort all the curve points corresponding to the current temperature x in ascending order of the curve point wavelengths to form a corresponding curve data sequence;

[0065] And perform a round of traversal on the N R2 first temperatures in the first temperature sequence; and during this round of traversal, use the currently traversed first temperature as the corresponding current temperature; and use the first molecular structure, the first set illumination intensity, and the curve data sequence corresponding to the current temperature as the corresponding current molecular structure, current set illumination intensity, and current curve data sequence; and perform analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range based on the current curve data sequence to obtain the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range; and use the current molecular structure, the current temperature, the current set illumination intensity, the current curve data sequence, the current maximum peak position, the current full width at half maximum, and the current half-peak wavelength range as the corresponding molecular structure field, temperature field, illumination intensity field, curve data sequence field, maximum peak position field, full width at half maximum field, and half-peak wavelength range field to form a corresponding second analysis data record;

[0066] and at the end of the current round of traversing the N R2 first temperatures in the first temperature sequence, the N R2 second analysis data records obtained form a corresponding second analysis data table.

[0067] Preferably, based on the emission intensity prediction model and the task configuration parameters, the emission spectral curves of two specified molecules are predicted respectively, and data analysis is performed on the maximum peak positions, full widths at half maximum, and half-peak wavelength ranges of the two predicted curves to obtain a corresponding third analysis data table, which specifically includes:

[0068] Extract the corresponding third molecular structure, fourth molecular structure, second set temperature, second set light intensity, and third wavelength sequence from the current task configuration parameters; and use the second set temperature and the second set light intensity as the current temperature x and light intensity y;

[0069] And use the third and fourth molecular structures as the corresponding current molecular structure M in sequence; and perform a round of traversal on all the third wavelengths in the third wavelength sequence; and during this round of traversal, use the currently traversed third wavelength as the corresponding current wavelength z; and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelength and curve point emission intensity to form a corresponding curve point; and at the end of this round of traversal, sort all the curve points corresponding to the current molecular structure M in ascending order of the curve point wavelength to form a corresponding curve data sequence;

[0070] And use the third and fourth molecular structures as the corresponding current molecular structure in sequence; and use the second set temperature, the second set light intensity, and the curve data sequence corresponding to the current molecular structure as the corresponding current set temperature, current set light intensity, and current curve data sequence; and perform analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range based on the current curve data sequence to obtain the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range; and use the current molecular structure, the current set temperature, the current set light intensity, the current curve data sequence, the current maximum peak position, the current full width at half maximum, and the current half-peak wavelength range as the corresponding molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak position field, full width at half maximum field, and half-peak wavelength range field to form a corresponding third analysis data record;

[0071] And the two obtained third analysis data records form the corresponding third analysis data table.

[0072] Further, analyzing the maximum peak position, full width at half maximum, and half-peak wavelength range based on the current curve data sequence to obtain the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range specifically includes:

[0073] Construct a two-dimensional coordinate plane with the emission intensity I as the ordinate and the wavelength λ as the abscissa as the corresponding emission spectrum curve plane; and perform two-dimensional curve fitting on the emission spectrum curve plane based on the current curve data sequence and use the obtained fitting curve as the corresponding current emission spectrum curve;

[0074] And take the maximum emission intensity position on the current emission spectrum curve as the corresponding maximum peak position P max ; And denote the emission intensity and wavelength corresponding to the maximum peak position P max as the corresponding emission intensity I max and wavelength λ max ;

[0075] And denote the curve parts on the left and right sides of the maximum peak position P max on the current emission spectrum curve as the corresponding left curve and right curve respectively; and denote the curve positions with the emission intensity of I max / 2 on the left and right curves as the corresponding left and right candidate positions; and take the left candidate position farthest from the maximum peak position P max as the corresponding left half-peak position p1, and the right candidate position farthest from the maximum peak position P max as the corresponding right half-peak position p2; and denote the corresponding wavelengths of the left and right half-peak positions p1, p2 as the corresponding wavelengths λ1, λ2; and identify the wavelength range width between the wavelengths λ1, λ2 to obtain the corresponding full width at half maximum W = λ2 - λ1; and take the wavelength range with the wavelengths λ1, λ2 as the left and right boundaries as the corresponding half-peak wavelength range [λ1, λ2];

[0076] And take the obtained maximum peak position P max , the full width at half maximum W, and the half-peak wavelength range [λ1, λ2] as the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range.

[0077] In the second aspect of the embodiments of the present invention, there is provided an apparatus for implementing the prediction analysis method of the emission spectral curve described in the first aspect above. The apparatus includes: a model construction module, a model training module, a prediction analysis task data receiving module, a prediction analysis task processing module, and a prediction analysis task feedback module;

[0078] The model construction module is used to construct an emission intensity prediction model for performing emission intensity prediction. The emission intensity prediction model is used to perform emission intensity prediction processing based on the molecular structure M, temperature x, light intensity y, and wavelength z input into the model and output the corresponding predicted emission intensity u;

[0079] The model training module is used to construct a first data set for model training through data collection; and perform model training on the emission intensity prediction model according to the first data set;

[0080] The prediction analysis task data receiving module is used to, after the model training is completed, receive the prediction analysis task input by the user and extract the corresponding task type and task configuration parameters from it. The prediction analysis task includes the task type and the task configuration parameters. The task type includes the first, second, and third types. When the task type is the first type, the corresponding task configuration parameters include the first molecular structure, the first set temperature, the first light intensity sequence, and the first wavelength sequence. When the task type is the second type, the corresponding task configuration parameters include the second molecular structure, the first set light intensity, the first temperature sequence, and the second wavelength sequence. When the task type is the third type, the corresponding task configuration parameters include the third molecular structure, the fourth molecular structure, the second set temperature, the second set light intensity, and the third wavelength sequence;

[0081] When the task type is the first type, the prediction analysis task processing module is used to, under the set conditions where multiple temperatures are the same but the light intensities are different, based on the emission intensity prediction model and the task configuration parameters, predict multiple emission spectral curves of a specified molecule and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding first analysis data table;

[0082] When the task type is the second type, the prediction analysis task processing module is also used to, under the set conditions where multiple light intensities are the same but the temperatures are different, based on the emission intensity prediction model and the task configuration parameters, predict multiple emission spectral curves of a specified molecule and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding second analysis data table;

[0083] When the task type is the third type, the prediction analysis task processing module is further configured to predict the emission spectral curves of two specified molecules respectively based on the emission intensity prediction model and the task configuration parameters under the set conditions where the temperature and the light intensity are both consistent, and perform data analysis on the maximum peak positions, full widths at half maximum, and half-peak wavelength ranges of the two predicted curves to obtain a corresponding third analysis data table;

[0084] The prediction analysis task feedback module is configured to feedback the first, second, or third analysis data table obtained this time to the current user.

[0085] A third aspect of the embodiments of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0086] The processor is configured to be coupled with the memory, read and execute instructions in the memory to implement the method steps described in the first aspect above;

[0087] The transceiver is coupled with the processor, and the processor controls the transceiver to perform message sending and receiving.

[0088] A fourth aspect of the embodiments of the present invention provides a computer-readable storage medium storing computer instructions, and when the computer instructions are executed by a computer, the computer is caused to execute the instructions of the method described in the first aspect above.

[0089] Embodiments of the present invention provide a method, apparatus, electronic device, and computer-readable storage medium for predicting and analyzing emission spectral curves. As can be seen from the above, embodiments of the present invention customize an emission intensity prediction model for predicting emission intensity based on the molecular structure, temperature, light intensity, and wavelength input into the model, and train it with a first data set obtained by big data collection; after the training is completed, the prediction analysis task (composed of a task type and task configuration parameters) input by the user is parsed; if the task type is the first type, multiple emission spectral curves of a specified molecule are predicted based on the emission intensity prediction model and task configuration parameters under a set of conditions where multiple temperatures are kept consistent but the light intensities are different, and three types of curve characteristics (maximum peak position, full width at half maximum, and half-peak wavelength range) of each predicted curve are analyzed to obtain a first analysis data table and feedback it to the current user; if the task type is the second type, multiple emission spectral curves of a specified molecule are predicted based on the emission intensity prediction model and task configuration parameters under a set of conditions where multiple light intensities are kept consistent but the temperatures are different, and three types of curve characteristics of each predicted curve are analyzed to obtain a second analysis data table and feedback it to the current user; if the task type is the third type, under a set of conditions where both the temperature and the light intensity are kept consistent, the emission spectral curves of two specified molecules are respectively predicted based on the emission intensity prediction model and task configuration parameters, and three types of curve characteristics of the two predicted curves are analyzed to obtain a third analysis data table and feedback it to the current user. On the one hand, embodiments of the present invention reduce the processing complexity, shorten the processing time, and improve the processing efficiency of the prediction analysis task through a customized emission intensity prediction model; on the other hand, the convenience of data analysis is improved through three types of customized emission spectral curve analysis means. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 FIG. 1 is a schematic diagram of a method for predicting and analyzing emission spectral curves provided in Embodiment 1 of the present invention;

[0091] Figure 2 FIG. 2 is a module structure diagram of an emission intensity prediction model provided in Embodiment 1 of the present invention;

[0092] Figure 3 FIG. 3 is a schematic diagram of the data structures of the first analysis data table, the second analysis data table, and the third analysis data table provided in Embodiment 1 of the present invention;

[0093] Figure 4 FIG. 4 is a schematic diagram of an emission spectral curve, a maximum peak position, a left half-peak position, a right half-peak position, a full width at half maximum, and a half-peak wavelength range provided in Embodiment 1 of the present invention;

[0094] Figure 5 FIG. 5 is a module structure diagram of an apparatus for predicting and analyzing emission spectral curves provided in Embodiment 2 of the present invention;

[0095] Figure 6 This is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present invention. Detailed implementation manners

[0096] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0097] Embodiment 1 of the present invention provides a method for predicting and analyzing an emission spectrum curve, as Figure 1 shown in the schematic diagram of the method for predicting and analyzing an emission spectrum curve provided in Embodiment 1 of the present invention. The method mainly includes the following steps:

[0098] Step 1: Construct an emission intensity prediction model for performing emission intensity prediction.

[0099] Here, the emission intensity prediction model of the embodiment of the present invention is used to perform emission intensity prediction processing based on the molecular structure M, temperature x, light intensity y, and wavelength z input to the model and output the corresponding predicted emission intensity u. Among them, the molecular structure M includes multiple atoms m i , 1 ≤ atomic index i ≤ N M , N M is the total number of atoms in the molecular structure M; the atomic parameters of each atom m i are composed of the corresponding atomic type and three-dimensional atomic coordinates.

[0100] As Figure 2 shown in the module structure diagram of the emission intensity prediction model provided in Embodiment 1 of the present invention, the first model input end of the emission intensity prediction model is used to receive the molecular structure M input to the model, the second model input end is used to receive the temperature x input to the model, the third model input end is used to receive the light intensity y input to the model, and the fourth model input end is used to receive the wavelength z input to the model; the model output end of the emission intensity prediction model is used to output the corresponding predicted emission intensity u.

[0101] As Figure 2As shown, the model components of the emission intensity prediction model include: an embedding encoding module, a molecular structure encoding module, a first Gaussian kernel module, a first vector mapping module, a second Gaussian kernel module, a second vector mapping module, a third Gaussian kernel module, a third vector mapping module, a feature fusion module, and an emission intensity prediction module. It should be noted that the molecular structure encoding module in the embodiments of the present invention is implemented based on the Uni-Mol model or a type of graph neural network that satisfies SE(3) equivariance; the first, second, and third vector mapping modules and the emission intensity prediction module in the embodiments of the present invention are each implemented based on the MLP model. The above-mentioned Uni-Mol model is an encoder model for atomic-level feature encoding of molecular structures. The detailed model structure, model inference principle, one-hot encoding rule for atomic types, pairwise feature initialization encoding rule, and model pre-training scheme of the Uni-Mol model are clearly described in the public technical literature A, "Uni-Mol: A Universal 3D Molecular Representation Learning Framework". The above-mentioned type of graph neural network that satisfies SE(3) equivariance can specifically be the EGNN model or the SE(3)-Transformers model, and the EGNN model is preferably used in the embodiments of the present invention. It should also be noted that regardless of whether the molecular structure encoding module in the embodiments of the present invention is implemented based on the Uni-Mol model or a type of graph neural network that satisfies SE(3) equivariance, the currently used Uni-Mol model or graph neural network has been pre-trained based on their respective corresponding conventional pre-training schemes, and the pre-training schemes of the Uni-Mol model, EGNN model, and SE(3)-Transformers model have also been published through public literature and will not be further elaborated here.

[0102] As Figure 2As shown in the figure, the connection relationships of the components of the emission intensity prediction model are as follows: the input end of the embedding encoding module is connected to the first model input end, and the output end is connected to the input end of the molecular structure encoding module; the output end of the molecular structure encoding module is connected to the first input end of the feature fusion module; the input end of the first Gaussian kernel module is connected to the second model input end, and the output end is connected to the input end of the first vector mapping module; the output end of the first vector mapping module is connected to the second input end of the feature fusion module; the input end of the second Gaussian kernel module is connected to the third model input end, and the output end is connected to the input end of the second vector mapping module; the output end of the second vector mapping module is connected to the third input end of the feature fusion module; the input end of the third Gaussian kernel module is connected to the fourth model input end, and the output end is connected to the input end of the third vector mapping module; the output end of the third vector mapping module is connected to the fourth input end of the feature fusion module; the output end of the feature fusion module is connected to the input end of the emission intensity prediction module; the output end of the emission intensity prediction module is connected to the model output end.

[0103] The functions of the components of the emission intensity prediction model are as follows.

[0104] 1) Embedding encoding module:

[0105] The embedding encoding module in the embodiment of the present invention is used to encode the molecular structure M input by the model according to the embedding encoding rule of the molecular structure encoding module and send the encoding result to the molecular structure encoding module. Specifically:

[0106] a. If the molecular structure encoding module is implemented based on the Uni-Mol model, then the embedding encoding module is used to perform one-hot encoding of the atomic types of each atom m of the molecular structure M according to the atomic type one-hot encoding rule of the Uni-Mol model i to obtain the corresponding atomic encoding vector, and the N M obtained atomic encoding vectors are used to form the corresponding atomic encoding tensor; and according to the pairwise feature initialization encoding rule of the Uni-Mol model, encoding is performed based on all three-dimensional atomic coordinates of the molecular structure M to obtain a tensor with a shape of N M ×N M pairwise encoding tensor; and the obtained atomic encoding tensor and pairwise encoding tensor are sent to the molecular structure encoding module;

[0107] b. If the molecular structure encoding module is implemented based on a class of graph neural networks that satisfy SE(3) equivariance, then the embedding encoding module uses each atom m of the molecular structure M i as a corresponding first node, and the N M obtained first nodes are used to form the corresponding first node set; and a corresponding directed edge is made from each first node to any other node and is denoted as the corresponding first edge; and the edge relationship of each first edge is set to the edge relationship of two corresponding atoms mi The coordinate difference vector between them; and from the obtained N M ×(N M -1) first edges form a corresponding first edge set; and a corresponding first graph structure is formed by the obtained first node set and the first edge set and sent to the molecular structure encoding module.

[0108] 2) Molecular structure encoding module:

[0109] If the molecular structure encoding module of the embodiment of the present invention is implemented based on the Uni-Mol model, then the molecular structure encoding module is used to perform atomic feature and pairwise feature extraction processing according to the encoding mechanism of the Uni-Mol model based on the input atomic encoding tensor and pairwise encoding tensor to obtain a corresponding hidden feature tensor H and send it to the feature fusion module; the hidden feature tensor H is composed of N M hidden feature vectors h i which are i in one-to-one correspondence with the atoms m i one by one.

[0110] If the molecular structure encoding module of the embodiment of the present invention is implemented based on a class of graph neural networks that satisfy SE(3) equivariance, then the molecular structure encoding module is used to perform feature update processing on the node features and edge features of each first node and each first edge in the input first graph structure through the message passing mechanism of the graph neural network to obtain a corresponding hidden feature tensor H and send it to the feature fusion module; the hidden feature tensor H is also composed of N M hidden feature vectors h i which are i in one-to-one correspondence with the atoms m i one by one.

[0111] 3) First Gaussian kernel module and first vector mapping module:

[0112] The first Gaussian kernel module of the embodiment of the present invention is used to preset a plurality of first Gaussian kernel functions in the module. The first Gaussian kernel module is further used to perform corresponding kernel function calculations on the temperature x input to the model according to each first Gaussian kernel function and use the calculated numerical result as the corresponding first result value; and all the obtained first result values form a corresponding temperature encoding vector and send it to the first vector mapping module; here, the vector length of the temperature encoding vector is consistent with the total number of the first Gaussian kernel functions.

[0113] The first vector mapping module of the embodiment of the present invention is used to use the vector space corresponding to the hidden feature vector h i as the target vector space; and perform a target vector space mapping on the temperature encoding vector to obtain a corresponding temperature mapping vector and send it to the feature fusion module. Here, the temperature mapping vector and the hidden feature vector hi The characteristic dimensions are kept consistent.

[0114] 4) The second Gaussian kernel module and the second vector mapping module:

[0115] The second Gaussian kernel module in the embodiment of the present invention is used to preset a plurality of second Gaussian kernel functions in the module. The second Gaussian kernel module is further used to perform corresponding kernel function calculations on the illumination intensity y input to the model according to each second Gaussian kernel function and use the calculated numerical result as the corresponding second result value; and all the obtained second result values are combined to form a corresponding illumination intensity encoding vector and sent to the second vector mapping module. Here, the vector length of the illumination intensity encoding vector is kept consistent with the total number of the second Gaussian kernel functions.

[0116] The second vector mapping module in the embodiment of the present invention is used to use the i corresponding vector space as the target vector space; and perform a target vector space mapping on the illumination intensity encoding vector to obtain a corresponding illumination intensity mapping vector and send it to the feature fusion module. Here, the illumination intensity mapping vector and the hidden feature vector h i have consistent characteristic dimensions.

[0117] 5) The third Gaussian kernel module and the third vector mapping module:

[0118] The third Gaussian kernel module in the embodiment of the present invention is used to preset a plurality of third Gaussian kernel functions in the module. The third Gaussian kernel module is further used to perform corresponding kernel function calculations on the wavelength z input to the model according to each third Gaussian kernel function and use the calculated numerical result as the corresponding third result value; and all the obtained third result values are combined to form a corresponding wavelength encoding vector and sent to the third vector mapping module. Here, the vector length of the wavelength encoding vector is kept consistent with the total number of the third Gaussian kernel functions.

[0119] The third vector mapping module in the embodiment of the present invention is used to use the hidden feature vector h i corresponding vector space as the target vector space; and perform a target vector space mapping on the wavelength encoding vector to obtain a corresponding wavelength mapping vector and send it to the feature fusion module. Here, the wavelength mapping vector and the hidden feature vector h i have consistent characteristic dimensions.

[0120] 6) The feature fusion module:

[0121] The feature fusion module in the embodiment of the present invention is used to sequentially splice each hidden feature vector h of the hidden feature tensor H i with the temperature mapping vector, the illumination intensity mapping vector and the wavelength mapping vector and use the obtained spliced vector as a corresponding fusion feature vector r i ; and by the obtained NM a fused feature vector r i Compose the corresponding fused feature tensor R and send it to the emission intensity prediction module.

[0122] 7) Emission intensity prediction module:

[0123] The emission intensity prediction module of the embodiment of the present invention is used to perform emission intensity regression calculation according to the input fused feature tensor R and output the calculation result as the corresponding predicted emission intensity u.

[0124] Step 2, construct a first data set for model training through data collection; and perform model training on the emission intensity prediction model according to the first data set;

[0125] Specifically, it includes: Step 21, construct a first data set for model training through data collection;

[0126] Among them, the first data set includes multiple first data records; the first data record includes the first training molecular structure, the first training temperature, the first training light intensity, the first training wavelength sequence, and the first emission intensity label sequence; the data structure type of the first training molecular structure is the same as that of the molecular structure M; the first training wavelength sequence is composed of multiple first training wavelengths; the first emission intensity label sequence is composed of multiple first emission intensity labels; the first emission intensity label corresponds to the first training wavelength one by one;

[0127] Specifically, it includes: Step 211, perform big data collection on the organic molecular structure of the organic photovoltaic material for the lighting device, its corresponding emission spectrum curve, and the temperature and light intensity information corresponding to the current curve through multiple data collection channels to obtain multiple first collection data;

[0128] Here, the multiple data collection channels of the embodiment of the present invention at least include multiple public molecular databases in the field of organic photovoltaic materials, all public literatures and experimental databases in the field of organic photovoltaic materials;

[0129] The first collection data of the embodiment of the present invention at least includes the first collection structure, the first collection curve data sequence, the first collection temperature, and the first collection light intensity; among them, the first collection structure is the molecular structure of an organic molecule, and its data structure is the same as that of the molecular structure M; the first collection curve data sequence is the discretized sampling point data sequence corresponding to the emission spectrum curve of the current organic molecule, which is composed of multiple first sampling points, and each first sampling point is composed of a sampling point wavelength and its corresponding sampling point emission intensity; the first collection temperature and the first collection light intensity are respectively the ambient temperature and ambient light intensity corresponding to the current emission spectrum curve;

[0130] Step 212, perform a round of traversal on all the first acquisition data; and during this round of traversal, take the currently traversed first acquisition data as the corresponding current acquisition data; and take the first acquisition structure, the first acquisition temperature, and the first acquisition light intensity of the current acquisition data as the corresponding first training molecular structure, the first training temperature, and the first training light intensity; and take the wavelengths of each sampling point of the current acquisition data as a corresponding first training wavelength, and the emission intensities of each sampling point as a corresponding first emission intensity label; and sort all the obtained first training wavelengths in the sorting order of the first sampling point to obtain the corresponding first training wavelength sequence; and sort all the obtained first emission intensity labels in the sorting order of the first sampling point to obtain the corresponding first emission intensity label sequence; and form a corresponding first data record from the first training molecular structure, the first training temperature, the first training light intensity, the first training wavelength sequence, and the first emission intensity label sequence corresponding to the current acquisition data; and at the end of this round of traversal, form the corresponding first data set from all the obtained first data records;

[0131] Step 22, and perform model training on the emission intensity prediction model according to the first data set;

[0132] Specifically, it includes: Step 221, divide the first data set into two sub-data sets based on a preset first division ratio, denoted as the corresponding first training set and first evaluation set;

[0133] Here, the first division ratio of the embodiment of the present invention is a preset ratio parameter, such as 8:2; both the first training set and the first evaluation set are composed of multiple first data records; the ratio of the total number of records in the first training set and the first evaluation set satisfies the first division ratio;

[0134] Step 222, take the first first data record of the first training set as the corresponding current training record;

[0135] Step 223, take the first training molecular structure, the first training temperature, and the first training light intensity of the current training record as the current molecular structure M, temperature x, and light intensity y;

[0136] Step 224, count the total number of the first training wavelengths of the current training record to obtain the corresponding first total number N1; and perform a round of traversal on the N1 first training wavelengths of the current training record; and during this round of traversal, take the currently traversed first training wavelength as the current wavelength z; and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and take the predicted emission intensity u output by this processing as a corresponding predicted intensity y pre,j, where 1 ≤ index j ≤ N1; and use the first emission intensity label corresponding to the currently traversed first training wavelength in the current training record as a corresponding label intensity y tag,j ; and from the predicted intensity y pre,j and the corresponding label intensity y tag,j to form a corresponding first prediction-label pair (y pre,j , y tag,j );

[0137] Step 225, bring the obtained N1 first prediction-label pairs (y pre,j , y tag,j ) into the preset first model loss function L a ; and based on the preset first model optimizer, modulate the model parameters of the emission intensity prediction model in the direction of minimizing the first model loss function L a for one round;

[0138] Here, the first model loss function L a of the embodiment of the present invention is:

[0139]

[0140] The first model optimizer of the embodiment of the present invention at least includes an Adam optimizer and an SGD optimizer;

[0141] Step 226, identify whether the current training record is the last first data record of the first training set. If so, go to Step 227. If not, use the next first data record of the first training set as the new current training record and return to Step 223 to continue training;

[0142] Step 227, perform one round of traversal on all first data records of the first evaluation set; and during this round of traversal, use the currently traversed first data record as the corresponding current evaluation record; and use the first training molecular structure, the first training temperature, and the first training light intensity of the current evaluation record as the current molecular structure M, temperature x, and light intensity y; and use each first training wavelength of the current evaluation record as the corresponding current wavelength z in turn, and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding first predicted intensity; and record the first emission intensity label corresponding to each first predicted intensity in the current evaluation record as the corresponding first label intensity; and form a corresponding second prediction-label pair from each first predicted intensity of the current evaluation record and its corresponding first label intensity;

[0143] Step 228, after the current round of traversal of all the first data records in the first evaluation set, count the total number of the obtained second prediction-label pairs to obtain the corresponding second total number N2; and record the first prediction strength and the first label strength of each second prediction-label pair as a group of corresponding prediction strengths y pre,k and label strengths y tag,k , 1 ≤ index k ≤ N2; and record the obtained N2 groups of prediction strengths y pre,k and label strengths y tag,k into the preset first model evaluation function S a for calculation to obtain the corresponding first evaluation value;

[0144] Here, the first model evaluation function S of the embodiment of the present invention a is:

[0145]

[0146] Step 229, identify whether the first evaluation value meets the preset first evaluation value range; if not, return to Step 222 to continue training; if so, stop training and confirm that the model training is completed.

[0147] Here, the first evaluation value range of the embodiment of the present invention is a preset numerical range.

[0148] Step 3, after the model training is completed, receive the prediction analysis task input by the user and extract the corresponding task type and task configuration parameters therefrom.

[0149] Here, the prediction analysis task of the embodiment of the present invention includes a task type and task configuration parameters; among them, the task type includes the first, second, and third types; the specific parameter content of the task configuration parameters has a corresponding relationship with the specific type of the task type:

[0150] 1) When the task type is the first type, the corresponding task configuration parameters include a first molecular structure, a first set temperature, a first light intensity sequence, and a first wavelength sequence; wherein the data structure type of the first molecular structure is the same as the data structure type of the molecular structure M; the first light intensity sequence is composed of multiple first light intensities sorted in ascending order of intensity; the first wavelength sequence is composed of multiple first wavelengths sorted in ascending order of wavelength;

[0151] When the task type is the second type, the corresponding task configuration parameters include a second molecular structure, a first set light intensity, a first temperature sequence, and a second wavelength sequence; wherein the data structure type of the second molecular structure is the same as the data structure type of the molecular structure M; the first temperature sequence is composed of multiple first temperatures sorted in ascending order of temperature; the second wavelength sequence is composed of multiple second wavelengths sorted in ascending order of wavelength;

[0152] When the task type is the third type, the corresponding task configuration parameters include a third molecular structure, a fourth molecular structure, a second set temperature, a second set light intensity, and a third wavelength sequence; the data structure types of the third and fourth molecular structures are both the same as the data structure type of the molecular structure M; the third wavelength sequence is formed by sorting multiple third wavelengths in ascending order of wavelength.

[0153] Step 4, if the task type is the first type, under a set of conditions where multiple temperatures are the same but the light intensities are different, based on the emission intensity prediction model and the task configuration parameters, predict multiple emission spectral curves of a specified molecule and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain a corresponding first analysis data table;

[0154] Among them, the first analysis data table consists of multiple first analysis data records; the record fields of the first analysis data record consist of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a full width at half maximum field, and a half-peak wavelength range field; the total number of the first analysis data records is the same as the total number of the first light intensities in the corresponding first light intensity sequence;

[0155] Specifically, it includes: Step 41, extract the corresponding first molecular structure, first set temperature, first light intensity sequence, and first wavelength sequence from the current task configuration parameters; and use the first molecular structure and the first set temperature as the current molecular structure M and temperature x; and count the total number of the first light intensities in the first light intensity sequence to obtain a corresponding total number N R1 ;

[0156] Step 42, and sequentially use each first light intensity in the first light intensity sequence as the corresponding current light intensity y; and perform a round of traversal on all the first wavelengths in the first wavelength sequence; and during this round of traversal, use the currently traversed first wavelength as the corresponding current wavelength z; and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelength and curve point emission intensity to form a corresponding curve point; and at the end of this round of traversal, sort all the curve points corresponding to the current light intensity y in ascending order of curve point wavelength to form a corresponding curve data sequence;

[0157] Step 43, and for the N of the first light intensity sequence R1Perform a round of traversal for a first light intensity; during this round of traversal, use the currently traversed first light intensity as the corresponding current light intensity; use the first molecular structure, the first set temperature, and the curve data sequence corresponding to the current light intensity as the corresponding current molecular structure, current set temperature, and current curve data sequence; perform analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range based on the current curve data sequence to obtain the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range; use the current molecular structure, current set temperature, current light intensity, current curve data sequence, current maximum peak position, current full width at half maximum, and current half-peak wavelength range to form a corresponding first analysis data record as the molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak position field, full width at half maximum field, and half-peak wavelength range field.

[0158] Step 44, and during the traversal of the N R1 first light intensities of the first light intensity sequence, after the end of this round of traversal of the N R1 first analysis data records obtained, form the corresponding first analysis data table.

[0159] Figure 3 FIG. 11 is a schematic diagram of the data structures of the first analysis data table, the second analysis data table, and the third analysis data table provided in Embodiment 1 of the present invention, where the data structure of the first analysis data table obtained in the embodiments of the present invention can be intuitively understood through Figure 3 this.

[0160] Step 5, if the task type is the second type, then under the set conditions where multiple light intensities are the same but the temperatures are different, based on the emission intensity prediction model and the task configuration parameters, predict multiple emission spectral curves of a specified molecule and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding second analysis data table;

[0161] Among them, the second analysis data table is composed of multiple second analysis data records; the record fields of the second analysis data record are composed of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a full width at half maximum field, and a half-peak wavelength range field; the total number of the second analysis data records is the same as the total number of the first temperatures of the corresponding first temperature sequence;

[0162] Specifically, it includes: Step 51, extract the corresponding second molecular structure, the first set light intensity, the first temperature sequence, and the second wavelength sequence from the current task configuration parameters; use the first molecular structure and the first set light intensity as the current molecular structure M and light intensity y; and count the total number of the first temperatures of the first temperature sequence to obtain the corresponding total number N R2 ;

[0163] Step 52, and sequentially use each first temperature in the first temperature sequence as the corresponding current temperature x; and perform a round of traversal on all second wavelengths in the second wavelength sequence; and during this round of traversal, use the currently traversed second wavelength as the corresponding current wavelength z; and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelengths and curve point emission intensities to form a corresponding curve point; and at the end of this round of traversal, sort all the curve points corresponding to the current temperature x in ascending order of the curve point wavelength to form a corresponding curve data sequence;

[0164] Step 53, and perform a round of traversal on the N R2 first temperatures in the first temperature sequence; and during this round of traversal, use the currently traversed first temperature as the corresponding current temperature; and use the first molecular structure, the first set light intensity, and the curve data sequence corresponding to the current temperature as the corresponding current molecular structure, current set light intensity, and current curve data sequence; and perform analysis on the maximum peak point, full width at half maximum, and half-peak wavelength range based on the current curve data sequence to obtain the corresponding current maximum peak point, current full width at half maximum, and current half-peak wavelength range; and use the current molecular structure, current temperature, current set light intensity, current curve data sequence, current maximum peak point, current full width at half maximum, and current half-peak wavelength range as the corresponding molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak point field, full width at half maximum field, and half-peak wavelength range field to form a corresponding second analysis data record;

[0165] Step 54, and at the end of this round of traversal of the N R2 first temperatures in the first temperature sequence, form a corresponding second analysis data table from the obtained N R2 second analysis data records.

[0166] Here, an intuitive understanding can be obtained through Figure 3 the data structure of the second analysis data table obtained in the embodiments of the present invention.

[0167] Step 6, if the task type is the third type, then under the set conditions where the temperature and light intensity are both kept consistent, respectively predict the emission spectral curves of two specified molecules based on the emission intensity prediction model and the task configuration parameters, and perform data analysis on the maximum peak point, full width at half maximum, and half-peak wavelength range of the two predicted curves to obtain a corresponding third analysis data table;

[0168] Among them, the third analysis data table consists of two third analysis data records; the record fields of the third analysis data records consist of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a full width at half maximum field, and a half-peak wavelength range field; the two third analysis data records correspond one by one to the third and fourth molecular structures;

[0169] Specifically, it includes: Step 61, extract the corresponding third molecular structure, fourth molecular structure, second set temperature, second set light intensity, and third wavelength sequence from the current task configuration parameters; and use the second set temperature and second set light intensity as the current temperature x and light intensity y;

[0170] Step 62, and use the third and fourth molecular structures as the corresponding current molecular structure M in turn; and perform a round of traversal on all the third wavelengths in the third wavelength sequence; and during this round of traversal, use the currently traversed third wavelength as the corresponding current wavelength z; and input the current molecular structure M, temperature x, light intensity y, and wavelength z into the emission intensity prediction model for emission intensity prediction processing, and use the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelength and curve point emission intensity to form a corresponding curve point; and at the end of this round of traversal, sort all the curve points corresponding to the current molecular structure M in ascending order of the curve point wavelength to form a corresponding curve data sequence;

[0171] Step 63, and use the third and fourth molecular structures as the corresponding current molecular structure in turn; and use the second set temperature, second set light intensity, and the curve data sequence corresponding to the current molecular structure as the corresponding current set temperature, current set light intensity, and current curve data sequence; and perform maximum peak position, full width at half maximum, and half-peak wavelength range analysis based on the current curve data sequence to obtain the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range; and use the current molecular structure, current set temperature, current set light intensity, current curve data sequence, current maximum peak position, current full width at half maximum, and current half-peak wavelength range as the corresponding molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak position field, full width at half maximum field, and half-peak wavelength range field to form a corresponding third analysis data record;

[0172] Step 64, and form the corresponding third analysis data table from the two obtained third analysis data records.

[0173] Here, the data structure of the third analysis data table obtained in the embodiments of the present invention can be intuitively understood through Figure 3 the following.

[0174] It should be noted that in sub-step 43 of step 4, sub-step 53 of step 5, and sub-step 63 of step 6 above, the same technical means for analyzing the maximum peak position, full width at half maximum, and half-peak wavelength range based on the curve data sequence is used. The detailed processing steps of this technical means are as follows:

[0175] Analyze the maximum peak position, full width at half maximum, and half-peak wavelength range based on the current curve data sequence to obtain the corresponding current maximum peak position, current full width at half maximum, and current half-peak wavelength range, specifically including:

[0176] Step A1, construct a two-dimensional coordinate plane with the emission intensity I as the ordinate and the wavelength λ as the abscissa as the corresponding emission spectrum curve plane; and perform two-dimensional curve fitting on the emission spectrum curve plane based on the current curve data sequence and use the obtained fitting curve as the corresponding current emission spectrum curve;

[0177] Figure 4 It is a schematic diagram of the emission spectrum curve, maximum peak position, left half-peak position, right half-peak position, full width at half maximum, and half-peak wavelength range provided in the first embodiment of the present invention. Here, it can be based on Figure 4 to have an intuitive understanding of the emission spectrum curve plane and the current emission spectrum curve;

[0178] Step A2, and use the maximum emission intensity position on the current emission spectrum curve as the corresponding maximum peak position P max ; and record the emission intensity and wavelength corresponding to the maximum peak position P max as the corresponding emission intensity I max and wavelength λ max ;

[0179] Here, it can be based on Figure 4 to have an intuitive understanding of the maximum peak position P max and its corresponding emission intensity I max and wavelength λ max ;

[0180] Step A3, and record the curve parts on the current emission spectrum curve on the left and right sides of the maximum peak position P max as the corresponding left curve and right curve respectively; and record the curve positions with the emission intensity of I max / 2 on the left and right curves as the corresponding left and right candidate positions; and use the left candidate position farthest from the maximum peak position P max as the corresponding left half-peak position p1, and the left candidate position farthest from the maximum peak position P maxThe rightmost candidate site among the farthest ones is used as the corresponding right half-peak site p2; and the corresponding wavelengths of the left and right half-peak sites p1 and p2 are denoted as the corresponding wavelengths λ1 and λ2; and the width of the wavelength range between the wavelengths λ1 and λ2 is identified to obtain the corresponding full width at half maximum W = λ2 - λ1; and the wavelength range with wavelengths λ1 and λ2 as the left and right boundaries is used as the corresponding half-peak wavelength range [λ1, λ2].

[0181] Here, it can be based on Figure 4 to have an intuitive understanding of the left half-peak site p1 and its corresponding wavelength λ1, the right half-peak site p2 and its corresponding wavelength λ2, the full width at half maximum W, and the half-peak wavelength range [λ1, λ2].

[0182] Step A4, and take the maximum peak site P obtained this time max , the full width at half maximum W, and the half-peak wavelength range [λ1, λ2] as the corresponding current maximum peak site, current full width at half maximum, and current half-peak wavelength range.

[0183] Step 7, feed back the first, second, or third analysis data table obtained this time to the current user.

[0184] Here, after obtaining the first analysis data table, the user can identify the shape of the emission spectrum curve of the organic molecule corresponding to the first molecular structure and the corresponding relationship between the maximum emission intensity / maximum peak position / full width at half maximum / half-peak wavelength range and the ambient light intensity. Based on such recognition results, the user can directly analyze whether the emission spectrum curve of the current organic molecule will undergo blue / red shift and the corresponding blue / red shift change trend when the ambient light intensity changes; after obtaining the second analysis data table, the user can identify the shape of the emission spectrum curve of the organic molecule corresponding to the second molecular structure and the corresponding relationship between the maximum emission intensity / maximum peak position / full width at half maximum / half-peak wavelength range and the ambient temperature. Based on such recognition results, the user can directly analyze whether the emission spectrum curve of the current organic molecule will undergo blue / red shift and the corresponding blue / red shift change trend when the ambient temperature changes; after obtaining the third analysis data table, the user can identify the differential characteristics of the emission spectrum curve shapes and characteristics such as the maximum emission intensity / maximum peak position / full width at half maximum / half-peak wavelength range of the two organic molecules corresponding to the third and fourth molecular structures. Based on such recognition results, the user can directly compare the advantages and disadvantages of the light emission properties of these two organic molecules.

[0185] Figure 5 This is the module structure diagram of a prediction and analysis device for an emission spectrum curve provided in the second embodiment of the present invention. This device is a terminal device or a server for implementing the foregoing method embodiment, or can also be a device that enables the foregoing terminal device or server to implement the foregoing method embodiment. For example, this device can be a device or a chip system of the foregoing terminal device or server. AsFigure 5 As shown in the figure, the device includes: a model construction module 201, a model training module 202, a prediction analysis task data receiving module 203, a prediction analysis task processing module 204, and a prediction analysis task feedback module 205.

[0186] The model construction module 201 is used to construct an emission intensity prediction model for predicting emission intensity; the emission intensity prediction model is used to perform emission intensity prediction processing based on the molecular structure M, temperature x, light intensity y, and wavelength z input into the model and output the corresponding predicted emission intensity u.

[0187] The model training module 202 is used to construct a first data set for model training through data collection; and perform model training on the emission intensity prediction model according to the first data set.

[0188] The prediction analysis task data receiving module 203 is used to, after the model training is completed, receive the prediction analysis task input by the user and extract the corresponding task type and task configuration parameters from it; the prediction analysis task includes a task type and task configuration parameters; the task type includes the first, second, and third types; when the task type is the first type, the corresponding task configuration parameters include the first molecular structure, the first set temperature, the first light intensity sequence, and the first wavelength sequence; when the task type is the second type, the corresponding task configuration parameters include the second molecular structure, the first set light intensity, the first temperature sequence, and the second wavelength sequence; when the task type is the third type, the corresponding task configuration parameters include the third molecular structure, the fourth molecular structure, the second set temperature, the second set light intensity, and the third wavelength sequence.

[0189] The prediction analysis task processing module 204 is used to, when the task type is the first type, under the set conditions where multiple temperatures are consistent but the light intensities are different, predict multiple emission spectral curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters, and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding first analysis data table.

[0190] The prediction analysis task processing module 204 is also used to, when the task type is the second type, under the set conditions where multiple light intensities are consistent but the temperatures are different, predict multiple emission spectral curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters, and perform data analysis on the maximum peak position, full width at half maximum, and half-peak wavelength range of each predicted curve to obtain the corresponding second analysis data table.

[0191] The prediction analysis task processing module 204 is further configured to, when the task type is the third type, respectively predict the emission spectral curves of two specified molecules based on the emission intensity prediction model and the task configuration parameters under the set conditions where the temperature and the light intensity are both kept consistent, and perform data analysis on the maximum peak sites, full width at half maximum, and half peak wavelength range of the two predicted curves to obtain the corresponding third analysis data table.

[0192] The prediction analysis task feedback module 205 is configured to feedback the first, second, or third analysis data table obtained this time to the current user.

[0193] The prediction analysis device for the emission spectral curve provided by the embodiment of the present invention can execute the method steps in the above method embodiment, and its implementation principle and technical effect are similar, which will not be elaborated here.

[0194] It should be noted that it should be understood that the division of each module of the above device is only a logical function division. In actual implementation, it can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by a processing element; they can also all be implemented in the form of hardware; or some modules can be implemented in the form of software called by a processing element, and some modules can be implemented in the form of hardware. For example, the model construction module can be a separately established processing element, or can be integrated in a certain chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a certain processing element of the above device to perform the functions of the above determined module. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together or can be independently implemented. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the integrated logic circuit in the processor element or the instruction in the form of software.

[0195] For example, the above modules can be one or more integrated circuits configured to implement the above methods. For example: one or more Application Specific Integrated Circuits (ASICs), or, one or more Digital Signal Processors (DSPs), or, one or more Field Programmable Gate Arrays (FPGAs), etc. Again, when a certain module above is implemented in the form of a processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processors that can call program code. Again, these modules can be integrated together and implemented in the form of a System-on-a-chip (SOC).

[0196] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the foregoing method embodiments are generated in whole or in part. The above computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the above computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.). The above computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The above available medium can be a magnetic medium (for example, floppy disk, hard disk, magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, solid state disk (SSD)), etc.

[0197] Figure 6 This is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present invention. The electronic device can be a terminal device or a server that implements the method of the foregoing embodiments, or a terminal device or a server that is connected to the foregoing terminal device or server and implements the method of the foregoing embodiments. As Figure 6As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver operations of the transceiver 303. Various instructions may be stored in the memory 302 for completing various processing functions and implementing the processing steps described in the method of the foregoing embodiments. Preferably, the electronic device according to the embodiment of the present invention further includes: a power supply 304, a system bus 305, and a communication port 306. The system bus 305 is used to implement communication connections between components. The above communication port 306 is used for the electronic device to connect and communicate with other peripherals.

[0198] The Figure 6 system bus 305 mentioned in may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to implement communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include a Random Access Memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0199] The above-mentioned processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), a Graphics Processing Unit (GPU), etc.; it may also be a Digital Signal Processor (DSP), 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.

[0200] It should be noted that the embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored, and when they run on a computer, the computer is caused to execute the methods and processing procedures provided in the above embodiments.

[0201] An embodiment of the present invention provides a method, apparatus, electronic device, and computer-readable storage medium for predicting and analyzing emission spectral curves. As can be seen from the above, an embodiment of the present invention customizes an emission intensity prediction model for predicting emission intensity based on the molecular structure, temperature, light intensity, and wavelength input into the model, and trains it with a first data set obtained by big data collection; after the training is completed, it parses the prediction and analysis tasks (composed of task types and task configuration parameters) input by the user; if the task type is the first type, under the set conditions where multiple temperatures are the same but the light intensities are different, based on the emission intensity prediction model and task configuration parameters, predict multiple emission spectral curves of a specified molecule and analyze the three types of curve characteristics (maximum peak position, full width at half maximum, and half-peak wavelength range) of each predicted curve to obtain a first analysis data table and feedback it to the current user; if the task type is the second type, under the set conditions where multiple light intensities are the same but the temperatures are different, based on the emission intensity prediction model and task configuration parameters, predict multiple emission spectral curves of a specified molecule and analyze the three types of curve characteristics of each predicted curve to obtain a second analysis data table and feedback it to the current user; if the task type is the third type, under the set conditions where the temperature and light intensity are both the same, based on the emission intensity prediction model and task configuration parameters, respectively predict the emission spectral curves of two specified molecules and analyze the three types of curve characteristics of the two predicted curves to obtain a third analysis data table and feedback it to the current user. On the one hand, an embodiment of the present invention reduces the processing complexity of the prediction and analysis task, shortens the processing time, and improves the processing efficiency through a customized emission intensity prediction model; on the other hand, it improves the convenience of data analysis through three types of customized emission spectral curve analysis means.

[0202] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented in hardware, software modules executed by a processor, or a combination of both. The software modules may be placed in a random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0203] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting and analyzing an emission spectrum curve, characterized in that: The method comprises: Constructing an emission intensity prediction model for emission intensity prediction; the emission intensity prediction model is used to perform emission intensity prediction processing according to the molecular structure M, temperature x, light intensity y and wavelength z input by the model and output the corresponding predicted emission intensity u; Constructing a first data set for model training through data collection; and performing model training on the emission intensity prediction model according to the first data set; After the model training is completed, the prediction and analysis task input by the user is received, and the corresponding task type and task configuration parameters are extracted therefrom; the prediction and analysis task includes the task type and the task configuration parameters; the task type includes the first, second and third types; when the task type is the first type, the corresponding task configuration parameters include the first molecular structure, the first set temperature, the first light intensity sequence and the first wavelength sequence; when the task type is the second type, the corresponding task configuration parameters include the second molecular structure, the first set light intensity, the first temperature sequence and the second wavelength sequence; when the task type is the third type, the corresponding task configuration parameters include the third molecular structure, the fourth molecular structure, the second set temperature, the second set light intensity and the third wavelength sequence; If the task type is the first type, then under the setting conditions that multiple temperatures are kept constant but the light intensities are different, multiple emission spectrum curves of a specified molecule are predicted based on the emission intensity prediction model and the task configuration parameters, and data analysis is performed on the maximum peak position, half-maximum full width and half-maximum wavelength range of each predicted curve to obtain a corresponding first analysis data table; If the task type is the second type, then under the setting conditions that multiple light intensities are kept consistent but temperatures are different, multiple emission spectrum curves of a specified molecule are predicted based on the emission intensity prediction model and the task configuration parameters, and data analysis is performed on the maximum peak position, half-maximum full width and half-maximum wavelength range of each predicted curve to obtain a corresponding second analysis data table; If the task type is the third type, then under the setting conditions that the temperature and the light intensity are kept consistent, the emission spectrum curves of the two specified molecules are predicted respectively based on the emission intensity prediction model and the task configuration parameters, and the maximum peak position, half-maximum full width and half-maximum wavelength range of the two predicted curves are analyzed to obtain the corresponding third analysis data table; The first, second or third analysis data table obtained this time is fed back to the current user.

2. The method for predicting and analyzing emission spectrum curve according to claim 1, characterized in that: The molecular structure M includes a plurality of atoms m i , 1≤atomic index i≤N M , N M is the total number of atoms in the molecular structure M; each of the atoms m i The atomic parameters of are composed of the corresponding atomic type and three-dimensional atomic coordinates; The first data set includes a plurality of first data records; the first data record includes a first training molecular structure, a first training temperature, a first training light intensity, a first training wavelength sequence and a first emission intensity label sequence; the data structure type of the first training molecular structure is consistent with the data structure type of the molecular structure M; the first training wavelength sequence consists of a plurality of first training wavelengths; the first emission intensity label sequence consists of a plurality of first emission intensity labels; the first emission intensity label corresponds one to one with the first training wavelength; When the task type is the first type, the first light intensity sequence of the task configuration parameters is formed by sorting a plurality of first light intensities in ascending order of intensity; when the task type is the second type, the first temperature sequence of the task configuration parameters is formed by sorting a plurality of first temperatures in ascending order of temperature; when the task type is the first, second or third type, the corresponding first, second or third wavelength sequence in the task configuration parameters is formed by sorting a plurality of corresponding first, second or third wavelengths in ascending order of wavelength; The data structure types of the first, second, third and fourth molecular structures are consistent with the data structure type of the molecular structure M; The first analysis data table is composed of a plurality of first analysis data records; the record fields of the first analysis data record are composed of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a half-peak full width field, and a half-peak wavelength range field; the total number of records of the first analysis data record is consistent with the total number of the first light intensities of the corresponding first light intensity sequence; The second analysis data table is composed of a plurality of second analysis data records; the record fields of the second analysis data records are composed of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a half-peak full width field and a half-peak wavelength range field; the total number of records of the second analysis data records is consistent with the total number of the first temperatures of the corresponding first temperature sequence; The third analysis data table consists of two third analysis data records; the record fields of the third analysis data record consist of a molecular structure field, a temperature field, a light intensity field, a curve data sequence field, a maximum peak position field, a half-peak full width field and a half-peak wavelength range field; the two third analysis data records correspond one-to-one to the third and fourth molecular structures.

3. The method for predicting and analyzing emission spectrum curve according to claim 1, characterized in that: The first model input end of the emission intensity prediction model is used to receive the molecular structure M input by the model, the second model input end is used to receive the temperature x input by the model, the third model input end is used to receive the light intensity y input by the model, and the fourth model input end is used to receive the wavelength z input by the model; the model output end of the emission intensity prediction model is used to output the corresponding predicted emission intensity u; The emission intensity prediction model includes an embedded coding module, a molecular structure coding module, a first Gaussian kernel module, a first vector mapping module, a second Gaussian kernel module, a second vector mapping module, a third Gaussian kernel module, a third vector mapping module, a feature fusion module and an emission intensity prediction module; the molecular structure coding module is implemented based on a Uni-Mol model or a graph neural network that satisfies SE (3) equivariance; the first, second and third vector mapping modules and the emission intensity prediction module are all implemented based on an MLP model; The input end of the embedded coding module is connected to the input end of the first model, and the output end is connected to the input end of the molecular structure coding module; the output end of the molecular structure coding module is connected to the first input end of the feature fusion module; the input end of the first Gaussian kernel module is connected to the input end of the second model, and the output end is connected to the input end of the first vector mapping module; the output end of the first vector mapping module is connected to the second input end of the feature fusion module; the input end of the second Gaussian kernel module is connected to the input end of the third model, and the output end is connected to the input end of the second vector mapping module; the output end of the second vector mapping module is connected to the third input end of the feature fusion module; the input end of the third Gaussian kernel module is connected to the input end of the fourth model, and the output end is connected to the input end of the third vector mapping module; the output end of the third vector mapping module is connected to the fourth input end of the feature fusion module; the output end of the feature fusion module is connected to the input end of the emission intensity prediction module; the output end of the emission intensity prediction module is connected to the output end of the model; The embedded coding module is used to encode the molecular structure M input by the model according to the embedded coding rule of the molecular structure coding module and send the coding result to the molecular structure coding module, specifically: when the molecular structure coding module is implemented based on the Uni-Mol model, the embedded coding module encodes each of the atoms m of the molecular structure M according to the one-hot coding rule of the atom type of the Uni-Mol model. i Perform one-hot encoding on the atom type to obtain the corresponding atom encoding vector, and use the obtained N M The atomic coding vectors form a corresponding atomic coding tensor, and according to the paired feature initialization coding rule of the Uni-Mol model, all three-dimensional atomic coordinates of the molecular structure M are encoded to obtain a tensor shape of N M ×N M The pairwise coding tensor of the molecular structure M is obtained, and the obtained atomic coding tensor and the paired coding tensor are sent to the molecular structure coding module; when the molecular structure coding module is implemented based on a graph neural network that satisfies SE (3) equivariance, the embedded coding module is composed of each of the atoms m of the molecular structure M. i As a corresponding first node and obtained by N M The first nodes form a corresponding first node set, and a corresponding directed edge is made from each of the first nodes to any other node as the corresponding first edge, and the edge relationship of each first edge is set to two corresponding atoms m i The coordinate difference vector between them is obtained by M ×(N M -1) first edges form a corresponding first edge set, and the obtained first node set and the first edge set form a corresponding first graph structure and send it to the molecular structure encoding module; When the molecular structure encoding module is based on the Uni-Mol model, it is used to extract atomic features and paired features according to the input atomic encoding tensor and the paired encoding tensor to obtain the corresponding latent feature tensor H and send it to the feature fusion module; the latent feature tensor H is composed of N M Hidden feature vector h i The latent feature vector h i With the atom m i One to one correspondence; When the molecular structure encoding module is implemented based on a graph neural network that satisfies SE (3) equivariance, it is used to perform feature update processing on the node features and edge features of each of the first nodes and each of the first edges in the input first graph structure through the message passing mechanism of the graph neural network to obtain the corresponding latent feature tensor H and send it to the feature fusion module; The first Gaussian kernel module is used to preset multiple first Gaussian kernel functions in the module; The first Gaussian kernel module is further used to perform a corresponding kernel function calculation on the temperature x input by the model according to each of the first Gaussian kernel functions and use the calculated numerical result as the corresponding first result numerical value; and form a corresponding temperature encoding vector from all the obtained first result numerical values ​​and send it to the first vector mapping module; the vector length of the temperature encoding vector is consistent with the total number of the first Gaussian kernel functions; The first vector mapping module is used to transform the latent feature vector h i The corresponding vector space is used as the target vector space; and the temperature encoding vector is mapped to the target vector space to obtain a corresponding temperature mapping vector and sent to the feature fusion module; the temperature mapping vector and the latent feature vector h i The characteristic dimensions remain consistent; The second Gaussian kernel module is used to preset multiple second Gaussian kernel functions in the module; The second Gaussian kernel module is further used to perform a corresponding kernel function calculation on the illumination intensity y input by the model according to each of the second Gaussian kernel functions and use the calculated numerical result as the corresponding second result numerical value; and form a corresponding illumination intensity encoding vector from all the obtained second result numerical values ​​and send it to the second vector mapping module; the vector length of the illumination intensity encoding vector is consistent with the total number of the second Gaussian kernel functions; The second vector mapping module is used to transform the latent feature vector h i The corresponding vector space is used as the target vector space; and the illumination intensity encoding vector is mapped into the target vector space to obtain a corresponding illumination intensity mapping vector and sent to the feature fusion module; the illumination intensity mapping vector and the latent feature vector h i The characteristic dimensions remain consistent; The third Gaussian kernel module is used to preset multiple third Gaussian kernel functions in the module; The third Gaussian kernel module is further used to perform a corresponding kernel function calculation on the wavelength z input by the model according to each of the third Gaussian kernel functions and use the calculated numerical result as the corresponding third result numerical value; and form a corresponding wavelength coding vector from all the obtained third result numerical values ​​and send it to the third vector mapping module; the vector length of the wavelength coding vector is consistent with the total number of the third Gaussian kernel functions; The third vector mapping module is used to transform the latent feature vector h i The corresponding vector space is used as the target vector space; and the wavelength encoding vector is mapped to the target vector space to obtain a corresponding wavelength mapping vector and send it to the feature fusion module; the wavelength mapping vector and the latent feature vector h i The characteristic dimensions remain consistent; The feature fusion module is used to combine the latent feature vectors h of the latent feature tensor H i The temperature mapping vector, the light intensity mapping vector and the wavelength mapping vector are sequentially spliced ​​and the obtained spliced ​​vector is used as a corresponding fusion feature vector r i ; and by the obtained N M The fused feature vector r i The corresponding fused feature tensor R is sent to the emission intensity prediction module; The emission intensity prediction module is used to perform emission intensity regression calculation according to the input fusion feature tensor R and output the calculation result as the corresponding predicted emission intensity u.

4. The method for predicting and analyzing emission spectrum curve according to claim 2, characterized in that: The step of constructing a first data set for model training through data collection specifically includes: Step 41, collecting big data of the organic molecular structure of the organic photovoltaic material used for the light-emitting device and its corresponding emission spectrum curve as well as the temperature and light intensity information corresponding to the current curve through multiple data collection channels to obtain a plurality of first collected data; The multiple data collection channels include at least multiple public molecular databases in the field of organic photovoltaic materials, all public literature and experimental databases in the field of organic photovoltaic materials; The first acquisition data at least includes a first acquisition structure, a first acquisition curve data sequence, a first acquisition temperature and a first acquisition light intensity; the first acquisition structure is a molecular structure of an organic molecule, and its data structure is consistent with the data structure type of the molecular structure M; the first acquisition curve data sequence is a discretized sampling point data sequence corresponding to the emission spectrum curve of the current organic molecule, and is composed of a plurality of first sampling points, each of which is composed of a sampling point wavelength and its corresponding sampling point emission intensity; the first acquisition temperature and the first acquisition light intensity are respectively the ambient temperature and the ambient light intensity corresponding to the current emission spectrum curve; Step 42, perform a round of traversal on all the first collected data; and in this round of traversal, take the first collected data currently traversed as the corresponding current collected data; and take the first collected structure, the first collected temperature and the first collected light intensity of the current collected data as the corresponding first training molecular structure, the first training temperature and the first training light intensity; and take the wavelength of each sampling point of the current collected data as a corresponding first training wavelength, and the emission intensity of each sampling point as a corresponding first emission intensity label; and sort all the first training wavelengths obtained in the sorting order of the first sampling points to obtain the corresponding first training wavelength sequence; and sort all the first emission intensity labels obtained in the sorting order of the first sampling points to obtain the corresponding first emission intensity label sequence; and form a corresponding first data record by the first training molecular structure, the first training temperature, the first training light intensity, the first training wavelength sequence and the first emission intensity label sequence corresponding to the current collected data; and at the end of this round of traversal, all the first data records obtained form the corresponding first data set.

5. The method for predicting and analyzing emission spectrum curve according to claim 2, characterized in that: The performing model training on the emission intensity prediction model according to the first data set specifically includes: Step 51, based on a preset first segmentation ratio, the first data set is divided into two sub-data sets which are recorded as a corresponding first training set and a first evaluation set; Wherein, both the first training set and the first evaluation set are composed of a plurality of the first data records; the ratio of the total number of records in the first training set and the first evaluation set satisfies the first segmentation ratio; Step 52, taking the first first data record of the first training set as the corresponding current training record; Step 53, using the first training molecular structure, the first training temperature and the first training light intensity recorded in the current training as the current molecular structure M, the temperature x and the light intensity y; Step 54, the total number of the first training wavelengths in the current training record is counted to obtain the corresponding first total number N1; and a round of traversal is performed on the N1 first training wavelengths in the current training record; and in this round of traversal, the first training wavelength currently traversed is used as the current wavelength z; and the current molecular structure M, the temperature x, the light intensity y and the wavelength z are input into the emission intensity prediction model to perform emission intensity prediction processing and the predicted emission intensity u output by this processing is used as a corresponding predicted intensity y pre,j , 1≤index j≤N1; and taking the first emission intensity label corresponding to the first training wavelength currently traversed in the current training record as a corresponding label intensity y tag,j ; and by the predicted intensity y pre,j and the corresponding label strength y tag,j Form a corresponding first prediction-label pair (y pre,j ,y tag,j ); Step 55: obtain the N1 first prediction-label pairs (y pre,j ,y tag,j ) into the preset first model loss function L a ; and based on the preset first model optimizer, move towards making the first model loss function L a Performing a round of modulation on the model parameters of the emission intensity prediction model in the direction of reaching the minimum value; Among them, the first model loss function L a for: The first model optimizer includes at least an Adam optimizer and an SGD optimizer; Step 56, identifying whether the current training record is the last first data record of the first training set, if yes, go to step 57, if no, take the next first data record of the first training set as the new current training record and return to step 53 to continue training; Step 57, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, take the first data record currently traversed as the corresponding current evaluation record; and take the first training molecular structure, the first training temperature and the first training light intensity of the current evaluation record as the current molecular structure M, the temperature x and the light intensity y; and take each of the first training wavelengths of the current evaluation record as the corresponding current wavelength z in turn, and input the current molecular structure M, the temperature x, the light intensity y and the wavelength z into the emission intensity prediction model to perform emission intensity prediction processing and take the predicted emission intensity u outputted by the current processing as a corresponding first predicted intensity; and record the first emission intensity label corresponding to each first predicted intensity in the current evaluation record as the corresponding first label intensity; and each of the first predicted intensities of the current evaluation record and its corresponding first label intensity constitute a corresponding second prediction-label pair; Step 58: After the current round of traversal of all the first data records of the first evaluation set is completed, the total number of the obtained second prediction-label pairs is counted to obtain a corresponding second total number N2; and the first prediction strength and the first label strength of each second prediction-label pair are recorded as a set of corresponding prediction strengths y pre,k and label intensity y tag,k , 1≤index k≤N2; and the obtained N2 group of prediction strength y pre,k and the label intensity y tag,k Substitute the preset first model evaluation function S a Performing calculation to obtain a corresponding first evaluation value; Among them, the first model evaluation function S a for: Step 59, identifying whether the first evaluation value meets the preset first evaluation value range; if not, returning to step 52 to continue training; if satisfied, stopping training and confirming the end of model training.

6. The method for predicting and analyzing emission spectrum curve according to claim 2, characterized in that: The step of predicting multiple emission spectrum curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters and performing data analysis on the maximum peak position, half-maximum full width and half-maximum wavelength range of each predicted curve to obtain a corresponding first analysis data table specifically includes: Extract the corresponding first molecular structure, the first set temperature, the first illumination intensity sequence and the first wavelength sequence from the current task configuration parameters; and use the first molecular structure and the first set temperature as the current molecular structure M and the temperature x; and count the total number of the first illumination intensities of the first illumination intensity sequence to obtain the corresponding total number N R1 ; And take each of the first illumination intensities of the first illumination intensity sequence as the corresponding current illumination intensity y in turn; and perform a round of traversal on all the first wavelengths of the first wavelength sequence; and during this round of traversal, take the currently traversed first wavelength as the corresponding current wavelength z; and input the current molecular structure M, the temperature x, the illumination intensity y and the wavelength z into the emission intensity prediction model to perform emission intensity prediction processing and take the predicted emission intensity u outputted by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelengths and curve point emission intensities to form a corresponding curve point; and at the end of this round of traversal, all the curve points corresponding to the current illumination intensity y are sorted in ascending order of the curve point wavelengths to form a corresponding curve data sequence; And for the first illumination intensity sequence N R1 The first illumination intensity is traversed for one round; and in this round of traversal, the first illumination intensity currently traversed is used as the corresponding current illumination intensity; and the curve data sequence corresponding to the first molecular structure, the first set temperature and the current illumination intensity is used as the corresponding current molecular structure, the current set temperature and the current curve data sequence; and based on the current curve data sequence, the maximum peak position, the half-peak full width and the half-peak wavelength range are analyzed to obtain the corresponding current maximum peak position, the current half-peak full width and the current half-peak wavelength range; and the current molecular structure, the current set temperature, the current illumination intensity, the current curve data sequence, the current maximum peak position, the current half-peak full width and the current half-peak wavelength range are used as the corresponding molecular structure field, temperature field, illumination intensity field, curve data sequence field, maximum peak position field, half-peak full width field and half-peak wavelength range field to form a corresponding first analysis data record; And in the first illumination intensity sequence N R1 After the current round of traversal of the first light intensity is completed, the obtained N R1 The first analysis data records constitute the corresponding first analysis data table.

7. The method for predicting and analyzing emission spectrum curve according to claim 2, characterized in that: The method of predicting multiple emission spectrum curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters and performing data analysis on the maximum peak position, half-maximum full width and half-maximum wavelength range of each predicted curve to obtain a corresponding second analysis data table specifically includes: Extract the corresponding second molecular structure, the first set light intensity, the first temperature sequence and the second wavelength sequence from the current task configuration parameters; and use the first molecular structure and the first set light intensity as the current molecular structure M and the light intensity y; and count the total number of the first temperatures in the first temperature sequence to obtain the corresponding total number N R2 ; And take each of the first temperatures of the first temperature sequence as the corresponding current temperature x in turn; and perform a round of traversal on all the second wavelengths of the second wavelength sequence; and during this round of traversal, take the currently traversed second wavelength as the corresponding current wavelength z; and input the current molecular structure M, the temperature x, the light intensity y and the wavelength z into the emission intensity prediction model to perform emission intensity prediction processing and take the predicted emission intensity u output by the current processing as a corresponding current predicted emission intensity; and use the current wavelength z and the current predicted emission intensity as a group of corresponding curve point wavelengths and curve point emission intensities to form a corresponding curve point; and at the end of this round of traversal, all the curve points corresponding to the current temperature x are sorted in the order of the curve point wavelengths from small to large to form a corresponding curve data sequence; And for the first temperature series N R2 The first temperature is traversed for one round; and in this round of traversal, the first temperature currently traversed is taken as the corresponding current temperature; and the first molecular structure, the first set light intensity and the curve data sequence corresponding to the current temperature are taken as the corresponding current molecular structure, the current set light intensity and the current curve data sequence; and based on the current curve data sequence, the maximum peak position, the half-peak full width and the half-peak wavelength range are analyzed to obtain the corresponding current maximum peak position, the current half-peak full width and the current half-peak wavelength range; and the current molecular structure, the current temperature, the current set light intensity, the current curve data sequence, the current maximum peak position, the current half-peak full width and the current half-peak wavelength range are used as the corresponding molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak position field, half-peak full width field and half-peak wavelength range field to form a corresponding second analysis data record; And in the first temperature series N R2 At the end of this round of traversal of the first temperature, the N R2 The second analysis data records constitute the corresponding second analysis data table.

8. The method for predicting and analyzing emission spectrum curve according to claim 2, characterized in that: The emission spectrum curves of the two specified molecules are predicted respectively based on the emission intensity prediction model and the task configuration parameters, and data analysis is performed on the maximum peak position, half-maximum full width and half-maximum wavelength range of the two predicted curves to obtain the corresponding third analysis data table, specifically including: Extracting the corresponding third molecular structure, fourth molecular structure, second set temperature, second set light intensity and third wavelength sequence from the current task configuration parameters; and using the second set temperature and the second set light intensity as the current temperature x and the light intensity y; And the third and fourth molecular structures are taken as the corresponding current molecular structure M in turn; and a round of traversal is performed on all the third wavelengths of the third wavelength sequence; and in this round of traversal, the currently traversed third wavelength is taken as the corresponding current wavelength z; and the current molecular structure M, the temperature x, the light intensity y and the wavelength z are input into the emission intensity prediction model to perform emission intensity prediction processing and the predicted emission intensity u outputted by the current processing is taken as a corresponding current predicted emission intensity; and the current wavelength z and the current predicted emission intensity are used as a group of corresponding curve point wavelengths and curve point emission intensities to form a corresponding curve point; and at the end of this round of traversal, all the curve points corresponding to the current molecular structure M are sorted in the order of the curve point wavelengths from small to large to form a corresponding curve data sequence; And the third and fourth molecular structures are used as the corresponding current molecular structures in sequence; and the second set temperature, the second set light intensity and the curve data sequence corresponding to the current molecular structure are used as the corresponding current set temperature, the current set light intensity and the current curve data sequence; and based on the current curve data sequence, the maximum peak position, the half-peak full width and the half-peak wavelength range are analyzed to obtain the corresponding current maximum peak position, the current half-peak full width and the current half-peak wavelength range; and the current molecular structure, the current set temperature, the current set light intensity, the current curve data sequence, the current maximum peak position, the current half-peak full width and the current half-peak wavelength range are used as the corresponding molecular structure field, temperature field, light intensity field, curve data sequence field, maximum peak position field, half-peak full width field and half-peak wavelength range field to form a corresponding third analysis data record; The two obtained third analysis data records form a corresponding third analysis data table.

9. The method for predicting and analyzing an emission spectrum curve according to any one of claims 6 to 8, characterized in that: The maximum peak position, half-maximum full width and half-maximum wavelength range analysis based on the current curve data sequence to obtain the corresponding current maximum peak position, current half-maximum full width and current half-maximum wavelength range specifically includes: A two-dimensional coordinate plane is constructed with the emission intensity I as the ordinate and the wavelength λ as the abscissa as the corresponding emission spectrum curve plane; and a two-dimensional curve fitting is performed on the emission spectrum curve plane based on the current curve data sequence and the obtained fitting curve is used as the corresponding current emission spectrum curve; The maximum emission intensity position on the current emission spectrum curve is taken as the corresponding maximum peak position P max ; and the maximum peak position P max The corresponding emission intensity and wavelength are recorded as the corresponding emission intensity I max and wavelength λ max ; and the maximum peak position P on the current emission spectrum curve max The curve parts on the left and right are respectively recorded as the corresponding left curve and right curve; and the emission intensity on the left and right curves is I max / 2 curve points are recorded as the corresponding left and right candidate points; and the distance from the maximum peak point P max The farthest left candidate site is used as the corresponding left half peak site p1, and the distance from the maximum peak site P max The farthest right candidate site is taken as the corresponding right half-peak site p2; and the corresponding wavelengths of the left and right half-peak sites p1 and p2 are recorded as the corresponding wavelengths λ1 and λ2; and the wavelength range width between the wavelengths λ1 and λ2 is identified to obtain the corresponding half-peak full width W=λ2-λ1; and the wavelength range with the wavelengths λ1 and λ2 as the left and right boundaries is taken as the corresponding half-peak wavelength range [λ1,λ2]; The maximum peak position P obtained this time max , the half-maximum full width W and the half-maximum wavelength range [λ1, λ2] as the corresponding current maximum peak position, the current half-maximum full width and the current half-maximum wavelength range.

10. A device for executing the prediction and analysis method of emission spectrum curve according to any one of claims 1 to 9, characterized in that: The device comprises: a model building module, a model training module, a prediction and analysis task data receiving module, a prediction and analysis task processing module and a prediction and analysis task feedback module; The model building module is used to build an emission intensity prediction model for emission intensity prediction; the emission intensity prediction model is used to perform emission intensity prediction processing according to the molecular structure M, temperature x, light intensity y and wavelength z input by the model and output the corresponding predicted emission intensity u; The model training module is used to construct a first data set for model training through data collection; and perform model training on the emission intensity prediction model according to the first data set; The prediction and analysis task data receiving module is used to receive the prediction and analysis task input by the user after the model training is completed, and extract the corresponding task type and task configuration parameters therefrom; the prediction and analysis task includes the task type and the task configuration parameters; the task type includes the first, second and third types; when the task type is the first type, the corresponding task configuration parameters include the first molecular structure, the first set temperature, the first light intensity sequence and the first wavelength sequence; when the task type is the second type, the corresponding task configuration parameters include the second molecular structure, the first set light intensity, the first temperature sequence and the second wavelength sequence; when the task type is the third type, the corresponding task configuration parameters include the third molecular structure, the fourth molecular structure, the second set temperature, the second set light intensity and the third wavelength sequence; The prediction and analysis task processing module is used for predicting multiple emission spectrum curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters when the task type is the first type, under the setting conditions that multiple temperatures are kept consistent but the light intensity is different, and performing data analysis on the maximum peak position, half-maximum full width and half-maximum wavelength range of each predicted curve to obtain a corresponding first analysis data table; The prediction and analysis task processing module is also used for predicting multiple emission spectrum curves of a specified molecule based on the emission intensity prediction model and the task configuration parameters when the task type is the second type, under the setting conditions that multiple light intensities are kept consistent but the temperatures are different, and performing data analysis on the maximum peak position, half-maximum full width and half-maximum wavelength range of each predicted curve to obtain a corresponding second analysis data table; The prediction and analysis task processing module is also used for, when the task type is the third type, under the set conditions that the temperature and the light intensity are kept consistent, respectively predicting the emission spectrum curves of two specified molecules based on the emission intensity prediction model and the task configuration parameters, and performing data analysis on the maximum peak position, half-maximum full width and half-maximum wavelength range of the two predicted curves to obtain a corresponding third analysis data table; The prediction analysis task feedback module is used to feed back the first, second or third analysis data table obtained this time to the current user.

11. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 9; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 9.

Images