CN117976094A - Recipe detection index prediction method, system, medium and equipment - Google Patents
Recipe detection index prediction method, system, medium and equipment Download PDFInfo
- Publication number
- CN117976094A CN117976094A CN202410084764.7A CN202410084764A CN117976094A CN 117976094 A CN117976094 A CN 117976094A CN 202410084764 A CN202410084764 A CN 202410084764A CN 117976094 A CN117976094 A CN 117976094A
- Authority
- CN
- China
- Prior art keywords
- detection
- data set
- model
- formula
- index
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 343
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000007781 pre-processing Methods 0.000 claims abstract description 26
- 239000000463 material Substances 0.000 claims description 62
- 238000011156 evaluation Methods 0.000 claims description 56
- 238000003860 storage Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 7
- 238000007637 random forest analysis Methods 0.000 claims description 5
- 230000001502 supplementing effect Effects 0.000 claims description 4
- 238000013136 deep learning model Methods 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000012628 principal component regression Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 abstract description 5
- 235000013305 food Nutrition 0.000 abstract description 4
- 238000012545 processing Methods 0.000 description 12
- 230000008569 process Effects 0.000 description 10
- 239000000047 product Substances 0.000 description 10
- 238000004891 communication Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 238000013461 design Methods 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 238000009472 formulation Methods 0.000 description 5
- 230000014509 gene expression Effects 0.000 description 5
- 230000010354 integration Effects 0.000 description 5
- 239000000203 mixture Substances 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 235000008429 bread Nutrition 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 238000007711 solidification Methods 0.000 description 2
- 230000008023 solidification Effects 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 108010068370 Glutens Proteins 0.000 description 1
- 240000004808 Saccharomyces cerevisiae Species 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000013604 expression vector Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 235000013312 flour Nutrition 0.000 description 1
- 235000021312 gluten Nutrition 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 230000009469 supplementation Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A formula detection index prediction method, a system, a medium and equipment relate to the technical field of food chemical formulas. The method comprises the following steps: acquiring a historical formula data set of a target type and a corresponding standard detection data set, wherein the standard detection data set comprises detection indexes and detection results corresponding to the detection indexes; preprocessing a historical formula data set and a standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data; inputting the sample data into a large model integrated by a plurality of predictive sub-models for training to obtain target predictive sub-models corresponding to all detection indexes; and acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, calculating the new formula data set through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result. By implementing the technical scheme provided by the application, the efficiency of predicting the formula detection index can be improved.
Description
Technical Field
The application relates to the technical field of food chemical industry formulas, in particular to a formula detection index prediction method, a formula detection index prediction system, a formula detection index prediction medium and formula detection index prediction equipment.
Background
Along with the development of society, the formula design of various products is more and more complex, and the requirements on the quality of the products are also higher and higher. In order to ensure the quality of the product, various key detection indexes of the product need to be predicted in the product formula design stage. Especially in the formulation process of food chemical formulas, the formulation and detection index prediction are key challenges for ensuring the quality and the production efficiency of products.
In the field of food chemical industry, the existing formula classification method mainly depends on rules and traditional statistical techniques, the prediction of the formula detection index is mainly obtained through physical and chemical detection in a laboratory, the traditional method for predicting the detection index involves complicated manual experiments, and the problem of low efficiency exists when large-scale and multi-dimensional formula data are processed, so that the formula detection index prediction efficiency is low.
Disclosure of Invention
The application provides a method, a system, a medium and equipment for predicting a formula detection index, which can improve the efficiency of predicting the formula detection index.
In a first aspect, the present application provides a method for predicting a recipe detection indicator, the method comprising:
Acquiring a historical formula data set of a target type and a corresponding standard detection data set, wherein the historical formula data set comprises historical formula materials and corresponding historical material consumption, and the standard detection data set comprises detection indexes and detection results corresponding to the detection indexes;
preprocessing the historical formula data set and the standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data;
Inputting the sample data into a large model integrated by a plurality of predictive sub-models for training to obtain target predictive sub-models corresponding to the detection indexes;
And acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, calculating the new formula data set through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result.
By adopting the technical scheme, the target prediction sub-model corresponding to each detection index is obtained by acquiring the historical formula data set of the target type and the standard detection data set corresponding to the historical formula data set, preprocessing the data set, inputting the preprocessed sample data into the large model integrated by the plurality of prediction sub-models for training. And then acquiring a new formula data set consistent with the target type, inputting the new formula data set into a trained large model, calculating the new formula data set through each target prediction sub-model to output each target detection index and a target detection result corresponding to the target detection index, and training a plurality of prediction sub-models and integrating the multiple prediction sub-models into one large model, so that the advantages of each prediction sub-model can be fully combined, the adaptability and the prediction precision of a complex formula are improved, and compared with a traditional complicated manual experiment mode, each key detection index of the new formula is rapidly predicted with high precision.
Optionally, the preprocessing the historical recipe data set and the standard detection data set, taking the preprocessed recipe data set and the preprocessed standard detection data set as sample data, includes: coding the historical formula materials and the corresponding historical material consumption in the historical formula data set to obtain a formula data set code; encoding each detection index in the standard detection data set and a detection result corresponding to each detection index to obtain a detection data set code; performing data preprocessing on the formula data set codes and the detection data set codes to obtain preprocessed formula data sets and standard detection data sets; and taking the preprocessed formula data set and standard detection data set as sample data.
By adopting the technical scheme, the text information can be digitized by carrying out coding processing on the original data set, so that the model is convenient for carrying out feature extraction and model training. The encoded sample data is preprocessed, invalid and redundant information can be filtered, data noise is reduced, and data quality is improved. The preprocessed data set is used as sample data of model training, so that the model training effect can be enhanced, and the model prediction performance can be improved.
Optionally, the data preprocessing is performed on the recipe data set code and the test data set code to obtain a preprocessed recipe data set and a standard test data set, which includes: judging whether missing values exist in the formula dataset codes and the detection dataset codes or not; and if the missing value exists in the formula data set code and/or the missing value exists in the detection data set code, supplementing each missing value with 0 to obtain a preprocessed formula data set and a standard detection data set.
By adopting the technical scheme, the misprediction caused by the missing of the sample data can be reduced through the processing of missing value supplementation, and the data volume available for model training can be reduced by directly discarding the sample with the missing value, so that the model performance is not facilitated. The missing value supplement can keep more samples, provide richer data for model training, avoid data waste, be favorable to improving the prediction accuracy of the model, adopt 0 to carry out the missing value supplement, can avoid introducing too big error. The accurate formula detection index prediction model can be trained better through the formula data set and the detection data set which are subjected to the deficiency value supplementing pretreatment.
Optionally, the large model is integrated by a plurality of predictor models, and the predictor models comprise a multiple linear regression model, a ridge regression model, a principal component regression model, a partial least squares model, a nonlinear regression model, a random forest model, a neural network model and a deep learning model.
By adopting the technical scheme, the advantages of each model can be fully utilized through the integration of the multiple types of prediction models, the adaptability to complex formula data is improved, meanwhile, the limitation of a single model can be reduced through the integrated model, and the accuracy and the stability of predicting complex formula detection indexes can be improved through the integration of multiple types of prediction sub-models, so that more accurate and reliable formula detection index prediction results are obtained.
Optionally, the inputting the sample data into a large model integrated by multiple predictor models for training to obtain a target predictor model corresponding to each detection index includes: sequentially inputting the sample data into each predictor model in the large model to obtain each index value corresponding to each detection index; substituting each index value into a model evaluation formula to obtain an evaluation value corresponding to each detection index; and taking the predictor model corresponding to the highest evaluation value as a target predictor model corresponding to each detection index.
By adopting the technical scheme, the model with the best prediction effect on each detection index can be screened from a plurality of candidate models by calculating the evaluation value and automatically selecting the target predictor model. Compared with manual selection, the technical means based on the automatic selection of the evaluation value can reduce subjective influence, determine a truly optimal target prediction sub-model from a more comprehensive angle, and finally integrate the target sub-model with the maximized evaluation value to form a large model, so that the advantages of each sub-model can be fully utilized, the prediction precision of each detection index of a complex formula is improved, and a more accurate and reliable formula detection index prediction result is output.
Optionally, the substituting each index value into a model evaluation formula to obtain an evaluation value corresponding to each detection index includes: acquiring a first predictor model corresponding to a first index value of a current detection index; the first index value is brought into a model evaluation formula to obtain an evaluation value of the first predictor model; the model evaluation formula is as follows: the mape is the evaluation value of the first predictor model, i is the i-th detection index in the standard detection set, y i is the detection result corresponding to the i-th detection index, and the/> An index value predicted by the first predictor model for the ith detection index; obtaining a second predictor model corresponding to a second index value of the current detection index, taking the second index value as the first index value, and executing the step of bringing the first index value into a model evaluation formula until the evaluation values corresponding to the current detection index in all predictor models are calculated; and taking the next detection index as the current detection index, and executing to acquire a first predictor model corresponding to the first index value of the current detection index until the evaluation values corresponding to all the detection indexes are obtained.
By adopting the technical scheme, the evaluation values of each detection index on each predictor model are calculated one by one through the model evaluation formula, the quantitative calculation of the evaluation values is carried out by adopting an accurate mathematical model, the evaluation formula considers the prediction error of each sample, the prediction accuracy degree can be comprehensively reflected, the evaluation values of all the submodels are calculated, the transverse comparison can be carried out, and the target predictor model with the best effect is selected. Compared with subjective evaluation, the technical means for quantitatively calculating the evaluation value can avoid artificial factors, evaluate the prediction effects of different models more accurately from the statistical perspective, and select the model which is truly optimal so as to improve the accuracy of the final prediction result.
Optionally, the data set of the new formulation includes a new material name and a corresponding new material usage, and the inputting the data set of the new formulation into the large model, so that the data set of the new formulation calculates through each target prediction sub-model, and outputs each target detection index and a corresponding target detection result, includes: inputting the new material name and the corresponding new material dosage into the large model, wherein the data set of the new formula is one of the historical formula data sets, and the new material dosage is in the range of the historical material dosage; acquiring the new formula data set, calculating through each target predictor model, and outputting each target detection index code and corresponding target detection result code; and decoding each target detection index code and the corresponding target detection result code to obtain each target detection index and the corresponding target detection result in the initial format.
By adopting the technical scheme, the new formula data set is input into the large model for prediction, so that the new formula can be rapidly and intelligently predicted, the new formula belongs to the range of the historical data set, the trained model can be directly applied for calculation without retraining, repeated labor is avoided, the prediction efficiency is improved, the numerical expression form of the data in the calculation process is maintained in the encoding and decoding process, additional errors are avoided, the human-readable prediction result of the initial format detection index is finally output, and the rapid and efficient prediction of the new formula detection index is realized.
In a second aspect of the present application, there is provided a recipe detection indicator prediction system, the system comprising:
The system comprises a historical data acquisition module, a target type detection module and a target type detection module, wherein the historical data acquisition module is used for acquiring a historical formula data set of a target type and a corresponding standard detection data set, the historical formula data set comprises historical formula materials and corresponding historical material consumption, and the standard detection data set comprises detection indexes and detection results corresponding to the detection indexes;
the data preprocessing module is used for preprocessing the historical formula data set and the standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data;
The sub-model selection module is used for inputting the sample data into a large model integrated by a plurality of prediction sub-models for training to obtain target prediction sub-models corresponding to the detection indexes;
The detection result determining module is used for acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, enabling the new formula data set to be calculated through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result.
In a third aspect the application provides a computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the above-described method steps.
In a fourth aspect of the application there is provided an electronic device comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the above-mentioned method steps.
In summary, one or more technical solutions provided in the embodiments of the present application at least have the following technical effects or advantages:
1. According to the method, the historical formula dataset of the target type and the standard detection dataset corresponding to the historical formula dataset are obtained, after preprocessing is carried out on the dataset, the preprocessed sample data are input into a large model integrated by a plurality of predictor models for training, and the target predictor model corresponding to each detection index is obtained. Then a new formula data set consistent with the target type is acquired and input into a trained large model, the new formula data set is calculated through each target prediction sub-model to output each target detection index and a corresponding target detection result, and the advantages of each prediction sub-model can be fully combined through training a plurality of prediction sub-models and integrating into one large model, so that the adaptability and the prediction precision to a complex formula are improved, and compared with a traditional complicated manual experiment mode, each key detection index of the new formula is rapidly predicted with high precision;
2. According to the application, the text information can be digitized by encoding the original data set, so that the model is convenient to extract the characteristics and train the model. The encoded sample data is preprocessed, invalid and redundant information can be filtered, data noise is reduced, and data quality is improved. The preprocessed data set is used as sample data of model training, so that the model training effect can be enhanced, and the model prediction performance can be improved;
3. the application can fully utilize the advantages of each model through the integration of multiple types of prediction models, improves the adaptability to complex formula data, simultaneously reduces the limitation of a single model through the integration model, and improves the accuracy and the stability of predicting complex formula detection indexes through the integration of multiple types of prediction sub-models, thereby obtaining more accurate and reliable formula detection index prediction results.
Drawings
FIG. 1 is a flow chart of a method for predicting a recipe detection index according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a system for predicting a recipe detection indicator according to an embodiment of the present application;
Fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Reference numerals illustrate: 300. an electronic device; 301. a processor; 302. a communication bus; 303. a user interface; 304. a network interface; 305. a memory.
Detailed Description
In order that those skilled in the art will better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments.
In describing embodiments of the present application, words such as "for example" or "for example" are used to mean serving as examples, illustrations, or descriptions. Any embodiment or design described herein as "such as" or "for example" in embodiments of the application should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "or" for example "is intended to present related concepts in a concrete fashion.
In the description of embodiments of the application, the term "plurality" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating an indicated technical feature. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless expressly specified otherwise.
The following description of the embodiments of the present application will be given in detail with reference to the accompanying drawings, and it is apparent that the embodiments described are only some, but not all embodiments of the present application.
Referring to fig. 1, a flow chart of a method for predicting a recipe detection index is specifically provided, the method may be implemented by a computer program, may be implemented by a single chip microcomputer, may also be run on a recipe detection index prediction system, the computer program may be integrated in an intelligent terminal, and may also be run as an independent tool application, and specifically the method includes steps 10 to 40, where the steps are as follows:
Step 10: and acquiring a historical formula data set of the target type and a corresponding standard detection data set, wherein the historical formula data set comprises all historical formula materials and corresponding historical material consumption, and the standard detection data set comprises all detection indexes and detection results corresponding to all detection indexes.
The embodiment of the application can be applied to the fields of design, production, quality control and the like of the formula type product, and can be used for realizing fast and efficient prediction of the key index of the formula, reducing trial and error cost and improving product design efficiency. For example, when designing a new product formula, the method can predict each detection index, thereby assisting research and development personnel in optimizing and determining the formula, reducing the times of trial production samples and shortening the research and development design period.
Specifically, a historical recipe data set of a target type and a corresponding standard test data set need to be acquired first. The historical formula data set of the target type in the embodiment of the application is consistent with the formula type to be predicted, and the historical formula data set of the target type is the formula data set of the same type, and the historical formula data set comprises all historical formula materials and corresponding historical material consumption. The standard detection data set refers to all key quality detection indexes which need to be predicted, and the standard detection data set comprises all detection indexes and corresponding detection results. The two establish a corresponding relation through the same batch of formulas, and each group of detection results in the standard detection data set corresponds to one formula record in the historical formula data set.
Illustratively, the historical recipe data set for the target type may be: x i＝[Mi1,Mi2,…,Mim ], wherein X i represents the feature vector name of the ith historical formula, and M ij represents the amount of the jth material in the representation vector of the ith historical formula. For example, for materials contained in the first historical recipe material: m 11、M12、M13、M14、M15 and M 16, wherein the amount of material M 11 is 0.3, the amount of material M 12 is 1, the amount of material M 11 is 12.7 … and the amount of material M 16 is 0.2. For a standard detection dataset may be: y i＝[Qi1,Qi2,…,Qim],Yi represents the feature vector name of the ith recipe, and Q ij represents the detection result of the jth detection index in the expression vector of the ith historical recipe material. For example, for a first historical recipe, the detection indicators include: q 11、Q12、Q13、Q14、Q15 and Q 16. The detection result of the detection index Q 11 is 52D, the detection result of the detection index Q 12 is 4 minutes 30 seconds …, and the detection result of the detection index Q 16 is coarse.
Step 20: preprocessing the historical formula data set and the standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data.
Specifically, various material names in the historical formula data set are encoded according to a certain rule, and the material consumption is also encoded in a numerical mode. Meanwhile, all detection indexes such as hardness and the like in the standard detection data set are also encoded, and the purpose of encoding is to convert word information into digital information which can be understood by a model so as to facilitate model training. On the encoded data set, data cleansing work such as processing missing values, outliers, deleting duplicates, etc. is performed. In addition, formatting processing, such as unified unit, can be performed, and the purpose of data cleaning is to improve the data quality and provide high-quality sample data. And finally, taking the encoded and preprocessed formula data set and the detection data set as training sample data of the model. The method can improve model training efficiency by using the coding expression form, and improve data quality by preprocessing, thereby helping model learning and ensuring more accurate prediction of the detection index.
On the basis of the above embodiment, as an optional embodiment, the step of preprocessing the historical recipe data set and the standard test data set, and taking the preprocessed recipe data set and the preprocessed standard test data set as sample data may further include the following steps:
Step 201: and encoding the historical formula materials in the historical formula data set and the corresponding historical material consumption to obtain a formula data set code.
Step 202: and encoding each detection index in the standard detection data set and the detection result corresponding to each detection index to obtain the detection data set encoding.
Step 203: and carrying out data preprocessing on the formula data set codes and the detection data set codes to obtain a preprocessed formula data set and a preprocessed standard detection data set.
Step 204: the preprocessed formula data set and the standard detection data set are used as sample data.
Specifically, the historical formula materials in the historical formula data set and the corresponding consumption of the historical materials are encoded, for example, various materials such as 'high gluten flour', 'yeast' and the like in the bread type historical formula data set are encoded, and the corresponding consumption of each material is encoded, so that a formula data set code only comprising coding expressions is obtained. Encoding each detection index in the standard detection data set, for example, encoding each detection index in the standard detection data set of bread, such as "hardness", "solidification", and the like, and encoding the detection result of each detection index to obtain a detection data set code. For example, the detection index is 52D, that is, the data of 52D is encoded to obtain corresponding encoded data 52; the detection index is that the detection result of solidification is 4 minutes 30 seconds, namely, the time-type data such as 4 minutes 30 seconds is encoded, and the corresponding encoded data is 270.
After the coding is finished, firstly judging whether missing values exist in the coding of the formula data set and the coding of the detection data set, namely whether the coding expression is complete or not, and whether all detection indexes appear or not. This is because the direct use of sample data with missing values can cause computational difficulties in the model training process, and if missing values are detected in the recipe data set code and/or missing values are detected in the recipe data set code, such as the recipe sample code lacking some code material or the detection sample lacking a detection value for some index, processing is required. In the embodiment of the application, a strategy of filling with 0 value is adopted, 0 is supplemented at the missing position so as to ensure that the sample expression is complete, a preprocessed formula data set and a standard detection data set are obtained, and the preprocessed formula data set and the standard detection data set are used as sample data. After the missing values are detected and filled, a formula and a detection data set with higher quality can be obtained, so that all sample characteristics and detection indexes are expressed completely, each sample can be ensured to participate in calculation stably in the model training process, the missing values are not stopped, and the convergence of the model is improved.
Step 30: and inputting the sample data into a large model integrated by a plurality of predictive sub-models for training to obtain a target predictive sub-model corresponding to each detection index.
Specifically, a large model is built that is integrated from multiple predictor models. Different types of machine learning algorithms can be used for different predictor models, and the predictor models in the embodiment of the application comprise, but are not limited to, a multiple linear regression model, a ridge regression model, a principal component regression model, a partial least square model, a nonlinear regression model, a random forest model, a neural network model, a deep learning model and the like. And then taking the preprocessed formula data set and the preprocessed detection data set as samples, inputting the samples into an integrated model for training, and in the training process, using formula information as input by each prediction sub-model, independently modeling and predicting each detection index, repeating multiple rounds of training, and optimizing the prediction effect of each sub-model. After model training, each material in the formula data set and each detection index in the detection data set establish a prediction association, and finally the target prediction sub-model with highest accuracy corresponding to each detection index can be obtained.
On the basis of the above embodiment, as an optional embodiment, the step of inputting the sample data into a large model integrated by a plurality of predictor models to train and obtain the target predictor model corresponding to each detection index may further include the following steps:
step 301: and sequentially inputting the sample data into each predictor model in the large model to obtain each index value corresponding to each detection index.
Step 302: substituting each index value into a model evaluation formula to obtain an evaluation value corresponding to each detection index.
Step 303: and taking the predictor model corresponding to the highest evaluation value as the target predictor model corresponding to each detection index.
Specifically, sample data is sequentially input into each prediction sub-model in the large model, the sub-model correspondingly outputs index values corresponding to each detection index, wherein the index values are coded values output by the sub-model, that is, the current formula corresponds to a plurality of detection indexes, and each detection index corresponds to an index value in the plurality of sub-models. Acquiring a first predictor model corresponding to a first index value of the current detection index, and bringing the first index value into a model evaluation formula to obtain an evaluation value of the first predictor model; the model evaluation formula is as follows: Wherein mape is the evaluation value of the first predictor model, i is the i-th detection index in the standard detection set, and yi is the detection result corresponding to the i-th detection index,/> An index value predicted in the first predictor model for the ith detection index; obtaining a second predictor model corresponding to a second index value of the current detection index, taking the second index value as a first index value, and executing the step of bringing the first index value into a model evaluation formula until the corresponding evaluation values of the current detection index in all predictor models are calculated; and taking the next detection index as the current detection index, and executing to acquire a first predictor model corresponding to the first index value of the current detection index until the evaluation values corresponding to all the detection indexes are obtained. For the same detection index, selecting the submodel with the highest evaluation value, namely the optimal prediction effect, as the final target prediction model of the detection index. For example, the highest evaluation value of the detection index Q1 is 98.3%, and the corresponding predictor model is a random forest model, and the random forest model is used as the target predictor model of the detection index Q1.
Step 40: and acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, calculating the new formula data set through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result.
Specifically, training out a target predictor model corresponding to each detection index through a historical formula data set and a corresponding standard detection data set. In practical application, if the detection index corresponding to the new formula with the same target type needs to be predicted, the new formula data needs to be input into a large model by adopting codes and formats consistent with the training data, so that the preprocessed new formula data can be directly adapted to the model, and a new formula sample can be subjected to reasoning calculation through the target predictor model corresponding to each detection index, so that each target detection index and a corresponding target detection result are obtained.
On the basis of the above embodiment, as an optional embodiment, the step of acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, so that the new formula data set is calculated by each target prediction sub-model, and outputting each target detection index and the corresponding target detection result may further include the following steps:
Step 401: and inputting the new material name and the corresponding new material consumption into the large model, wherein the data set of the new formula is one of the historical formula data sets, and the new material consumption is in the range of the historical material consumption.
Step 402: and acquiring a new formula data set, calculating through each target predictor model, and outputting each target detection index code and a corresponding target detection result code.
Step 403: and decoding each target detection index code and the corresponding target detection result code to obtain each target detection index and the corresponding target detection result in the initial format.
Specifically, when the detection index corresponding to the new formula needs to be predicted, the new material name and the corresponding new material usage in the new formula data set are input into the large model, and it is required that the new formula data set is one of the historical formula data sets, and the new material usage is in the range of the historical material usage. And the same coding rule is used for coding a new formula data set and preprocessing data, the preprocessed new formula data set is input into a large model, the new formula data set is subjected to reasoning calculation through a target prediction sub-model corresponding to each detection index, each prediction target prediction sub-model outputs a corresponding target detection index code and a corresponding target detection result code, and then each target detection index code and the corresponding target detection result code are decoded to obtain each target detection index and a corresponding target detection result in an initial format. For example, the target detection result is encoded into 270 data and decoded, and the obtained detection result in the initial format is 4 minutes and 30 seconds, so that the readable decoding prediction result of each detection index of the new formula on the pre-training model can be obtained.
Referring to fig. 2, a schematic diagram of a formula detection index prediction system according to an embodiment of the application is provided, where the formula detection index prediction system may include: the device comprises a historical data acquisition module, a data preprocessing module, a submodel selection module and a detection result determination module, wherein:
The system comprises a historical data acquisition module, a target type detection module and a target type detection module, wherein the historical data acquisition module is used for acquiring a historical formula data set of a target type and a corresponding standard detection data set, the historical formula data set comprises historical formula materials and corresponding historical material consumption, and the standard detection data set comprises detection indexes and detection results corresponding to the detection indexes;
the data preprocessing module is used for preprocessing the historical formula data set and the standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data;
The sub-model selection module is used for inputting the sample data into a large model integrated by a plurality of prediction sub-models for training to obtain target prediction sub-models corresponding to the detection indexes;
The detection result determining module is used for acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, enabling the new formula data set to be calculated through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result.
Optionally, the data preprocessing module is further configured to encode the historical formula materials in the historical formula data set and the corresponding historical material usage amounts to obtain a formula data set code; encoding each detection index in the standard detection data set and a detection result corresponding to each detection index to obtain a detection data set code; performing data preprocessing on the formula data set codes and the detection data set codes to obtain preprocessed formula data sets and standard detection data sets; and taking the preprocessed formula data set and standard detection data set as sample data.
Optionally, the data preprocessing module is further configured to determine whether missing values exist in the formula dataset encoding and the detection dataset encoding; and if the missing value exists in the formula data set code and/or the missing value exists in the detection data set code, supplementing each missing value with 0 to obtain a preprocessed formula data set and a standard detection data set.
Optionally, the sub-model selection module is further configured to sequentially input the sample data into each of the predictor sub-models in the large model, so as to obtain each index value corresponding to each detection index; substituting each index value into a model evaluation formula to obtain an evaluation value corresponding to each detection index; and taking the predictor model corresponding to the highest evaluation value as a target predictor model corresponding to each detection index.
Optionally, the sub-model selection module is further configured to obtain a first prediction sub-model corresponding to a first index value of the current detection index; the first index value is brought into a model evaluation formula to obtain an evaluation value of the first predictor model; the model evaluation formula is as follows: the mape is the evaluation value of the first predictor model, i is the i-th detection index in the standard detection set, y i is the detection result corresponding to the i-th detection index, and the/> An index value predicted by the first predictor model for the ith detection index; obtaining a second predictor model corresponding to a second index value of the current detection index, taking the second index value as the first index value, and executing the step of bringing the first index value into a model evaluation formula until the evaluation values corresponding to the current detection index in all predictor models are calculated; and taking the next detection index as the current detection index, and executing to acquire a first predictor model corresponding to the first index value of the current detection index until the evaluation values corresponding to all the detection indexes are obtained.
Optionally, the detection result determining module is further configured to input the new material name and a corresponding new material usage amount into the large model, where the data set of the new formula is one of the historical formula data sets, and the new material usage amount is in a range of the historical material usage amount; acquiring the new formula data set, calculating through each target predictor model, and outputting each target detection index code and corresponding target detection result code; and decoding each target detection index code and the corresponding target detection result code to obtain each target detection index and the corresponding target detection result in the initial format.
It should be noted that: in the system provided in the above embodiment, when implementing the functions thereof, only the division of the above functional modules is used as an example, in practical application, the above functional allocation may be implemented by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to implement all or part of the functions described above. In addition, the system and method embodiments provided in the foregoing embodiments belong to the same concept, and specific implementation processes of the system and method embodiments are detailed in the method embodiments, which are not repeated herein.
The embodiment of the present application further provides a computer storage medium, where the computer storage medium may store a plurality of instructions, where the instructions are suitable for being loaded by a processor and executed by a processor, and a specific execution process may refer to a specific description of the foregoing embodiment, and will not be described herein.
Referring to fig. 3, the application also discloses an electronic device. Fig. 3 is a schematic structural diagram of an electronic device according to the disclosure. The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, at least one communication bus 302.
Wherein the communication bus 302 is used to enable connected communication between these components.
The user interface 303 may include a Display screen (Display), a Camera (Camera), and the optional user interface 303 may further include a standard wired interface, and a wireless interface.
The network interface 304 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface), among others.
Wherein the processor 301 may include one or more processing cores. The processor 301 utilizes various interfaces and lines to connect various portions of the overall server, perform various functions of the server and process data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and invoking data stored in the memory 305. Alternatively, the processor 301 may be implemented in at least one hardware form of digital signal Processing (DIGITAL SIGNAL Processing, DSP), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 301 may integrate one or a combination of several of a central processing unit (Central Processing Unit, CPU), an image processor (Graphics Processing Unit, GPU), and a modem, etc. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 301 and may be implemented by a single chip.
The Memory 305 may include a random access Memory (Random Access Memory, RAM) or a Read-Only Memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer readable medium (non-transitory computer-readable storage medium). Memory 305 may be used to store instructions, programs, code, sets of codes, or sets of instructions. The memory 305 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-described respective method embodiments, etc.; the storage data area may store data or the like involved in the above respective method embodiments. Memory 305 may also optionally be at least one storage device located remotely from the aforementioned processor 301. Referring to fig. 3, an operating system, a network communication module, a user interface module, and an application program of a recipe detection index prediction method may be included in the memory 305 as a computer storage medium.
In the electronic device 300 shown in fig. 3, the user interface 303 is mainly used for providing an input interface for a user, and acquiring data input by the user; and the processor 301 may be configured to invoke an application program in the memory 305 that stores a recipe detection indicator prediction method that, when executed by the one or more processors 301, causes the electronic device 300 to perform the method as described in one or more of the embodiments above. It should be noted that, for simplicity of description, the foregoing method embodiments are all described as a series of acts, but it should be understood by those skilled in the art that the present application is not limited by the order of acts described, as some steps may be performed in other orders or concurrently in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all of the preferred embodiments, and that the acts and modules referred to are not necessarily required for the present application.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.
In the several embodiments provided by the present application, it should be understood that the disclosed apparatus may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, such as a division of units, merely a division of logic functions, and there may be additional divisions in actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some service interface, device or unit indirect coupling or communication connection, electrical or otherwise.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable memory. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in whole or in part in the form of a software product stored in a memory, comprising several instructions for causing a computer device (which may be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method of the various embodiments of the present application. And the aforementioned memory includes: various media capable of storing program codes, such as a U disk, a mobile hard disk, a magnetic disk or an optical disk.
The foregoing is merely exemplary embodiments of the present disclosure and is not intended to limit the scope of the present disclosure. That is, equivalent changes and modifications are contemplated by the teachings of this disclosure, which fall within the scope of the present disclosure. Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure.
This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a scope and spirit of the disclosure being indicated by the claims.
Claims (10)
1. A method for predicting a recipe detection indicator, the method comprising:
Acquiring a historical formula data set of a target type and a corresponding standard detection data set, wherein the historical formula data set comprises historical formula materials and corresponding historical material consumption, and the standard detection data set comprises detection indexes and detection results corresponding to the detection indexes;
preprocessing the historical formula data set and the standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data;
Inputting the sample data into a large model integrated by a plurality of predictive sub-models for training to obtain target predictive sub-models corresponding to the detection indexes;
And acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, calculating the new formula data set through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result.
2. The recipe detection index prediction method according to claim 1, wherein the preprocessing the historical recipe data set and the standard detection data set, taking the preprocessed recipe data set and standard detection data set as sample data, comprises:
coding the historical formula materials and the corresponding historical material consumption in the historical formula data set to obtain a formula data set code;
encoding each detection index in the standard detection data set and a detection result corresponding to each detection index to obtain a detection data set code;
performing data preprocessing on the formula data set codes and the detection data set codes to obtain preprocessed formula data sets and standard detection data sets;
and taking the preprocessed formula data set and standard detection data set as sample data.
3. The method of claim 2, wherein the data preprocessing the recipe dataset encoding and the test dataset encoding to obtain a preprocessed recipe dataset and a standard test dataset, comprises:
judging whether missing values exist in the formula dataset codes and the detection dataset codes or not;
And if the missing value exists in the formula data set code and/or the missing value exists in the detection data set code, supplementing each missing value with 0 to obtain a preprocessed formula data set and a standard detection data set.
4. The method of claim 1, wherein the large model is integrated with a plurality of predictor models, the predictor models including a multiple linear regression model, a ridge regression model, a principal component regression model, a partial least squares model, a nonlinear regression model, a random forest model, a neural network model, and a deep learning model.
5. The method for predicting a recipe detection index according to claim 1, wherein the step of inputting the sample data into a large model integrated by a plurality of predictor models for training to obtain target predictor models corresponding to the detection indexes comprises:
sequentially inputting the sample data into each predictor model in the large model to obtain each index value corresponding to each detection index;
Substituting each index value into a model evaluation formula to obtain an evaluation value corresponding to each detection index;
and taking the predictor model corresponding to the highest evaluation value as a target predictor model corresponding to each detection index.
6. The method according to claim 5, wherein substituting each index value into a model evaluation formula to obtain an evaluation value corresponding to each detection index comprises:
Acquiring a first predictor model corresponding to a first index value of a current detection index;
the first index value is brought into a model evaluation formula to obtain an evaluation value of the first predictor model;
The model evaluation formula is as follows: the mape is the evaluation value of the first predictor model, i is the i-th detection index in the standard detection set, y i is the detection result corresponding to the i-th detection index, and the/> An index value predicted by the first predictor model for the ith detection index;
obtaining a second predictor model corresponding to a second index value of the current detection index, taking the second index value as the first index value, and executing the step of bringing the first index value into a model evaluation formula until the evaluation values corresponding to the current detection index in all predictor models are calculated;
And taking the next detection index as the current detection index, and executing to acquire a first predictor model corresponding to the first index value of the current detection index until the evaluation values corresponding to all the detection indexes are obtained.
7. The method according to claim 1, wherein the new recipe data set includes new material names and corresponding new material amounts, the inputting the new recipe data set into the large model, so that the new recipe data set is calculated by each target prediction sub-model, and each target detection index and corresponding target detection result are output, including:
inputting the new material name and the corresponding new material dosage into the large model, wherein the data set of the new formula is one of the historical formula data sets, and the new material dosage is in the range of the historical material dosage;
acquiring the new formula data set, calculating through each target predictor model, and outputting each target detection index code and corresponding target detection result code;
and decoding each target detection index code and the corresponding target detection result code to obtain each target detection index and the corresponding target detection result in the initial format.
8. A system for predicting a recipe detection indicator, the system comprising:
The system comprises a historical data acquisition module, a target type detection module and a target type detection module, wherein the historical data acquisition module is used for acquiring a historical formula data set of a target type and a corresponding standard detection data set, the historical formula data set comprises historical formula materials and corresponding historical material consumption, and the standard detection data set comprises detection indexes and detection results corresponding to the detection indexes;
the data preprocessing module is used for preprocessing the historical formula data set and the standard detection data set, and taking the preprocessed formula data set and the preprocessed standard detection data set as sample data;
The sub-model selection module is used for inputting the sample data into a large model integrated by a plurality of prediction sub-models for training to obtain target prediction sub-models corresponding to the detection indexes;
The detection result determining module is used for acquiring a new formula data set consistent with the target type, inputting the new formula data set into the large model, enabling the new formula data set to be calculated through each target prediction sub-model, and outputting each target detection index and a corresponding target detection result.
9. A computer readable storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method of any one of claims 1 to 7.
10. An electronic device comprising a processor, a memory, a user interface, and a network interface, the memory for storing instructions, the user interface and the network interface for communicating to other devices, the processor for executing the instructions stored in the memory to cause the electronic device to perform the method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410084764.7A CN117976094A (en) | 2024-01-19 | 2024-01-19 | Recipe detection index prediction method, system, medium and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410084764.7A CN117976094A (en) | 2024-01-19 | 2024-01-19 | Recipe detection index prediction method, system, medium and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117976094A true CN117976094A (en) | 2024-05-03 |
Family
ID=90847117
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410084764.7A Pending CN117976094A (en) | 2024-01-19 | 2024-01-19 | Recipe detection index prediction method, system, medium and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117976094A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021004198A1 (en) * | 2019-07-10 | 2021-01-14 | 江苏金恒信息科技股份有限公司 | Plate performance prediction method and apparatus |
| CN115238590A (en) * | 2022-08-11 | 2022-10-25 | 中科云谷科技有限公司 | Welding quality prediction method, device and computer storage medium |
| CN115829157A (en) * | 2022-12-28 | 2023-03-21 | 淮阴工学院 | Chemical water quality index prediction method based on variational modal decomposition and auto former model |
| CN116776753A (en) * | 2023-08-25 | 2023-09-19 | 北京科技大学 | Soft measurement method and system for mechanical property index in hot continuous rolling process of strip steel |
-
2024
- 2024-01-19 CN CN202410084764.7A patent/CN117976094A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021004198A1 (en) * | 2019-07-10 | 2021-01-14 | 江苏金恒信息科技股份有限公司 | Plate performance prediction method and apparatus |
| CN115238590A (en) * | 2022-08-11 | 2022-10-25 | 中科云谷科技有限公司 | Welding quality prediction method, device and computer storage medium |
| CN115829157A (en) * | 2022-12-28 | 2023-03-21 | 淮阴工学院 | Chemical water quality index prediction method based on variational modal decomposition and auto former model |
| CN116776753A (en) * | 2023-08-25 | 2023-09-19 | 北京科技大学 | Soft measurement method and system for mechanical property index in hot continuous rolling process of strip steel |
Non-Patent Citations (1)
| Title |
|---|
| 宋剑杰;: "污水处理过程COD指标集成软测量模型", 计算机工程与应用, no. 17, 25 July 2011 (2011-07-25), pages 1 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Galeotti et al. | On the robustness of robustness checks of the environmental Kuznets curve hypothesis | |
| CN112668822B (en) | Scientific and technological achievement transformation platform sharing system, method, storage medium and mobile phone APP | |
| CN113268883A (en) | Method for predicting corrosion rate of submarine crude oil pipeline based on PCA-ABC-SVM model | |
| CN115146607B (en) | Comment information emotion preference recognition model training method, recognition method and device | |
| CN115034430A (en) | Carbon emission prediction method, device, terminal and storage medium | |
| CN115436591A (en) | Method, device, medium and system for detecting chemical oxygen demand of wastewater | |
| CN114239949A (en) | Website access amount prediction method and system based on two-stage attention mechanism | |
| CN113627977A (en) | House value prediction method based on heteromorphic graph | |
| CN117976094A (en) | Recipe detection index prediction method, system, medium and equipment | |
| CN117150237A (en) | Time sequence data prediction method, device, equipment and computer readable storage medium | |
| CN112257958A (en) | Power saturation load prediction method and device | |
| CN114240519A (en) | Comprehensive evaluation-based intelligent brand promotion sequencing method and system | |
| CN113111588B (en) | NO of gas turbine X Emission concentration prediction method and device | |
| CN114840983A (en) | Power storage battery SOH calculation model construction method and calculation method | |
| CN114971354A (en) | Water quality grade evaluation method, device and medium | |
| CN115115469A (en) | Method and system for evaluating scheduling efficiency of three-dimensional library, storage medium and terminal | |
| CN113779391A (en) | Intelligent lock unlocking recommendation method, system and device based on modeling and storage medium | |
| CN113706347A (en) | Multitask model distillation method, multitask model distillation system, multitask model distillation medium and electronic terminal | |
| CN113128907A (en) | Patent value online evaluation method and system | |
| CN112100367A (en) | Public opinion early warning method and device for scenic spot | |
| CN114386196B (en) | Method for evaluating mechanical property prediction accuracy of plate strip | |
| CN114519307A (en) | Information prediction method, device, system, storage medium and electronic equipment | |
| CN113706212A (en) | Prediction model-based housing valuation method, device, equipment and storage medium | |
| CN117612603A (en) | Protein mutation effect prediction method, device, equipment and medium | |
| CN117035852A (en) | Medicine market data analysis method and system based on sample data |
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 |