CN111768096A - Rating method and device based on algorithm model, electronic equipment and storage medium - Google Patents
Rating method and device based on algorithm model, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN111768096A CN111768096A CN202010593982.5A CN202010593982A CN111768096A CN 111768096 A CN111768096 A CN 111768096A CN 202010593982 A CN202010593982 A CN 202010593982A CN 111768096 A CN111768096 A CN 111768096A
- Authority
- CN
- China
- Prior art keywords
- model
- user data
- data set
- models
- sub
- 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
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 115
- 238000000034 method Methods 0.000 title claims abstract description 52
- 229920000642 polymer Polymers 0.000 claims abstract description 60
- 238000012549 training Methods 0.000 claims abstract description 46
- 238000011156 evaluation Methods 0.000 claims abstract description 35
- 238000010801 machine learning Methods 0.000 claims description 29
- 238000012216 screening Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 7
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 9
- 238000011835 investigation Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/02—Banking, e.g. interest calculation or account maintenance
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Software Systems (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Operations Research (AREA)
- Artificial Intelligence (AREA)
- Tourism & Hospitality (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Technology Law (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an artificial intelligence technology, and discloses a rating method based on an algorithm model, which comprises the following steps: acquiring a user data set, constructing a plurality of original sub-models according to data characteristics, and training to obtain a training result; performing parameter tuning on the plurality of original sub-models to obtain a plurality of optimal parameter sub-models; selecting the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models; merging the polymer models into a main model according to a preset accurate merging algorithm; and acquiring a user data set to be judged, inputting the user data set to be judged to the main model, and acquiring a grade rating result of the user data set to be judged according to an output result of the main model and outputting the grade rating result to a user. The invention also relates to a blockchain technique, in which user data sets can be stored. The invention can use various algorithms to comprehensively evaluate the enterprise grade, improve the accuracy of the rating result and enhance the overall flexibility of the evaluation mode.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a rating method and device based on an algorithm model and a computer readable storage medium.
Background
At present, the evaluation of the enterprise grade of the small and micro enterprises mainly adopts a scoring card mode to calculate the scoring of the small and micro enterprises. And if the score meets a threshold value of a certain grade, the small micro-enterprise is considered to belong to the grade. Meanwhile, a corresponding rating model is constructed by adopting a machine learning algorithm, so that the output of the rating score of a single data source is realized. The above method mainly has the following defects: the accuracy of the result is low by using only a single algorithm, so the method does not achieve the desired effect under the condition of occupying too many computer resources and human resources.
Furthermore, the existing rating method also has the following problems: when the internal business of the small and micro enterprise is changed, the rating of the small and micro enterprise cannot be updated.
Disclosure of Invention
The invention provides a rating method and device based on an algorithm model and a computer readable storage medium, and mainly aims to solve the problems that the enterprise grade is comprehensively rated by using multiple algorithms, the accuracy of a rating result is improved, and the overall flexibility of a rating mode is enhanced.
In order to achieve the above object, the present invention provides a rating method based on an algorithm model, comprising:
acquiring a user data set, constructing a plurality of original submodels according to the data characteristics of the user data set, and training the original submodels to obtain a model training result;
performing parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models;
selecting the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models;
merging the polymer models into a main model according to a preset accurate merging algorithm;
and acquiring a user data set to be judged, inputting the user data set to be judged to the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user.
Optionally, the user data set is stored in a block chain, and the constructing a plurality of original sub-models according to the data characteristics of the user data set includes:
screening the user data set to obtain an effective user data set;
dividing the user data set into a plurality of sub-user data sets according to the data type of the data in the effective user data set;
and respectively selecting corresponding machine learning algorithms to construct a plurality of original sub-models according to the data characteristics of the sub-user data sets.
Optionally, the performing parameter tuning on the multiple original sub-models according to the model training result to obtain multiple optimal parameter sub-models includes:
taking the ratio of the training result to the original correct result as the judgment accuracy of the original sub-model;
comparing the judgment accuracy with a preset accuracy standard threshold;
if the judgment accuracy is lower than the standard threshold, adjusting the parameters of the plurality of original sub-models by using a preset parameter adjustment method;
and obtaining a plurality of optimal parameter submodels until the judgment accuracy is greater than or equal to the standard threshold.
Optionally, the selecting the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models includes:
obtaining the score of each optimal parameter submodel according to a preset evaluation algorithm;
and reserving the optimal parameter submodel with the score value larger than or equal to a preset score threshold value to obtain the polymer model.
Optionally, the obtaining the score of each optimal parameter submodel according to a preset evaluation algorithm includes:
calculating the score of each optimal parameter submodel by adopting the following algorithm:
Score＝avg(auc,ks,-log(psi))
wherein, Score is the Score of each optimal parameter submodel under the unified evaluation index, auc is the judgment accuracy of each original submodel, ks is the maximum adjustment value when parameter adjustment is executed when each original submodel is trained, and psi is the training frequency when each original submodel is trained.
Optionally, the merging the plurality of polymer models into a master model according to a preset precise merging algorithm includes:
calculating the weight of each polymer model according to the judgment accuracy of the plurality of polymer models;
merging the plurality of polymer models into a master model by using the following exact merging algorithm:
wherein P is the main model, wkIs the weight, p, of each of the polymer modelskIs the output result, AUC, of each of the polymer modelskIs the judgment accuracy, AUC, of the kth Polymer modeliIs the judgment accuracy of all the polymer models, and n is the total number of the polymer models.
Optionally, the method further comprises:
when a new user data set with different data characteristics from those in the user data set is obtained, selecting a corresponding original sub-model according to the new user data set, and performing parameter adjustment on the corresponding original sub-model to obtain an updated original sub-model;
and merging the updated original sub-model into the main model.
In order to solve the above problems, the present invention also provides an algorithmic model-based rating apparatus, comprising:
the submodel building module is used for obtaining a user data set, building a plurality of original submodels according to the data characteristics of the user data set, and training the plurality of original submodels to obtain a training result;
the parameter tuning module is used for performing parameter tuning on the plurality of original sub-models according to the training result to obtain a plurality of optimal parameter sub-models;
the submodel screening module is used for selecting the optimal parameter submodels by utilizing a preset evaluation algorithm to obtain a plurality of polymer models;
the master model building module is used for merging the plurality of polymer models into a master model according to a preset accurate merging algorithm;
and the grade analysis module is used for acquiring a user data set to be judged, inputting the user data set to be judged into the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user.
In order to solve the above problem, the present invention also provides an electronic device, including:
a memory storing at least one instruction; and
and the processor executes the instructions stored in the memory to realize the algorithm model-based rating method.
In order to solve the above problems, the present invention also provides a computer-readable storage medium comprising a storage data area storing data created according to the use of blockchain nodes and a storage program area storing a computer program, wherein the computer program, when executed by a processor, implements the above algorithm model-based rating method.
The method comprises the steps of obtaining a user data set, constructing a plurality of original submodels according to data characteristics of the user data set, training the original submodels to obtain model training results, and comprehensively evaluating enterprise grades by using a plurality of algorithms when the original submodels are constructed; according to the model training result, parameter tuning is carried out on the plurality of original sub-models, so that the accuracy of the rating result can be improved; selecting the optimal parameter submodels by using a preset evaluation algorithm, further improving the accuracy of the rating result and reducing the consumption of computing resources; the polymer models are combined into a main model according to a preset accurate combination algorithm, so that the overall flexibility of a rating mode is improved; and acquiring a user data set to be judged, inputting the user data set to be judged to the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user, so that the usability of the grade rating result is ensured. Therefore, the rating method, the rating device, the electronic equipment and the computer readable storage medium based on the algorithm model can realize the functions of comprehensively evaluating the enterprise grade by using various algorithms, improving the accuracy of the rating result and enhancing the overall flexibility of the evaluation mode.
Drawings
FIG. 1 is a flow chart of a rating method based on an algorithm model according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of constructing a plurality of original sub-models according to an embodiment of the present invention;
FIG. 3 is a block diagram of an algorithmic model based rating device according to an embodiment of the present invention;
fig. 4 is a schematic diagram of an internal structure of an electronic device implementing a rating method based on an algorithm model according to an embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The execution subject of the rating method based on the algorithm model provided by the embodiment of the present application includes, but is not limited to, at least one of electronic devices, such as a server and a terminal, which can be configured to execute the method provided by the embodiment of the present application. In other words, the rating method based on the algorithm model may be performed by software or hardware installed in the terminal device or the server device, and the software may be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like.
Referring to fig. 1, a flowchart of a rating method based on an algorithm model according to an embodiment of the present invention is shown. In the embodiment, the rating method based on the algorithm model comprises the following steps:
s1, obtaining a user data set, constructing a plurality of original sub-models according to the data characteristics of the user data set, and training the original sub-models to obtain model training results.
Preferably, in the embodiment of the present invention, the user data set may be a data set of financial attributes of an enterprise, including, for example, social security data, credit rating, cash flow balance, loan query record, account settlement information, and the like of the enterprise. The user data set may be stored in a node of a block chain.
In detail, as shown in fig. 2, the obtaining a user data set and constructing a plurality of original sub-models according to data characteristics of the user data set includes:
and S21, screening the user data set to obtain an effective user data set.
Preferably, the embodiment of the present invention obtains the user data set stored in the blockchain node based on a pre-joined intelligent contract. The smart contract is a computer protocol intended to propagate, verify or execute contracts in an informational manner.
Further, the embodiment of the invention screens the user data set according to a preset screening standard. The preset screening standard is to set different limiting conditions for different data types and remove invalid data, repeated data and the like in the user data set, so that resource consumption is reduced, data processing efficiency is improved, and accuracy of an analysis result is improved.
For example, the social security data of one enterprise is screened in the time range of the last three months, and the cash flow balance data is screened in the condition of the minimum amount of more than 100 yuan.
S22, dividing the user data set into a plurality of sub-user data sets according to the data type of the effective user data set.
For example, the valid data set of an enterprise includes social security data, credit rating, cash flow balance, loan inquiry record, account settlement information, etc., and the embodiment of the present invention may use the social security data of an enterprise as a data set of business and industry information, the credit rating and cash flow balance as a data set of credit investigation, and the loan inquiry record and account settlement information as a data set of credit investigation.
And S23, respectively selecting corresponding machine learning algorithms according to the data characteristics of the sub-user data sets to construct a plurality of original sub-models.
For example, the embodiment of the invention can use a first machine learning algorithm to construct a first sub-model for the data of the industry and commerce information data set, use a second machine learning algorithm to construct a second sub-model for the data of the credit investigation data set, and use a third machine learning algorithm to construct a third sub-model for the data of the credit investigation data set, so as to ensure that the data with different characteristics are fully utilized and improve the accuracy of the analysis result.
In detail, the first machine learning algorithm, the second machine learning algorithm, and the third machine learning algorithm may adopt currently disclosed machine learning algorithms, such as logistic regression, random forest, support vector machine, GBDT, naive bayes, neural network, XGboost, and the like, and the first machine learning algorithm, the second machine learning algorithm, and the third machine learning algorithm may all adopt the same machine learning algorithm, or may adopt different machine learning algorithms.
Preferably, the embodiment of the present invention inputs the plurality of sub-user data sets into the corresponding original sub-models respectively for training, so as to obtain a plurality of model training results corresponding to the original sub-models.
And S2, performing parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models.
In detail, the embodiment of the present invention compares the model training result with the original correct result, and uses the ratio of the training result to the original correct result as the judgment accuracy of the original sub-model.
Further, the embodiment of the present invention compares the judgment accuracy with a preset accuracy standard threshold, and if the judgment accuracy is lower than the standard threshold, the preset parameter adjustment method is used to adjust the parameters of the plurality of original submodels until the judgment accuracy is greater than or equal to the standard threshold, so as to obtain a plurality of optimal parameter submodels.
The embodiment of the invention adjusts the parameters of the plurality of original sub-models, can fully utilize the data characteristics of the sub-user data set and improve the accuracy of the original sub-models. The preset parameter adjusting method can adopt a random parameter adjusting method and a grid searching method which are disclosed at present, so as to obtain the optimal parameters.
And S3, selecting the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models.
The preset evaluation algorithm is a scoring algorithm for selecting the most elegant model according to evaluation indexes integrating accuracy, stability and orderliness, and the accuracy of an analysis result can be improved due to comprehensive performance of the model.
In the preferred embodiment of the present invention, the score of each optimal parameter submodel can be calculated by using the following scoring algorithm:
Score＝avg(auc,ks,-log(psi))
wherein, Score is a Score of each optimal parameter submodel under the unified evaluation index, auc is a determination accuracy of each original submodel after parameter tuning in the step S2, ks is a maximum tuning value when parameter tuning is performed when each original submodel is trained in the step S1, and psi is a number of times of training when each original submodel is trained in the step S1.
Preferably, the selecting the optimal parameter submodels by using a preset evaluation algorithm includes:
obtaining the score of each optimal parameter submodel according to a preset evaluation algorithm;
and reserving the optimal parameter submodel with the score value larger than or equal to a preset score threshold value to obtain the polymer model.
And S4, merging the polymer models into a main model according to a preset accurate merging algorithm.
In detail, the step S4 includes:
obtaining the judgment accuracy of the plurality of polymer models according to step S2, and calculating the weight of each polymer model according to the judgment accuracy;
merging the plurality of polymer models into a master model by using the following exact merging algorithm:
wherein P is the main model, wkIs the weight, p, of each of the polymer modelskIs the output result, AUC, of each of the polymer modelskIs the judgment accuracy, AUC, of the kth Polymer modeliIs the judgment accuracy of all the polymer models, and n is the total number of the polymer models.
S5, obtaining a user data set to be judged, inputting the user data set to be judged to the main model, obtaining a grade rating result of the user data set to be judged according to the output result of the main model, and outputting the grade rating result to a user.
In detail, the embodiment of the present invention obtains the user data set to be determined, inputs the user data set to be determined to the main model, compares an output result of the main model with a preset standard threshold according to an output result of the main model to obtain a level rating result of the user data set to be determined, and outputs the level rating result to a user. It should be emphasized that, in order to further ensure the privacy and security of the user data set to be determined, the user data set to be determined may also be stored in a node of a block chain.
In the embodiment of the invention, the main model is a scalable model structure, so that the robustness, the accuracy and the integrity of a model result are ensured; the submodels respectively analyze and process data from different evaluation dimensions, accurately analyze user data under each sub-dimension, find the best balance point in accuracy and robustness, merge the main models, traverse all algorithm submodel combinations of applicable sub-user data, select a merged main model of the optimal combination by integrating effectiveness, stability and promotion indexes, and promote stable operation of the main model when analyzing different user data.
Preferably, when a new user data set is introduced, the method does not need to adjust the main model, only needs to adjust the sub-model corresponding to the user data set, and then incorporates the new sub-model into the main model according to the merging algorithm, thereby simplifying the adjusting steps of the main model, reducing the repeated steps, and being beneficial to the subsequent business scene adjustment and the sub-model iterative optimization.
Further, the master model in the preferred embodiment of the present invention is constructed based on each submodel by a merge algorithm, and thus, operations of adding a new submodel and deleting an invalid submodel can be flexibly performed according to actual conditions. Under different types of data dimensions, different submodels are established, and when different types of data sources are owned due to different business scenes of different industries, scales, regions, products and the like, corresponding main models can be merged, so that the sources of user data sets are expanded, and the accuracy of the rating result based on the algorithm model is improved.
The method comprises the steps of obtaining a user data set, constructing a plurality of original submodels according to data characteristics of the user data set, training the original submodels to obtain model training results, and comprehensively evaluating enterprise grades by using a plurality of algorithms when the original submodels are constructed; according to the model training result, parameter tuning is carried out on the plurality of original sub-models, so that the accuracy of the rating result can be improved; selecting the optimal parameter submodels by using a preset evaluation algorithm, further improving the accuracy of the rating result and reducing the consumption of computing resources; the polymer models are combined into a main model according to a preset accurate combination algorithm, so that the overall flexibility of a rating mode is improved; and acquiring a user data set to be judged, inputting the user data set to be judged to the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user, so that the usability of the grade rating result is ensured. Therefore, the rating method, the rating device and the computer readable storage medium based on the algorithm model can realize the functions of comprehensively evaluating the enterprise grade by using various algorithms, improving the accuracy of the rating result and enhancing the overall flexibility of the evaluation mode.
Fig. 3 is a functional block diagram of the rating device based on the algorithm model according to the present invention.
The rating apparatus 100 based on the algorithm model according to the present invention may be installed in an electronic device. According to the implemented functions, the ranking device based on the algorithm model can comprise a sub-model building module 101, a parameter tuning module 102, a sub-model screening module 103, a main model building module 104 and a grade analyzing module 105. A module according to the present invention, which may also be referred to as a unit, refers to a series of computer program segments that can be executed by a processor of an electronic device and that can perform a fixed function, and that are stored in a memory of the electronic device.
In the present embodiment, the functions regarding the respective modules/units are as follows:
the submodel building module 101 is configured to obtain a user data set, build a plurality of original submodels according to data characteristics of the user data set, and train the plurality of original submodels to obtain a model training result;
the parameter tuning module 102 is configured to perform parameter tuning on the multiple original sub-models according to the model training result to obtain multiple optimal parameter sub-models;
the submodel screening module 103 is configured to select the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models;
the main model building module 104 is configured to merge the plurality of polymer models into a main model according to a preset accurate merging algorithm;
the level analysis module 105 is configured to obtain a user data set to be determined, input the user data set to be determined to the main model, obtain a level rating result of the user data set to be determined according to an output result of the main model, and output the level rating result to a user.
In detail, the specific implementation steps of each module of the rating device based on the algorithm model are as follows:
the submodel building module 101 obtains a user data set, builds a plurality of original submodels according to data characteristics of the user data set, and trains the original submodels to obtain a model training result.
Preferably, in the embodiment of the present invention, the user data set may be a data set of financial attributes of an enterprise, including, for example, social security data, credit rating, cash flow balance, loan query record, account settlement information, and the like of the enterprise. The user data set may be stored in a node of a block chain.
In detail, the sub-model building module 101 performs the building of the plurality of original sub-models from the data characteristics of the user data set by:
step one, acquiring the user data set, and screening the user data set to obtain an effective user data set;
preferably, the embodiment of the present invention obtains the user data set stored in the blockchain node based on a pre-joined intelligent contract. The smart contract is a computer protocol intended to propagate, verify or execute contracts in an informational manner.
Further, the embodiment of the invention screens the user data set according to a preset screening standard. The preset screening standard is that different limiting conditions are set for different data types, invalid data, repeated data and the like in the user data set are removed, resource consumption is reduced, data processing efficiency is improved, and meanwhile accuracy of analysis results is improved.
For example, the social security data of one enterprise is screened in the time range of the last three months, and the cash flow balance data is screened in the condition of the minimum amount of more than 100 yuan.
Dividing the user data set into a plurality of sub-user data sets according to the data type of the effective user data set;
for example, the valid data set of an enterprise includes social security data, credit rating, cash flow balance, loan inquiry record, account settlement information, etc., and the embodiment of the present invention may use the social security data of an enterprise as a data set of business and industry information, the credit rating and cash flow balance as a data set of credit investigation, and the loan inquiry record and account settlement information as a data set of credit investigation.
Thirdly, respectively selecting corresponding machine learning algorithms to construct a plurality of original sub-models according to the data characteristics of the sub-user data sets;
for example, the embodiment of the invention can use a first machine learning algorithm to construct a first sub-model for the data of the industry and commerce information data set, use a second machine learning algorithm to construct a second sub-model for the data of the credit investigation data set, and use a third machine learning algorithm to construct a third sub-model for the data of the credit investigation data set, so as to ensure that the data with different characteristics are fully utilized and improve the accuracy of the analysis result.
In detail, the first machine learning algorithm, the second machine learning algorithm, and the third machine learning algorithm may adopt currently disclosed machine learning algorithms, such as logistic regression, random forest, support vector machine, GBDT, naive bayes, neural network, XGboost, and the like, and the first machine learning algorithm, the second machine learning algorithm, and the third machine learning algorithm may all adopt the same machine learning algorithm, or may adopt different machine learning algorithms.
Preferably, the embodiment of the present invention inputs the plurality of sub-user data sets into the corresponding original sub-models respectively for training, so as to obtain a plurality of training results corresponding to the original sub-models.
And the parameter tuning module 102 performs parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models.
In detail, the parameter tuning module 102 in the embodiment of the present invention compares the model training result with the original correct result, and uses a ratio of the training result to the original correct result as the determination accuracy of the original sub-model.
Further, the parameter tuning module 102 in the embodiment of the present invention compares the determination accuracy with a preset accuracy standard threshold, and if the determination accuracy is lower than the standard threshold, adjusts the parameters of the multiple original sub-models by using a preset parameter tuning method until the determination accuracy is greater than or equal to the standard threshold, so as to obtain multiple optimal parameter sub-models.
The embodiment of the invention adjusts the parameters of the plurality of original sub-models, can fully utilize the data characteristics of the sub-user data set and improve the accuracy of the original sub-models. The preset parameter adjusting method can adopt a random parameter adjusting method and a grid searching method which are disclosed at present, so as to obtain the optimal parameters.
The submodel screening module 103 selects the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models.
The preset evaluation algorithm is a scoring algorithm for selecting the most elegant model according to evaluation indexes integrating accuracy, stability and orderliness, and the accuracy of an analysis result can be improved due to comprehensive performance of the model.
In the preferred embodiment of the present invention, the score of each optimal parameter submodel can be calculated by using the following scoring algorithm:
Score＝avg(auc,ks,-log(psi))
wherein, Score is the Score of each optimal parameter submodel under the unified evaluation index, auc is the judgment accuracy of each original submodel after parameter tuning in the parameter tuning module 102, ks is the maximum adjustment value when parameter tuning is performed when each original submodel is trained in the submodel building module 101, and psi is the training frequency when each original submodel is trained in the submodel building module 101.
Preferably, the sub-model screening module 103 performs the selection of the optimal parameter sub-models by using a preset evaluation algorithm by:
obtaining the score of each optimal parameter submodel according to a preset evaluation algorithm;
and reserving the optimal parameter submodel with the score value larger than or equal to a preset score threshold value to obtain the polymer model.
The main model building module 104 merges the plurality of polymer models into a main model according to a preset precise merging algorithm.
In detail, the master model building module 104 performs the merging of the plurality of polymer models into the master model according to a preset exact merging algorithm by:
calculating the weight of each polymer model according to the judgment accuracy of the plurality of polymer models obtained by the parameter tuning module 102;
merging the plurality of polymer models into a master model by using the following exact merging algorithm:
wherein P is the main model, wkIs the weight, p, of each of the polymer modelskIs the output result, AUC, of each of the polymer modelskIs the judgment accuracy, AUC, of the kth Polymer modeliIs the judgment accuracy of all polymer models, and n is the total number of the polymer models
The level analysis module 105 obtains a user data set to be judged, inputs the user data set to be judged to the main model, obtains a level rating result of the user data set to be judged according to an output result of the main model, and outputs the level rating result to a user.
In detail, in the embodiment of the present invention, the level analysis module 105 acquires the user data set to be determined, inputs the user data set to be determined into the main model, compares an output result of the main model with a preset standard threshold according to an output result of the main model, obtains a level rating result of the user data set to be determined, and outputs the level rating result to the user. It should be emphasized that, in order to further ensure the privacy and security of the user data set to be determined, the user data set to be determined may also be stored in a node of a block chain.
In the embodiment of the invention, the main model is a scalable model structure, so that the robustness, the accuracy and the integrity of a model result are ensured; the submodels respectively analyze and process data from different evaluation dimensions, accurately analyze user data under each sub-dimension, find the best balance point in accuracy and robustness, merge the main models, traverse all algorithm submodel combinations of applicable sub-user data, select a merged main model of the optimal combination by integrating effectiveness, stability and promotion indexes, and promote stable operation of the main model when analyzing different user data.
Preferably, when a new user data set is introduced, the method does not need to adjust the main model, only needs to adjust the sub-model corresponding to the user data set, and then incorporates the new sub-model into the main model according to the merging algorithm, thereby simplifying the adjusting steps of the main model, reducing the repeated steps, and being beneficial to the subsequent business scene adjustment and the sub-model iterative optimization.
Further, the master model in the preferred embodiment of the present invention is constructed based on each submodel by a merge algorithm, and thus, operations of adding a new submodel and deleting an invalid submodel can be flexibly performed according to actual conditions. Under different types of data dimensions, different submodels are established, and when different types of data sources are owned due to different business scenes of different industries, scales, regions, products and the like, corresponding main models can be merged, so that the sources of user data sets are expanded, and the accuracy of the rating result based on the algorithm model is improved.
Fig. 4 is a schematic structural diagram of an electronic device implementing the rating method based on the algorithm model according to the present invention.
The electronic device 1 may comprise a processor 10, a memory 11 and a bus, and may further comprise a computer program, such as an algorithmic model based rating program 12, stored in the memory 11 and operable on the processor 10.
The memory 11 includes at least one type of readable storage medium, which includes flash memory, removable hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc. The memory 11 may in some embodiments be an internal storage unit of the electronic device 1, such as a removable hard disk of the electronic device 1. The memory 11 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in mobile hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the electronic device 1. Further, the memory 11 may also include both an internal storage unit and an external storage device of the electronic device 1. The memory 11 may be used not only for storing application software installed in the electronic device 1 and various types of data, such as codes of the rating program 12 based on an algorithm model, etc., but also for temporarily storing data that has been output or is to be output.
The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips. The processor 10 is a Control Unit (Control Unit) of the electronic device, connects various components of the whole electronic device by using various interfaces and lines, and executes various functions and processes data of the electronic device 1 by running or executing programs or modules (for example, executing a rating program based on an algorithm model, etc.) stored in the memory 11 and calling data stored in the memory 11.
The bus may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is arranged to enable connection communication between the memory 11 and at least one processor 10 or the like.
Fig. 4 only shows an electronic device with components, and it will be understood by a person skilled in the art that the structure shown in fig. 4 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than shown, or a combination of certain components, or a different arrangement of components.
For example, although not shown, the electronic device 1 may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 10 through a power management device, so as to implement functions of charge management, discharge management, power consumption management, and the like through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device 1 may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
Further, the electronic device 1 may further include a network interface, and optionally, the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a bluetooth interface, etc.), which are generally used for establishing a communication connection between the electronic device 1 and other electronic devices.
Optionally, the electronic device 1 may further comprise a user interface, which may be a Display (Display), an input unit (such as a Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable for displaying information processed in the electronic device 1 and for displaying a visualized user interface, among other things.
It is to be understood that the described embodiments are for purposes of illustration only and that the scope of the appended claims is not limited to such structures.
The algorithmic model based rating program 12 stored by the memory 11 in the electronic device 1 is a combination of instructions that, when executed in the processor 10, may implement:
acquiring a user data set, constructing a plurality of original submodels according to the data characteristics of the user data set, and training the original submodels to obtain a model training result;
performing parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models;
selecting the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models;
merging the polymer models into a main model according to a preset accurate merging algorithm;
and acquiring a user data set to be judged, inputting the user data set to be judged to the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user.
Further, the integrated modules/units of the electronic device 1, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. The computer-readable medium may include: any entity or device capable of carrying said computer program code, recording medium, U-disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM).
Further, the computer usable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method can be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof.
The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any accompanying claims should not be construed as limiting the claim concerned.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the system claims may also be implemented by one unit or means in software or hardware. The terms second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (10)
1. An algorithmic model based rating method, the method comprising:
acquiring a user data set, constructing a plurality of original submodels according to the data characteristics of the user data set, and training the original submodels to obtain a model training result;
performing parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models;
selecting the optimal parameter submodels by using a preset evaluation algorithm to obtain a plurality of polymer models;
merging the polymer models into a main model according to a preset accurate merging algorithm;
and acquiring a user data set to be judged, inputting the user data set to be judged to the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user.
2. An algorithmic model based rating method as defined in claim 1, wherein the user data set is stored in a block chain, and the constructing of the plurality of primitive submodels from the data features of the user data set comprises:
screening the user data set to obtain an effective user data set;
dividing the user data set into a plurality of sub-user data sets according to the data type of the data in the effective user data set;
and respectively selecting corresponding machine learning algorithms to construct a plurality of original sub-models according to the data characteristics of the sub-user data sets.
3. The algorithmic-model-based rating method of claim 1, wherein the performing parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models comprises:
taking the ratio of the training result to the original correct result as the judgment accuracy of the original sub-model;
comparing the judgment accuracy with a preset accuracy standard threshold;
if the judgment accuracy is lower than the standard threshold, adjusting the parameters of the plurality of original sub-models by using a preset parameter adjustment method;
and obtaining a plurality of optimal parameter submodels until the judgment accuracy is greater than or equal to the standard threshold.
4. The algorithm model-based rating method of claim 3, wherein the selecting the optimal parameter submodels using a predetermined evaluation algorithm to obtain a plurality of polymer models comprises:
obtaining the score of each optimal parameter submodel according to a preset evaluation algorithm;
and reserving the optimal parameter submodel with the score value larger than or equal to a preset score threshold value to obtain the polymer model.
5. The algorithmic model based rating method of claim 4, wherein the deriving a score for each of the optimal parameter submodels according to a predetermined evaluation algorithm comprises:
calculating the score of each optimal parameter submodel by adopting the following algorithm:
Score＝avg(auc,ks,-log(psi))
wherein, Score is the Score of each optimal parameter submodel under the unified evaluation index, auc is the judgment accuracy of each original submodel, ks is the maximum adjustment value when parameter adjustment is executed when each original submodel is trained, and psi is the training frequency when each original submodel is trained.
6. The algorithmic-model-based rating method of claim 3, wherein the merging the plurality of polymer models into a master model according to a preset exact merge algorithm comprises:
calculating the weight of each polymer model according to the judgment accuracy of the plurality of polymer models;
merging the plurality of polymer models into a master model by using the following exact merging algorithm:
wherein P is the main model, wkIs the weight, p, of each of the polymer modelskIs the output result, AUC, of each of the polymer modelskIs the judgment accuracy, AUC, of the kth Polymer modeliIs all high scoreAnd judging accuracy of the sub-model, wherein n is the total number of the polymer models.
7. An algorithmic model based rating method as defined in any of claims 1 to 6, the method further comprising:
when a new user data set with different data characteristics from those in the user data set is obtained, selecting a corresponding original sub-model according to the new user data set, and performing parameter adjustment on the corresponding original sub-model to obtain an updated original sub-model;
and merging the updated original sub-model into the main model.
8. An algorithmic model based rating apparatus, the apparatus comprising:
the submodel building module is used for obtaining a user data set, building a plurality of original submodels according to the data characteristics of the user data set, and training the plurality of original submodels to obtain a model training result;
the parameter tuning module is used for performing parameter tuning on the plurality of original sub-models according to the model training result to obtain a plurality of optimal parameter sub-models;
the submodel screening module is used for selecting the optimal parameter submodels by utilizing a preset evaluation algorithm to obtain a plurality of polymer models;
the master model building module is used for merging the plurality of polymer models into a master model according to a preset accurate merging algorithm;
and the grade analysis module is used for acquiring a user data set to be judged, inputting the user data set to be judged into the main model, obtaining a grade rating result of the user data set to be judged according to an output result of the main model, and outputting the grade rating result to a user.
9. An electronic device, characterized in that the electronic device comprises:
a memory storing at least one instruction; and
a processor executing instructions stored in the memory to perform an algorithmic model based rating method as claimed in any of claims 1 to 7.
10. A computer-readable storage medium comprising a storage data area storing data created from use of blockchain nodes and a storage program area storing a computer program, wherein the computer program, when executed by a processor, implements an algorithmic model based rating method as defined in any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010593982.5A CN111768096A (en) | 2020-06-24 | 2020-06-24 | Rating method and device based on algorithm model, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010593982.5A CN111768096A (en) | 2020-06-24 | 2020-06-24 | Rating method and device based on algorithm model, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111768096A true CN111768096A (en) | 2020-10-13 |
Family
ID=72722041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010593982.5A Pending CN111768096A (en) | 2020-06-24 | 2020-06-24 | Rating method and device based on algorithm model, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111768096A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113569975A (en) * | 2021-08-04 | 2021-10-29 | 华南师范大学 | Sketch work rating method and device based on model fusion |
| CN113607683A (en) * | 2021-08-09 | 2021-11-05 | 天津九光科技发展有限责任公司 | Automatic modeling method for near infrared spectrum quantitative analysis |
| CN113792872A (en) * | 2021-08-24 | 2021-12-14 | 浙江数秦科技有限公司 | Neural network training container and hosting training method based on block chain |
| CN113866638A (en) * | 2021-08-24 | 2021-12-31 | 陈九廷 | Battery parameter inference method, device, equipment and medium |
| CN113935400A (en) * | 2021-09-10 | 2022-01-14 | 东风商用车有限公司 | Vehicle fault diagnosis method, device and system and storage medium |
-
2020
- 2020-06-24 CN CN202010593982.5A patent/CN111768096A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113569975A (en) * | 2021-08-04 | 2021-10-29 | 华南师范大学 | Sketch work rating method and device based on model fusion |
| CN113607683A (en) * | 2021-08-09 | 2021-11-05 | 天津九光科技发展有限责任公司 | Automatic modeling method for near infrared spectrum quantitative analysis |
| CN113792872A (en) * | 2021-08-24 | 2021-12-14 | 浙江数秦科技有限公司 | Neural network training container and hosting training method based on block chain |
| CN113866638A (en) * | 2021-08-24 | 2021-12-31 | 陈九廷 | Battery parameter inference method, device, equipment and medium |
| CN113792872B (en) * | 2021-08-24 | 2024-05-28 | 浙江数秦科技有限公司 | Neural network training container and managed training method based on block chain |
| CN113935400A (en) * | 2021-09-10 | 2022-01-14 | 东风商用车有限公司 | Vehicle fault diagnosis method, device and system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111768096A (en) | Rating method and device based on algorithm model, electronic equipment and storage medium | |
| CN111652278B (en) | User behavior detection method, device, electronic equipment and medium | |
| CN111340240A (en) | Method and device for realizing automatic machine learning | |
| CN111339072B (en) | User behavior-based change value analysis method and device, electronic equipment and medium | |
| CN111694844A (en) | Enterprise operation data analysis method and device based on configuration algorithm and electronic equipment | |
| CN113516417A (en) | Service evaluation method and device based on intelligent modeling, electronic equipment and medium | |
| CN114612194A (en) | Product recommendation method and device, electronic equipment and storage medium | |
| CN112699142A (en) | Cold and hot data processing method and device, electronic equipment and storage medium | |
| CN114491047A (en) | Multi-label text classification method and device, electronic equipment and storage medium | |
| CN114881616A (en) | Business process execution method and device, electronic equipment and storage medium | |
| CN113268665A (en) | Information recommendation method, device and equipment based on random forest and storage medium | |
| CN112465141A (en) | Model compression method, model compression device, electronic device and medium | |
| CN113656690B (en) | Product recommendation method and device, electronic equipment and readable storage medium | |
| CN113742069A (en) | Capacity prediction method and device based on artificial intelligence and storage medium | |
| CN112214602B (en) | Humor-based text classification method and device, electronic equipment and storage medium | |
| CN112269875B (en) | Text classification method, device, electronic equipment and storage medium | |
| CN112700006A (en) | Network architecture searching method, device, electronic equipment and medium | |
| CN116562588A (en) | Enterprise supply chain analysis system, method and equipment based on ERP | |
| CN113705201B (en) | Text-based event probability prediction evaluation algorithm, electronic device and storage medium | |
| CN111652281B (en) | Information data classification method, device and readable storage medium | |
| CN111652741B (en) | User preference analysis method, device and readable storage medium | |
| WO2022160443A1 (en) | Lineage mining method and apparatus, electronic device and computer-readable storage medium | |
| CN114862140A (en) | Behavior analysis-based potential evaluation method, device, equipment and storage medium | |
| CN113657546A (en) | Information classification method and device, electronic equipment and readable storage medium | |
| CN113343102A (en) | Data recommendation method and device based on feature screening, electronic equipment and medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201013 |