CN116110505A - Flow chemistry process optimization method based on multi-objective Bayesian optimization - Google Patents
Flow chemistry process optimization method based on multi-objective Bayesian optimization Download PDFInfo
- Publication number
- CN116110505A CN116110505A CN202211535581.XA CN202211535581A CN116110505A CN 116110505 A CN116110505 A CN 116110505A CN 202211535581 A CN202211535581 A CN 202211535581A CN 116110505 A CN116110505 A CN 116110505A
- Authority
- CN
- China
- Prior art keywords
- objective
- function
- experimental
- initial
- reaction parameters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 48
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000005111 flow chemistry technique Methods 0.000 title claims abstract description 18
- 238000006243 chemical reaction Methods 0.000 claims abstract description 56
- 238000005070 sampling Methods 0.000 claims abstract description 32
- 238000002474 experimental method Methods 0.000 claims abstract description 29
- 238000001311 chemical methods and process Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 4
- 239000000376 reactant Substances 0.000 claims description 3
- 238000013401 experimental design Methods 0.000 claims description 2
- 238000013383 initial experiment Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 46
- 238000011161 development Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- WZCQRUWWHSTZEM-UHFFFAOYSA-N 1,3-phenylenediamine Chemical compound NC1=CC=CC(N)=C1 WZCQRUWWHSTZEM-UHFFFAOYSA-N 0.000 description 1
- NOGFHTGYPKWWRX-UHFFFAOYSA-N 2,2,6,6-tetramethyloxan-4-one Chemical compound CC1(C)CC(=O)CC(C)(C)O1 NOGFHTGYPKWWRX-UHFFFAOYSA-N 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 239000003054 catalyst Substances 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 229940018564 m-phenylenediamine Drugs 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012803 optimization experiment Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 239000002904 solvent Substances 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/10—Analysis or design of chemical reactions, syntheses or processes
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
Landscapes
- Chemical & Material Sciences (AREA)
- Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Theoretical Computer Science (AREA)
- Crystallography & Structural Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a flow chemistry process optimization method based on multi-objective Bayesian optimization, which can optimize two objectives, and comprises the following steps: determining the value range of each reaction parameter according to different flow chemistry experiments, constructing an initial sampling set according to the preset value range of the reaction parameter, obtaining an initial data set by using an objective function or a manual experiment, constructing a prediction model by a Gaussian model according to initial experimental data, updating the model according to experimental data updated by Bayesian posterior, recommending the next sampling point in the preset range by an acquisition function according to a prediction result, and outputting optimized parameters in a form of a chart after the acquisition function reaches the preset requirement. The invention can reduce the experiment times to a certain extent, lower the experiment cost, and obtain the experiment information as much as possible by using the least experiment times.
Description
Technical Field
The application relates to the field of flow chemical process optimization, in particular to a flow chemical process optimization method based on multi-objective Bayesian optimization.
Background
The chemical synthesis becomes more efficient, safe and automatic due to the appearance of flow chemistry, the continuous reactor can accurately control the residence time, the reaction temperature and the molar flow ratio of reactants, so that the repeatability of experiments is greatly improved, and the continuous flow technology can be used together with online analysis and detection, so that the degree of automation of the continuous flow technology is greatly improved, and the application field is wider. However, how to obtain the optimal experimental conditions is still a complex and expensive problem, and the chemist needs to spend a lot of time to test and optimize, and needs to continuously screen the reaction conditions, by changing the temperature, the reaction time, the substrate concentration, the catalyst, the solvent, etc., wherein the combination amount is very huge, and it takes a lot of time, and when the experiment needs to use expensive reagents, the budget limits the evaluation amount of the optimization experiment, and how to obtain as much experimental information as possible with the least number of experiments is critical, so that more and more machine learning methods are applied to optimizing the process parameters. Because there is a clear functional relationship (kinetic equation) between the reaction conditions and the experimental results, we can optimize the reaction parameters using bayesian optimization.
Bayesian optimization is a global optimization strategy, which is applicable to optimizing black box functions, and has two core parts: the probability agent model consists of prior distribution and an observation model, is used for agent unknown objective functions, the acquisition function selects the next evaluation point after balanced development and exploration, exploration refers to selecting a point far from a known point as a point of the next iteration, development refers to selecting a point close to the known point as a point of the next iteration, if the acquisition function is biased to exploration, the convergence is poor, the accuracy of an optimal solution cannot be ensured, and if the acquisition function is biased to development, the acquisition function is easy to fall into local optimal.
Bayesian optimization can be applied to multi-objective optimization, which refers to mutual constraint among objects, the improvement of one object often comes at the cost of losing the performance of other objects, so that we cannot find a set of experimental conditions to obtain the optimal value of both objects, in most cases, the multi-objective evaluation cost is high, the two objects must be balanced by using the least experiment as much as possible, the solution of the multi-objective optimization problem is a set of non-dominant points, and the optimization of one object does not have adverse effect on the other object.
When a new evaluation point is selected by the traditional acquisition function, unbalanced development and exploration are easy to occur, and the situation of over development or over exploration occurs, so that the optimization efficiency is reduced, and the Bayesian optimization is caused to fall into local optimum.
Disclosure of Invention
The invention provides a flow chemistry process optimization method based on multi-objective Bayesian optimization, which adopts advanced acquisition functions and aims at improving the problems.
In order to achieve the above object, the specific steps of the present invention include the steps of:
s1, determining the range of the values of experimental reaction parameters.
S2, carrying out initial sampling according to the value range of the reaction parameters in the step S1.
And S3, carrying out actual experiments or calculating by utilizing an objective function according to the initial sampling in the step S2 to obtain experimental results, and summarizing the sampled reaction parameters and the corresponding experimental results to obtain an initial experimental data set.
S4, constructing a Gaussian model according to the initial experimental data set in the step S3, and predicting experimental results
S5, searching optimal reaction parameters by utilizing an acquisition function according to the prediction result obtained in the step S4, carrying out an actual experiment on the selected new sampling points or calculating by utilizing an objective function to obtain an experiment result, updating an initial experiment data set, detecting whether the iteration number reaches a preset requirement, returning to the step S4 if the detection result is not met, and otherwise outputting the optimal reaction parameters in a chart format.
The method can optimize two targets, and comprises the following steps: determining the value range of each reaction parameter according to different flow chemistry experiments, constructing an initial sampling set according to the preset value range of the reaction parameter, obtaining an initial data set by using an objective function or a manual experiment, constructing a prediction model by a Gaussian model according to initial experimental data, updating the model according to experimental data updated by Bayesian posterior, recommending the next sampling point in the preset range by an acquisition function according to a prediction result, and outputting optimized parameters in a form of a chart after the acquisition function reaches the preset requirement. The invention can reduce the experiment times to a certain extent, lower the experiment cost, and obtain the experiment information as much as possible by using the least experiment times.
In step S1, the reaction parameters are experimental reaction conditions, and the reaction parameters comprise residence time, temperature and reactant concentration.
Further, the step S2 specifically includes: and (3) extracting initial reaction parameters by using the value range of the reaction parameters determined in the step S1 as a reference through a pull Ding Chao cubic experimental design. And extracting an initial sampling set within a preset reaction parameter range by using Latin hypercube sampling, wherein the sampling process stops sampling after a preset termination condition is reached. And (3) carrying out layered sampling, so that the structure of the sample is similar to the overall structure, the estimation precision is improved, and the sampling is terminated after the sampling process reaches the termination condition.
Further, the step S3 specifically includes:
and (2) according to the sampling result in the step (S2), obtaining a corresponding experimental result by utilizing an objective function or an actual experiment, and summarizing the reaction parameters and the corresponding experimental result to obtain an initial experimental data set. The objective function used is the kinetic equation of the flow chemistry. And obtaining an experimental result corresponding to the initial reaction parameter by using an actual experiment or an objective function. The experimental results are the experimental selectivity and conversion rate.
Further, the step S4 specifically includes:
s41, performing format preprocessing on the initial experimental data set in the step S3.
S42, constructing a Gaussian model for the data in the step S3 by adopting a Gaussian function, and predicting an experimental result.
And (3) carrying out format pretreatment on the initial experimental data set in the step (S3), and then carrying out Gaussian modeling on the treated data by combining a Gaussian function to predict an experimental result.
Further, the formula of the gaussian function in step S4 is expressed as:
wherein f (x) represents the predicted result of the gaussian function, d represents the distance between two points, a represents the standard deviation, Γ is the gamma function, K v Is a modified Bessel function and v is a non-negative parameter.
Further, the linear combination of random variables in the gaussian process in step S4 is subjected to normal distribution, each finite dimensional distribution is a joint normal distribution, and the probability density function of the finite dimensional distribution on the continuous index set is a gaussian measure of all random variables, which is a functional distribution on the continuous domain.
Further, the step S5 specifically includes:
s51, setting two optimization targets, and selecting maximization or minimization.
S52, setting a reference point slightly worse than the lower limit of the acceptable value of each optimization target.
S53, setting the iteration times and the number of candidate points.
S54, recommending new sampling points by the acquisition function according to the prediction result of the Gaussian function in the step S4, wherein the number of the new sampling points is the number of the candidate points set in the step S53, and the value range of the sampling points is the reaction parameter range set in the step S1.
S55, judging whether the iteration times in the step S53 reach a preset value, if so, entering the step S56, otherwise, returning to the step S4.
And S56, summarizing all the recommended reaction parameters of the acquisition function and the corresponding experimental results, taking the reaction parameters corresponding to the optimal experimental results as optimized reaction parameters, wherein at least one group of optimal parameters exist, and outputting the optimal parameters in a form of a graph.
Further, the calculation formula of the collection function in step S5 is expressed as:
wherein 
Representing the recommended results of the acquisition function, HVI representing the hypervolume improvement, X representing the candidate point, P representing the pareto boundary, t representing the number of trials, N representing the number of samples, +.>
Representing the true objective value.
Further, in step S5, the acquisition function calculates the probability that the value of the unknown point is greater than the value of the known point after balanced development and exploration according to the result of gaussian process prediction, and then selects a new test point, which brings about the maximum expected improvement. After the iteration is terminated, at least one set of reaction parameters is close to the optimal experimental conditions within a preset range.
Compared with the prior art, the scheme has the following beneficial effects:
after Latin hypercube sampling is carried out according to a preset reaction parameter value range, an initial experimental data set is obtained by utilizing an objective function or an actual experiment, a prediction model is constructed by utilizing a Gaussian model, a next sampling point is recommended by an acquisition function according to the prediction model, and the parallel multi-noise target expected hypercube improved acquisition function is adopted, so that the prediction accuracy is high, the efficiency is high, and the robustness is higher.
Drawings
FIG. 1 is an exemplary chemical reaction equation for flow chemistry optimization based on multi-objective Bayesian optimization provided by embodiments of the present invention.
Fig. 2 is a schematic step diagram of a flow chemical process optimization method based on multi-objective bayesian optimization.
Detailed Description
The following detailed description of the embodiments of the present invention, given by way of example only, is presented to aid those skilled in the art in a more complete, accurate and thorough understanding of the present invention's inventive concepts and technical solutions, and is not limited to the scope of the embodiments, but is intended to be protected by all the inventions utilizing the present invention's concepts.
Fig. 2 is a flowchart of a flow chemical process optimization method based on multi-objective bayesian optimization, which specifically includes the following steps:
s1, determining the value range of experimental reaction parameters, wherein the optimized flow chemical reaction in the example is a continuous flow monoacylation reaction of m-phenylenediamine and benzoic anhydride, as shown in figure 1.
The reaction parameters in the invention are experimental reaction conditions and process parameters to be optimized, the reaction parameters at least comprise one reaction condition, four parameters are used in the example, the optimization is performed in four-dimensional continuous space, and the range of the values of the reaction parameters is shown in table 1.
Table 1 reaction parameter value ranges
S2, carrying out initial sampling according to the value range of the reaction parameters in the step S1, and carrying out sampling through Latin hypercube test design.
In this embodiment, we extract ten initial sample sets, and adopt hierarchical sampling, so that the structure of the samples is similar to that of the whole body, and estimation accuracy is improved.
And S3, carrying out actual experiments or calculating experimental results by utilizing an objective function according to the initial sampling in the step S2, and summarizing to obtain an initial experimental data set.
In this embodiment, the experimental result is selectivity and conversion rate, the dynamic equation is used as an objective function to calculate the experimental result, and if the dynamic equation is unknown, the actual experiment can be used to obtain the result.
S4, constructing a Gaussian model according to the initial experimental data set in the step S3, wherein the Gaussian model is used for predicting experimental results, and the calculation formula of the Gaussian function is expressed as follows:
where f (x) represents the predicted outcome of the gaussian function,d represents the distance between two points, A represents the standard deviation, Γ is the gamma function, K v Is a modified Bessel function and v is a non-negative parameter.
S5, searching optimal reaction parameters by utilizing an acquisition function according to the prediction result obtained in the step S4, calculating an experimental result by utilizing an objective function for the selected new sampling point, updating an initial data set, detecting whether the iteration number reaches a preset iteration, returning to the step S4 if the detection result is not, and otherwise, outputting the optimal reaction parameters. The expression of the acquisition function in this example:
wherein 
Representing the recommended results of the acquisition function, HVI representing the hypervolume improvement, X representing the candidate point, P representing the pareto boundary, t representing the number of trials, N representing the number of samples, +.>
Representing the true objective value.
S51, an optimization objective is set, which in this example is conversion and selectivity, and both maximization is selected.
S52, setting a reference point slightly worse than the lower limit of the acceptable value of each optimization target.
S53, setting iteration times and the number of candidate points, wherein in practice, the iteration is set for 10 times, and 1 new candidate point is recommended each time.
S54, updating the recommended reaction parameters of the acquisition function and the experimental result obtained through the objective function calculation to the set of initial experimental data.
S55, judging whether the iteration times meet the preset value, if so, entering a step S56, otherwise, returning to the step S4.
And S56, outputting the optimal experimental result and the corresponding reaction parameters in a form of a chart, namely, the optimized reaction parameter combination.
In practice, the acquisition function can be explored and developed in a balanced way, the sampling data and the iteration times are greatly reduced, the optimization efficiency is greatly improved, and the experimental results obtained after 10 iterations are shown in table 2.
TABLE 2 results of experiments after 10 iterations
The flow chemistry process optimization method based on the multi-objective Bayesian optimization has the following beneficial technical effects:
the Bayesian optimization and the mobile chemistry are combined, history data are fully utilized, the optimization efficiency is improved, the experiment cost is reduced, meanwhile Latin hypercube sampling is adopted, and therefore the sampling quality is improved.
The principles and embodiments of the present invention have been described with reference to specific examples, which are provided herein to facilitate understanding of the method and core ideas of the present invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.
The embodiments described herein are presented to aid the reader in understanding the principles of the invention and should be understood that the scope of the invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific modifications and combinations from the teachings of the present disclosure without departing from the spirit thereof, and such modifications and combinations remain within the scope of the present disclosure.
Claims (8)
1. A flow chemistry process optimizing method based on multi-objective Bayesian optimization is characterized by comprising the following steps:
s1, determining a value range of experimental reaction parameters;
s2, carrying out initial sampling according to the value range of the reaction parameters in the step S1;
s3, carrying out actual experiments or obtaining experimental results by utilizing an objective function according to the initial sampling in the step S2, and summarizing the sampled reaction parameters and the corresponding experimental results to obtain an initial experimental data set;
s4, constructing a Gaussian model according to the initial experimental data set in the step S3 and combining the Gaussian function, and predicting an experimental result by using the Gaussian model to obtain a predicted result;
s5, searching optimal reaction parameters by utilizing an acquisition function according to the prediction result obtained in the step S4, carrying out an actual experiment on the selected new sampling points or calculating by utilizing an objective function to obtain an experiment result, updating an initial experiment data set, detecting whether the iteration number reaches a preset requirement, returning to the step S4 if the detection result is positive, and otherwise outputting the optimal reaction parameters in a chart format.
2. The method of optimizing a flow chemistry process based on multi-objective bayesian optimization according to claim 1, wherein the reaction parameters include residence time, temperature, reactant concentration in step S1.
3. The flow chemistry process optimization method based on multi-objective bayesian optimization according to claim 1, wherein step S2 is specifically: and (3) taking the value range of the reaction parameters determined in the step (S1) as a benchmark, extracting initial reaction parameters through a pull Ding Chao cubic experimental design, and completing initial sampling.
4. The method for optimizing a flow chemical process based on multi-objective bayesian optimization according to claim 1, wherein in the step S3, the experimental result is the experimental selectivity and conversion rate.
5. The method for optimizing a flow chemistry process based on multi-objective bayesian optimization according to claim 1, wherein in step S3, the objective function is a kinetic equation of a flow chemistry reaction.
6. The flow chemistry process optimization method based on the multi-objective bayesian optimization according to claim 1, wherein in step S4, step S4 specifically includes:
s41, carrying out format pretreatment on the initial experimental data set in the step S3;
s42, constructing a Gaussian model for the preprocessed data by adopting a Gaussian function, and predicting an experimental result to obtain a prediction result.
7. The flow chemistry process optimization method based on the multi-objective bayesian optimization according to claim 1, wherein in step S4, a calculation formula of a gaussian function of a gaussian model is expressed as:
wherein f (x) represents the prediction result of the Gaussian function, d represents the distance between two points, A represents the standard deviation, Γ is the gamma function, K v Is a modified Bessel function and v is a non-negative parameter.
8. The flow chemistry optimization method based on the multi-objective bayesian optimization according to claim 1, wherein in step S5, the expression of the collection function is:
wherein ,
representing the recommended results of the acquisition function, HVI representing the hypervolume improvement, X representing the candidate point, P representing the pareto boundary, t representing the number of trials, N representing the number of samples, +.>
Representing the true objective value. />
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211535581.XA CN116110505A (en) | 2022-11-30 | 2022-11-30 | Flow chemistry process optimization method based on multi-objective Bayesian optimization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211535581.XA CN116110505A (en) | 2022-11-30 | 2022-11-30 | Flow chemistry process optimization method based on multi-objective Bayesian optimization |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116110505A true CN116110505A (en) | 2023-05-12 |
Family
ID=86255154
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211535581.XA Pending CN116110505A (en) | 2022-11-30 | 2022-11-30 | Flow chemistry process optimization method based on multi-objective Bayesian optimization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116110505A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116502566A (en) * | 2023-06-27 | 2023-07-28 | 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) | Multi-objective optimization method for performance of combustion chamber of gas turbine based on Bayesian optimization |
| CN116959629A (en) * | 2023-09-21 | 2023-10-27 | 烟台国工智能科技有限公司 | Multi-index optimization method and system for chemical experiment, storage medium and electronic equipment |
| CN117744894A (en) * | 2024-02-19 | 2024-03-22 | 中国科学院电工研究所 | Active learning agent optimization method of comprehensive energy system |
-
2022
- 2022-11-30 CN CN202211535581.XA patent/CN116110505A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116502566A (en) * | 2023-06-27 | 2023-07-28 | 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) | Multi-objective optimization method for performance of combustion chamber of gas turbine based on Bayesian optimization |
| CN116959629A (en) * | 2023-09-21 | 2023-10-27 | 烟台国工智能科技有限公司 | Multi-index optimization method and system for chemical experiment, storage medium and electronic equipment |
| CN116959629B (en) * | 2023-09-21 | 2023-12-29 | 烟台国工智能科技有限公司 | Multi-index optimization method and system for chemical experiment, storage medium and electronic equipment |
| CN117744894A (en) * | 2024-02-19 | 2024-03-22 | 中国科学院电工研究所 | Active learning agent optimization method of comprehensive energy system |
| CN117744894B (en) * | 2024-02-19 | 2024-05-28 | 中国科学院电工研究所 | Active learning agent optimization method of comprehensive energy system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116110505A (en) | Flow chemistry process optimization method based on multi-objective Bayesian optimization | |
| CN110320335B (en) | Multi-element robust soft measurement method for effluent quality index of sewage treatment | |
| CN113433086B (en) | Method for predicting water quality COD (chemical oxygen demand) by combining fuzzy neural network with spectrophotometry | |
| CN109033524A (en) | A kind of chemical process concentration variable On-line Estimation method based on robust mixed model | |
| CN116777066B (en) | Photovoltaic power probability prediction method and device based on foundation cloud image guided depth decomposition | |
| Zhang et al. | A data‐driven soft sensor based on weighted probabilistic slow feature analysis for nonlinear dynamic chemical processes | |
| CN116612814A (en) | Regression model-based batch detection method, device, equipment and medium for gene sample pollution | |
| CN116153404A (en) | Single-cell ATAC-seq data analysis method | |
| CN113065281B (en) | TE process time sequence prediction method based on transfer entropy and long-short-term memory network | |
| CN115561282A (en) | Detection method and system for heavy metal in underground water | |
| CN114970674A (en) | Time sequence data concept drift adaptation method based on relevance alignment | |
| Zhang et al. | Active learning strategy for online prediction of particle size distribution in cobalt oxalate synthesis process | |
| Liu et al. | Soft sensor modelling of acrolein conversion based on hidden Markov model of principle component analysis and fireworks algorithm | |
| CN114354666A (en) | Method for extracting and optimizing spectral characteristics of soil heavy metal based on wavelength frequency selection | |
| CN112116091A (en) | On-line forecasting method for rapidly forecasting band gap of organic-inorganic hybrid perovskite based on machine learning | |
| Lu et al. | Neural network based on windowed convolutional transformation to extract features in time domain and its application on soft sensing | |
| CN116223480B (en) | Method and system for detecting heavy metal content in coal chemical industry sludge | |
| CN116049719B (en) | Fault detection and estimation method, system and medium based on migration type generation model | |
| CN116451812B (en) | Wettability prediction method and device based on multi-granularity cascade forest and super-parameter optimization | |
| CN111948155B (en) | Fine full spectrum combined GS-SVR improved complex water body nitrate quantitative analysis method | |
| CN116793991B (en) | Glutamic acid concentration measurement method based on near infrared spectrum and mixing loss | |
| CN114036948B (en) | Named entity identification method based on uncertainty quantification | |
| CN117076943A (en) | Method and device for predicting alkane gas content | |
| CN117909656A (en) | Method, system and storage medium for predicting temperature of rectifying tower based on ResNet-LSTM neural network model | |
| Carrington et al. | Post-selection inference for quantifying uncertainty in changes in variance |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |