CN116580249A - Method, system and storage medium for classifying beats based on ensemble learning model - Google Patents
Method, system and storage medium for classifying beats based on ensemble learning model Download PDFInfo
- Publication number
- CN116580249A CN116580249A CN202310664036.9A CN202310664036A CN116580249A CN 116580249 A CN116580249 A CN 116580249A CN 202310664036 A CN202310664036 A CN 202310664036A CN 116580249 A CN116580249 A CN 116580249A
- Authority
- CN
- China
- Prior art keywords
- classification
- target
- beat
- learning model
- beats
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000012549 training Methods 0.000 claims description 53
- 238000012795 verification Methods 0.000 claims description 30
- 238000004590 computer program Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000011084 recovery Methods 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 15
- 238000000605 extraction Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 239000002699 waste material Substances 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000008901 benefit Effects 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
- 238000004458 analytical method Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004064 recycling Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007726 management method 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
- 238000000513 principal component analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble 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
- G06Q10/00—Administration; Management
- G06Q10/30—Administration of product recycling or disposal
-
- 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
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/08—Auctions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Mathematical Physics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Sustainable Development (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides a beat classification method, a system and a storage medium based on an integrated learning model. The method is applied to an integrated learning model system, the integrated learning model system comprises an information blockchain, and the method comprises the following steps: acquiring first image information of a target beat product from a target information node, and extracting a plurality of first image features of the target beat product according to the first image information; calculating a first classification degree of the target beat product according to the plurality of first image features, wherein the first classification degree is used for indicating the actual classification value of the target beat product; based on the trained integrated learning model, predicting a second classification degree of the target beat according to the plurality of first image features, wherein the second classification degree is used for indicating the predicted classification value of the target beat; and classifying the target beats according to the first classification degree and the second classification degree. The recovery efficiency of the clapping products can be improved.
Description
Technical Field
The application relates to the technical field of intelligent recognition, in particular to a beat classification method, a system and a storage medium based on an integrated learning model.
Background
Along with the continuous development of science and technology, enterprises can generate a large amount of waste materials or waste articles such as second-hand equipment in the process of continuously updating and upgrading hardware facilities, and if the waste articles are directly discarded, a large amount of resource waste can be caused. Through the mode of resource recovery and then auction, not only can bring certain recovery benefit for former enterprise, can sell old and useless article for suitable new enterprise again, promote resource recycling and recycle, avoid extravagant, promote green development.
Most of the waste articles (namely auction articles, short for beats) are collected on site by the staff carrying the auction enterprises, and the waste articles are classified by manual analysis, so that a great deal of time is required, and the labor cost is high.
The applicant found that for some small beats, value assessment and classification can be achieved by appearance basically, without the need for personnel to evaluate on site. However, in the above-mentioned claps, classification is usually performed by manual recognition or simple image recognition, so that the recognition effect is poor and the accuracy is low.
Disclosure of Invention
The application provides a beat classification method, a system and a storage medium based on an integrated learning model, which are used for solving the problems that part of beats are required to be classified by mostly simple image recognition, the recognition effect is poor and the accuracy is low.
In a first aspect, the application provides a beat classification method based on an integrated learning model, which is applied to an integrated learning model system, wherein the integrated learning model system comprises an information block chain, and each information node of the information block chain correspondingly stores image information of an uploaded beat; the method comprises the following steps:
acquiring first image information of a target beat from a target information node, extracting a plurality of first image features of the target beat according to the first image information, wherein the target information node is any information node of an information block chain;
calculating a first classification degree of the target beat product according to the plurality of first image features, wherein the first classification degree is used for indicating the actual classification value of the target beat product;
based on the trained integrated learning model, predicting a second classification degree of the target beat according to the plurality of first image features, wherein the second classification degree is used for indicating the predicted classification value of the target beat;
and classifying the target beats according to the first classification degree and the second classification degree.
In one possible implementation, classifying the target beat according to the first classification degree and the second classification degree includes:
calculating the ratio of the first classification degree to the second classification degree as a target deviation value;
And classifying the target beats according to the target deviation value.
In one possible implementation, the integrated learning model system further includes a classification blockchain including a first sub-chain, a second sub-chain, and a third sub-chain;
classifying the target beats according to the target deviation value, including:
when the target deviation value is smaller than a first preset deviation value, marking the target beat as stable, and uploading the image information of the target beat to a first sub-chain;
when the target deviation value is not smaller than the first preset deviation value and smaller than the second preset deviation value, marking the target beat as reasonable, and uploading the image information of the target beat to a second sub-chain;
when the target deviation value is larger than a second preset deviation value, marking the target beat as a risk, and uploading the image information of the target beat to a third sub-chain;
the first preset deviation value is smaller than the second preset deviation value.
In one possible implementation, the integrated learning model system further includes a record blockchain, each record node of the record blockchain storing second image information and selling price of a sold beat product;
calculating a first degree of classification of the target beat product according to the plurality of first image features, including:
Acquiring second image information and selling price of the sold beats of each recording node;
extracting a plurality of second image features of each sold beat product according to the second image information;
calculating M similarity of each first image feature and each second image feature;
based on a predetermined first formula, a first degree of classification of the target beat product is calculated from the M degrees of similarity and the selling prices of the respective sold beat products.
In one possible implementation, the first formula includes:
wherein C is z A first classification degree of the target beat z, A zq Representing the similarity between the target beat z and the sold beat Q, Q q Shows the selling price of the sold beats q, P zi,qj The similarity between the ith first image feature of the target racket and the jth second image feature of the sold racket q is represented, and lambda represents a selling price fluctuation parameter, and 0 < lambda < 1.
In one possible implementation, the integrated learning model system further includes a standard blockchain in which third image information and standard selling prices of the first number of standard beats are stored;
the method further includes, prior to acquiring the first image information of the target beat from the target information node:
acquiring third image information and selling price of a second number of standard beats from the standard blockchain, wherein the second number is not more than the first number;
Calculating a third image characteristic of each standard racket, and calculating a third classification degree of each standard racket according to the third image characteristic, wherein the third classification degree is used for indicating the theoretical classification value of the standard racket;
and training the integrated learning model according to the third image characteristics and the third classification degree of the second number of standard beats to obtain a trained integrated learning model.
In one possible implementation, the ensemble learning model includes a first prediction layer, a second prediction layer, and a third prediction layer, where the models included in the first prediction layer, the second prediction layer, and the third prediction layer are not identical;
training the integrated learning model according to the third image features and the third classification degree of the second number of standard beats to obtain a trained integrated learning model, including:
dividing the second number of standard beats into a first training set and a first verification set, wherein the first training set comprises image features and third classification degrees of X standard beats, and the first verification set comprises image features and third classification degrees of Y standard beats;
training the first prediction layer according to the first training set and the first verification set to obtain a trained first prediction layer;
dividing the output result of the first prediction layer into a second training set and a second verification set, and training the second prediction layer according to the second training set and the second verification set to obtain a trained second prediction layer;
Dividing the output result of the second prediction layer into a third training set and a third verification set, and training the third prediction layer according to the third training set and the third verification set to obtain a trained third prediction layer;
and constructing a trained integrated learning model according to the trained first prediction layer, the trained second prediction layer and the trained third prediction layer.
In one possible implementation, the formula for calculating the third classification degree of each standard beat product is:
wherein C is y A third classification degree of the standard racket y, B yp The similarity between the standard racket y and other standard racket p is represented, lambda represents the price fluctuation parameter, and 0 is less than lambda<1。
In a second aspect, the present application provides a beat classification device based on an ensemble learning model, applied to an ensemble learning model system, where the ensemble learning model system includes an information blockchain and a value blockchain, and each information node of the information blockchain correspondingly stores image information of an uploaded beat, the device includes:
the first acquisition module is used for acquiring first image information of the target beat from a target information node, extracting a plurality of first image features of the target beat according to the first image information, wherein the target information node is an information node of an information blockchain;
The first calculating module is used for calculating a first classification degree of the target beat product according to the plurality of first image features, and the first classification degree is used for indicating the actual classification value of the target beat product;
the second calculation module is used for predicting a second classification degree of the target beat according to the plurality of first image features based on the trained integrated learning model, and the second classification degree is used for indicating the predicted classification value of the target beat;
and the classification module is used for classifying the target beats according to the first classification degree and the second classification degree.
In a third aspect, the present application provides an AI classification system, including a memory and a processor, where the memory stores a computer program executable on the processor, and the processor implements the steps of the integrated learning model-based beat classification method as described above in the first aspect or any one of the possible implementations of the first aspect when the processor executes the computer program.
In a fourth aspect, the present application provides a computer readable storage medium storing a computer program which when executed by a processor implements the steps of the beat classification method according to the first aspect or any one of the possible implementations of the first aspect.
The application provides a method, a system and a storage medium for classifying beats based on an ensemble learning model. The first image characteristics are extracted by acquiring the image information of the target beat, then the first classification degree of the target beat is calculated according to the plurality of first image characteristics, and the second classification degree of the target beat is predicted according to the plurality of first image characteristics based on the trained integrated learning model. Finally, classifying the target beats according to the first classification degree and the second classification degree, and compared with simple image recognition, the integrated learning model is introduced, the whole process of judgment is automatic and intelligent, recognition is more accurate, efficiency is higher, the judging efficiency of the beats can be greatly improved, and further the efficiency of resource circulation is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of an implementation of a beat classification method based on an ensemble learning model provided by an embodiment of the present application;
fig. 2 is a schematic structural diagram of a beat classification device based on an integrated learning model according to an embodiment of the present application;
fig. 3 is a schematic diagram of an AI classification system according to an embodiment of the application.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
The embodiment of the application mainly aims at some small beats, and can evaluate the value of the auction articles basically through the appearance images. Such as a second-hand electronic device or a small second-hand industrial device, etc. The AI intelligent (Artificial Intelligence ) recognition system of embodiments of the present application may include a plurality of blockchains, with different blockchains having different functions to ensure reliability of information.
Alternatively, the integrated learning model system in embodiments of the present application may include an information blockchain, a classification blockchain, a record blockchain, and a standard blockchain.
The information blockchain may include a plurality of information nodes, each of which may store image information of an uploaded beat. The seller user can upload the image information of the bat to the idle information node of the information blockchain, or the seller user sends the image information of the bat to the bat, and the bat uploads the image information of the corresponding bat to the idle information node.
Each information node can store the image information of a plurality of beats of the same seller, and can also store the image information of one beat. The setting can be specifically performed according to actual conditions. In order to ensure the reliability of calculation, in the embodiment of the application, each information node only stores the image information of one uploaded beat.
The classification blockchain comprises a plurality of subchains, the classification blockchain is used for storing classification information of each bat for an auctioneer, and different subchains can store information of different classes of bat, including image information, vendor identity information and the like. The classes of the beats can be divided according to actual conditions, for example, the beats can be divided into general, stable, reasonable and risk, primary, secondary and tertiary, class A, class B, class C and the like, and specific classes can be determined according to the classification degree.
Each classification range or classification level can correspond to one sub-chain, different sub-chains store different types of beats, and the auctioneer can determine the types of different beats from different sub-chains, so that subsequent further evaluation or direct setting of a reserve price is facilitated.
The record blockchain may include a plurality of record nodes, each of which may store image information and a selling price of a sold beat, and the record blockchain may also be referred to as a history blockchain. The auctioneer can upload the image information and the selling price of each sold item to the corresponding recording node.
The standard blockchain can comprise a plurality of standard nodes, each standard node can store image information of a standard racket and standard selling price of the standard racket, and data in the standard blockchain can be uploaded by a racket.
According to the embodiment of the application, by arranging the plurality of block chains and storing different information in different block chains, the reliability of the information can be ensured, the information is prevented from being tampered, and the reliability of the classification method provided by the embodiment of the application is ensured.
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the following description will be made by way of specific embodiments with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of an implementation of a beat classification method based on an ensemble learning model according to an embodiment of the present application is shown. As shown in fig. 1, a beat classification method based on an ensemble learning model is applied to an ensemble learning model system, where the ensemble learning model system includes an information blockchain, and each information node of the information blockchain correspondingly stores image information of an uploaded beat, and the method may include S101 to S104.
S101, acquiring first image information of a target beat from a target information node, and extracting a plurality of first image features of the target beat according to the first image information, wherein the target information node is any information node of an information block chain.
The execution subject of the embodiment of the application is an integrated learning model system, and the integrated learning model system can execute the classification process after the user of the seller uploads the beat. Or searching whether a new beat is uploaded by the integrated learning model system every preset time interval, and if so, executing a classification process. Specifically, the selection can be performed according to actual conditions.
The target bat is a bat uploaded by a seller user, and the first image information can comprise color picture information such as a front view, a left view, a right view, a top view, a partial detail view and the like of the target bat.
Based on related image feature extraction means, feature extraction is carried out on a plurality of pictures, a plurality of first image features of a target beat can be obtained, the first image features can comprise corner features, edge features, gradient point features, texture features and the like, and specific extraction means can comprise principal component analysis, singular value decomposition, linear discriminant analysis and the like.
Each picture of the target beat product can correspondingly extract a first image feature, and each image feature can be expressed as a feature vector value, namely, the feature values of a plurality of first image features of the target beat product can be extracted according to the first image information.
S102, calculating a first classification degree of the target beat product according to the plurality of first image features, wherein the first classification degree is used for indicating the actual classification value of the target beat product.
Alternatively, after obtaining the plurality of first image features of the target beat product, the first classification degree of the target beat product may be calculated according to the plurality of first image features. The feature values of the plurality of first image features may be added or multiplied as the first degree of classification of the target beat. Or the different first image features correspond to different weights, and the first classification degree of the target beat is calculated according to the different weights.
The first classification degree is used for indicating the actual classification value of the target bat, namely, the target bat is obtained through real-time calculation according to the image information of the target bat uploaded by a seller user in the target information section.
The integrated learning model system further comprises a record block chain, wherein each record node of the record block chain stores second image information and selling price of one sold bat, and the first classification degree of the target bat can be calculated by combining the information of the sold bat in the record block chain.
Specifically, the process of calculating the first degree of classification may include S1021 to S1024.
S1021, second image information and selling prices of the sold beats of each recording node are acquired.
The ensemble learning model system may read relevant information of sold beats from the record blockchain. Second image information and selling prices of sold beats are extracted from the recording node that records the blockchain. The second image information and selling price of all sold beats can be extracted, and the second image information and selling price of a certain number of sold beats can be extracted, and can be determined according to the importance level of the target beat vendor user.
Specifically, the user authority level can be determined according to the seller user information of the target beats, and the number of the sold beats is determined and extracted according to the user authority level. The user permission level is proportional to the number of extractions.
And when the user authority level is smaller than the first preset level, extracting second image information and selling prices of the S sold beats from the authentication blockchain.
And when the user authority level is not smaller than the first preset level and smaller than the second preset level, extracting second image information and selling prices of F sold beats from the authentication block chain.
And when the user authority level is not smaller than a second preset level, extracting second image information and selling prices of the G sold beats from the authentication blockchain. Wherein S < F < G and S, F, G are positive integers not less than 3.
Different numbers of sold beats are extracted according to different user rights, so that the occupation amount of resources can be reduced while the user requirements are met, and the overall calculation efficiency is improved.
S1022, extracting a plurality of second image features of each sold bat product according to the second image information.
The second image features of all sold beats can be extracted by adopting the image feature extraction mode same as that of the target beat, and the feature values of the second image features are determined so as to ensure that the dimensions of the second image features and the first image features are the same and ensure the reliability of calculation.
S1023, calculating M similarity of each first image feature and each second image feature.
For each first image feature, the similarity of the first image feature and the respective second image feature may be calculated separately.
In view of the fact that the first image feature and the second image feature are feature vectors, the similarity between the first image feature and the second image feature can be selected according to actual situations by taking one of cosine similarity, correlation coefficient, ming's distance, euclidean distance and Mahalanobis distance as the similarity between the first image feature and the second image feature.
According to the embodiment of the application, the similarity of the target bat and the sold bat is evaluated by calculating the characteristic vector value, so that the comparison process is vectorized, and the target bat can be evaluated more accurately and reliably.
S1024, calculating the first classification degree of the target beats according to the M similarities and the selling prices of the sold beats.
For each sold beat item whose similarity is calculated with the target beat item, all the similarity of the sold beat item and the target beat item can be added to obtain the total similarity of the sold beat item and the target beat item, then the total similarity is multiplied by the selling price of the sold product, and the obtained value is taken as the preselected classification degree of the sold beat item and the target beat item.
Alternatively, all the pre-selected classifications may be added as the first classification of the target beat. Alternatively, the average value of all the pre-selected classifications is taken as the first classification of the target beat. Alternatively, the maximum value of all the preselected classifications is selected as the first classification of the target beat. Alternatively, the smallest value among all the preselected classifications is selected as the first classification of the target beat. Specifically, the selection can be performed according to actual conditions.
For example, the first degree of classification of the target beat item may be calculated based on a predetermined first formula from the M degrees of similarity and the selling prices of the respective sold beat items.
The first formula includes:
C z a first classification degree of the target beat z, A zq Representing the similarity between the target beat z and the sold beat Q, Q q Shows the selling price of the sold beats q, P zi,qj The similarity between the ith first image feature of the target racket and the jth second image feature of the sold racket q is represented, and lambda represents a selling price fluctuation parameter, and 0 < lambda < 1. Where λ may be preset, for example, λ=0.5, or λ=0.7, with a larger λ indicating easier fluctuation of the current selling price.
Illustratively, λ=0.5, m=2, a=1, b=2, q 1 =0.1w,Q 2 =0.2w,P z1,11 =0.5,P z1,12 =0.9,P z1,21 =0.3,P z1,22 =0.7。
A z1 =0.7,A z2 =0.5
C z =0.17
The first degree of classification of the target beat piece z at this time may be 0.17.
According to the embodiment of the application, the first classification degree of the target beat product is calculated through the sold beat product, so that the real-time classification value of the target beat product can be truly reflected.
S103, based on the trained integrated learning model, predicting a second classification degree of the target beat according to the plurality of first image features, wherein the second classification degree is used for indicating the predicted classification value of the target beat.
The integrated learning model comprises a plurality of prediction layers, each prediction layer can comprise a prediction model, the prediction models of the prediction layers are not identical, the plurality of prediction layers are in linear series connection, the output of one prediction layer can be used as the input of the next prediction layer, and the second classification degree of the target shooting can be obtained through linear series connection and multi-layer prediction. The second classification degree is obtained according to the integrated learning model prediction and is a predicted value.
Specifically, the integrated learning model system further includes a standard blockchain in which third image information of the first number of standard beats and a standard selling price are stored. The third image information of the standard racket is the standard image information shot by the standard racket in an ideal state, the standard selling price is the auction price of the standard racket in the ideal state, the standard racket can be used as an ideal standard racket, and the integrated learning model can be reasonably trained. The third image information and the standard selling price of the standard racket can be uploaded to the standard blockchain by the racket, so that the reliability of the standard, the stability of calculation and the non-falsification of data transmission are guaranteed, and the unification of calculation standards is guaranteed.
Before classifying the target beats, the integrated learning model is trained, so that a trained integrated learning model is obtained. The training process may be as follows:
and acquiring third image information and selling price of a second number of standard beats from the standard blockchain, wherein the second number is not more than the first number. The second number can be determined according to the user rights, and the corresponding integrated learning model can be trained according to different user rights, so that the unification of target beat calculation standards corresponding to the user rights is ensured, and the calculation reliability is ensured.
And calculating a third image characteristic of each standard racket, and calculating a third classification degree of each standard racket according to the third image characteristic, wherein the third classification degree is used for indicating the theoretical classification value of the standard racket. The third image feature of the standard racket can be obtained by adopting the image feature extraction mode which is the same as that of the target racket, and the third image feature is expressed as a feature vector.
The formula for calculating the third classification degree of each standard beat product is as follows:
wherein C is y A third classification degree of the standard racket y, B yp The similarity of the standard racket y and other standard racket p is represented, lambda represents the selling price fluctuation parameter, and lambda is more than 0 and less than 1.
And training the integrated learning model according to the third image characteristics and the third classification degree of the second number of standard beats to obtain a trained integrated learning model.
The second number of standard beats can be divided into a verification set and a training set, each prediction layer in the integrated learning model is trained by adopting the training set, then each prediction layer in the integrated learning model is verified by adopting the verification set, and the trained integrated learning model can be obtained after the standard beats are qualified.
The ensemble learning model may include one or more of an MLP prediction layer, an XGB prediction layer, a CNN prediction layer, and an RF prediction layer, and may specifically be selected according to practical situations.
S104, classifying the target beats according to the first classification degree and the second classification degree.
The first classification degree is used for representing the actual classification value of the target beat product, and the second classification degree is used for representing the prediction classification value of the target beat product. When the first classification degree and the second classification degree have larger differences, the classification value of the target beat product is shown to have larger fluctuation, and certain risks exist in the trade classification, so that the classification risks can be reduced by manually determining the classification risks again. When the difference between the first classification degree and the second classification degree is smaller, the classification value fluctuation of the target beat is smaller, the classification risk is smaller, and classification can be performed according to actual conditions without manual determination.
The classification is finally to distribute the target beats to the corresponding reasonable auction reserve price or auction bid price category, and the classification is carried out primarily through the integrated learning model system, so that the unreliability of manual subjective classification can be reduced, the time consumption of manual classification is reduced, the running efficiency of auction enterprises can be improved, the circulation efficiency of the beats is accelerated, and the resource circulation efficiency is further improved.
Alternatively, the classification process may include: and calculating the ratio of the first classification degree to the second classification degree as a target deviation value. Alternatively, the absolute value of the difference between the first classification degree and the second classification degree is calculated as the target deviation value. Alternatively, a sum of the first classification degree and the second classification degree is calculated as the target deviation value. And finally classifying the target beats according to the target deviation value.
For target beats with high classification risk, the trade classification may result in too high or too low a setting of the reserve price, resulting in a streaming beat of the target beat or a sale of the reserve price, resulting in a loss of the benefit of the seller.
According to the embodiment of the application, the reliability of data storage is ensured by combining a blockchain storage technology, the use requirements of seller users and clapping parties are met, and the classification degree of the target clapping articles is automatically calculated by utilizing the integrated learning model system, so that the target clapping articles are reasonably classified, the labor investment is reduced, the clapping article classification efficiency can be greatly improved, the resource recycling is promoted, and the method has wide practicability.
In some embodiments of the application, the integrated learning model system further comprises a classification blockchain including a first sub-chain for storing beats classified as stable, a second sub-chain for storing beats classified as reasonable, and a third sub-chain for storing beats classified as risky. The S104 may include:
s1041, calculating a ratio of the first classification degree to the second classification degree as a target deviation value.
The target deviation value is used for indicating the deviation degree between the actual calculated value and the predicted calculated value of the target beat product. The smaller the target deviation value is, the closer the actual calculated value and the predicted calculated value are, and the more stable the classification result of the target beat product is. The larger the target deviation value is, the more the actual calculated value and the predicted calculated value are, the more the classification result of the target beat is at risk
S1042, when the target deviation value is smaller than the first preset deviation value, marking the target beat as stable, and uploading the image information of the target beat to the first sub-chain.
When the target deviation value is smaller than the first preset deviation value, the target beat product is indicated to be stable in classification result, and the preliminary estimated base price is stable, so that the image information of the target beat product can be uploaded to the first sub-chain. The stable beats marked as stable can be auctioned preferentially later.
S1043, marking the target beat as reasonable when the target deviation value is not smaller than the first preset deviation value and smaller than the second preset deviation value, and uploading the image information of the target beat to the second sub-chain.
When the target deviation value is not smaller than the first preset deviation value and smaller than the second preset deviation value, the target beats are reasonably classified, and the primarily estimated reserve price is relatively stable, so that the image information of the target beats can be uploaded to the second sub-chain, and later auctions which can be prioritized can be marked as reasonable beats.
S1044, when the target deviation value is larger than the second preset deviation value, marking the target beat as a risk, and uploading the image information of the target beat to a third sub-chain. The first preset deviation value is smaller than the second preset deviation value.
And when the target deviation value is larger than the second preset deviation value. The classification result of the target beats is indicated to be risk, the image information of the target beats can be uploaded to the third sub-chain, and each beat in the third sub-chain is required to be evaluated again by manpower later so as to reduce the possibility of streaming beats or reserve auction of the beats and ensure the rights and interests of sellers.
According to the embodiment of the application, the reliability of classification result storage can be ensured by introducing the classification block chain, and the beats of different classification results are stored in different sub-chains, so that the subsequent unified management and processing are facilitated, the standardization of enterprise data processing is improved, and the running efficiency of auction enterprises is improved.
In some embodiments of the present application, the ensemble learning model includes a first prediction layer, a second prediction layer, and a third prediction layer, the models included in the first prediction layer, the second prediction layer, and the third prediction layer are not identical.
The training the integrated learning model according to the third image features and the third classification degree of the second number of standard beats to obtain a trained integrated learning model may include:
dividing the second number of standard beats into a first training set and a first verification set, wherein the first training set comprises image features and third classifications of X standard beats, and the first verification set comprises image features and third classifications of Y standard beats.
And training the first prediction layer according to the first training set and the first verification set to obtain a trained first prediction layer.
Dividing the output result of the first prediction layer into a second training set and a second verification set, and training the second prediction layer according to the second training set and the second verification set to obtain a trained second prediction layer.
Dividing the output result of the second prediction layer into a third training set and a third verification set, and training the third prediction layer according to the third training set and the third verification set to obtain a trained third prediction layer.
And constructing a trained integrated learning model according to the trained first prediction layer, the trained second prediction layer and the trained third prediction layer.
In the embodiment of the application, the first image characteristics of the target beat can be input into a trained integrated learning model, and the second classification degree of the target beat for indicating the prediction classification value is obtained through three prediction layers and prediction.
According to the embodiment of the application, the three prediction layers are connected in series to construct the integrated learning model, so that the prediction accuracy of the second classification degree of the target beat can be ensured, and the classification accuracy is improved.
In the embodiment of the application, the process of calculating the second classification degree is similar to the process of calculating the first classification degree, except that the first classification degree of the target beat product is calculated according to the related information of the sold beat product in the record block chain, and is the actual classification value of the target beat product. The second classification degree of the target beats is obtained by prediction according to an integrated learning model established by standard beat training in a standard block chain, and is the predicted classification value of the target beats, and the classification result of the target beats can be accurately obtained by comparing the two classification values.
According to the embodiment of the application, the first classification degree and the second classification degree of the target beat product are calculated through two dimensions, and then the classification result of the target beat product is obtained through comparison, so that the manual participation is reduced in the whole process, the labor cost is reduced, the circulation efficiency of the beat product is accelerated, and the resource recovery efficiency is improved. And the related data are stored in the block chain, so that the reliability and the non-tamper property of data transmission are ensured. Meanwhile, all the beats are evaluated according to unified standards, so that the credibility and the publicity of the integrated learning model system are ensured.
It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present application.
The following are device embodiments of the application, for details not described in detail therein, reference may be made to the corresponding method embodiments described above.
Fig. 2 shows a schematic structural diagram of a beat classification device based on an ensemble learning model according to an embodiment of the present application, and for convenience of explanation, only the portions relevant to the embodiment of the present application are shown, which are described in detail below:
as shown in fig. 2, the beat classification device 20 based on the ensemble learning model is applied to an ensemble learning model system, where the ensemble learning model system includes an information blockchain, and each information node of the information blockchain correspondingly stores image information of an uploaded beat, and the device 20 may include:
A first obtaining module 201, configured to obtain first image information of a target beat from a target information node, and extract a plurality of first image features of the target beat according to the first image information, where the target information node is any information node of an information blockchain;
a first calculating module 202, configured to calculate a first classification degree of the target beat product according to the plurality of first image features, where the first classification degree is used to indicate an actual classification value of the target beat product;
the second calculation module 203 is configured to predict, based on the trained ensemble learning model, a second classification degree of the target beat according to the plurality of first image features, where the second classification degree is used to indicate a predicted classification value of the target beat;
the classification module 204 is configured to classify the target beat product according to the first classification degree and the second classification degree.
May include:
in some embodiments of the application, classification module 204 may include:
the ratio unit is used for calculating the ratio of the first classification degree to the second classification degree to be used as a target deviation value;
and the classification unit is used for classifying the target beats according to the target deviation value.
In some embodiments of the application, the integrated learning model system further includes a classification blockchain including a first sub-chain, a second sub-chain, and a third sub-chain; the classification unit may include:
The first classification subunit is used for marking the target beat as stable when the target deviation value is smaller than a first preset deviation value, and uploading the image information of the target beat to the first sub-chain;
the second classification subunit is used for marking the target beat as reasonable when the target deviation value is not smaller than the first preset deviation value and smaller than the second preset deviation value, and uploading the image information of the target beat to the second sub-chain;
the third classification subunit is used for marking the target beat as a risk when the target deviation value is larger than a second preset deviation value, and uploading the image information of the target beat to a third sub-chain;
the first preset deviation value is smaller than the second preset deviation value.
In some embodiments of the application, the integrated learning model system further comprises a record blockchain, wherein each record node of the record blockchain stores second image information and selling price of one sold beat product;
the first computing module 202 may include:
the acquisition unit is used for acquiring second image information and selling price of the sold beats of each recording node;
a first calculation unit configured to extract, for each sold beat item, a plurality of second image features of the sold beat item based on the second image information;
A second calculation unit configured to calculate M similarities for each of the first image features and the respective second image features;
and a third calculation unit for calculating a first degree of classification of the target beat product according to the M degrees of similarity and the selling prices of the respective sold beat products based on a predetermined first formula.
In some embodiments of the application, the first formula comprises:
wherein C is z A first classification degree of the target beat z, A zq Representing the similarity between the target beat z and the sold beat Q, Q q Shows the selling price of the sold beats q, P zi,qj The similarity between the ith first image feature of the target racket and the jth second image feature of the sold racket q is represented, and lambda represents a selling price fluctuation parameter, and 0 < lambda < 1.
In some embodiments of the application, the integrated learning model system further comprises a standard blockchain in which third image information and standard selling prices of the first number of standard beats are stored; the apparatus 20 may further include:
the second acquisition module is used for acquiring third image information and selling price of a second number of standard beats from the standard block chain before acquiring the first image information of the target beats from the target information node, wherein the second number is not more than the first number;
The third calculation module is used for calculating a third image characteristic of each standard racket, and calculating a third classification degree of each standard racket according to the third image characteristic, wherein the third classification degree is used for indicating the theoretical classification value of the standard racket;
and the training module is used for training the integrated learning model according to the third image characteristics and the third classification degree of the second number of standard beats to obtain a trained integrated learning model.
In some embodiments of the present application, the ensemble learning model includes a first prediction layer, a second prediction layer, and a third prediction layer, the models included in the first prediction layer, the second prediction layer, and the third prediction layer are not identical; the training module may include:
dividing the second number of standard beats into a first training set and a first verification set, wherein the first training set comprises image features and third classification degrees of X standard beats, and the first verification set comprises image features and third classification degrees of Y standard beats;
training the first prediction layer according to the first training set and the first verification set to obtain a trained first prediction layer;
dividing the output result of the first prediction layer into a second training set and a second verification set, and training the second prediction layer according to the second training set and the second verification set to obtain a trained second prediction layer;
Dividing the output result of the second prediction layer into a third training set and a third verification set, and training the third prediction layer according to the third training set and the third verification set to obtain a trained third prediction layer;
and constructing a trained integrated learning model according to the trained first prediction layer, the trained second prediction layer and the trained third prediction layer.
In some embodiments of the present application, the formula for calculating the third degree of classification for each standard beat product is:
/>
wherein C is y A third classification degree of the standard racket y, B yp The similarity of the standard racket y and other standard racket p is represented, lambda represents the selling price fluctuation parameter, and lambda is more than 0 and less than 1.
Fig. 3 is a schematic diagram of an AI classification system according to an embodiment of the application. As shown in fig. 3, the AI classification system 30 of this embodiment includes: a processor 300 and a memory 301, the memory 301 having stored therein a computer program 302 executable on the processor 300. The processor 300, when executing the computer program 302, implements the steps in the above-described embodiment of the beat classification method based on the ensemble learning model, for example, S101 to S104 shown in fig. 1. Alternatively, the processor 300, when executing the computer program 302, performs the functions of the modules/units of the apparatus embodiments described above, such as the functions of the modules 201 to 204 shown in fig. 2.
By way of example, the computer program 302 may be partitioned into one or more modules/units, which are stored in the memory 301 and executed by the processor 300 to accomplish the present application. One or more of the modules/units may be a series of computer program instruction segments capable of performing particular functions to describe the execution of the computer program 302 in the AI classification system 30. For example, the computer program 302 may be partitioned into modules 201 through 204 shown in FIG. 2.
The AI classification system 30 may be a computing device such as a desktop computer, a notebook computer, a palm top computer, and a cloud server. The AI classification system 30 can include, but is not limited to, a processor 300, a memory 301. It will be appreciated by those skilled in the art that fig. 3 is merely an example of the AI classification system 30 and is not limiting of the AI classification system 30, and may include more or fewer components than shown, or may combine certain components, or different components, e.g., the AI classification system may further include input-output devices, network access devices, buses, etc.
The processor 300 may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 301 may be an internal storage unit of the AI classification system 30, such as a hard disk or a memory of the AI classification system 30. The memory 301 may also be an external storage device of the AI classification system 30, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash Card (Flash Card) or the like, which are provided on the AI classification system 30. Further, the memory 301 may also include both internal storage units and external storage devices of the AI classification system 30. The memory 301 is used to store a computer program and other programs and data required for the AI classification system. The memory 301 may also be used to temporarily store data that has been output or is to be output.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment 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, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, the specific names of the functional units and modules are only for distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiments provided herein, it should be understood that the disclosed apparatus/AI classification systems and methods may be implemented in other ways. For example, the apparatus/AI classification system embodiments described above are merely illustrative, e.g., the division of modules or elements is merely a logical functional division, and there may be additional divisions of actual implementation, e.g., multiple elements 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 may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.
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 modules/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 storage medium. Based on such understanding, the present application may implement all or part of the procedures in the above embodiment method, or may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium, and when the computer program is executed by a processor, the computer program may implement the steps of each embodiment of the above integrated learning model-based shot classification method. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, executable files or in some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth.
The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.
Claims (10)
1. The integrated learning model system comprises an information block chain, wherein each information node of the information block chain correspondingly stores image information of an uploaded beat; the method comprises the following steps:
acquiring first image information of a target beat from a target information node, and extracting a plurality of first image features of the target beat according to the first image information, wherein the target information node is any information node of the information blockchain;
calculating a first classification degree of the target beat product according to the plurality of first image features, wherein the first classification degree is used for indicating the actual classification value of the target beat product;
Predicting a second classification degree of the target beat according to the plurality of first image features based on the trained ensemble learning model, wherein the second classification degree is used for indicating the predicted classification value of the target beat;
and classifying the target beats according to the first classification degree and the second classification degree.
2. The integrated learning model-based beat classification method of claim 1, wherein the classifying the target beat according to the first classification degree and the second classification degree comprises:
calculating the ratio of the first classification degree to the second classification degree as a target deviation value;
and classifying the target beats according to the target deviation value.
3. The integrated learning model based beat classification method of claim 2, wherein the integrated learning model system further comprises a classification blockchain comprising a first subchain, a second subchain, and a third subchain;
the classifying the target beat according to the target deviation value includes:
when the target deviation value is smaller than a first preset deviation value, marking the target beat as stable, and uploading the image information of the target beat to the first sub-chain;
When the target deviation value is not smaller than the first preset deviation value and smaller than a second preset deviation value, marking the target beat as reasonable, and uploading the image information of the target beat to the second sub-chain;
when the target deviation value is larger than the second preset deviation value, marking the target beat as a risk, and uploading the image information of the target beat to the third sub-chain;
wherein the first preset deviation value is smaller than the second preset deviation value.
4. The integrated learning model based beat classification method of claim 1, wherein the integrated learning model system further comprises a record blockchain, each record node of the record blockchain storing second image information and selling prices of one sold beat;
the calculating a first classification degree of the target beat product according to the plurality of first image features includes:
acquiring second image information and selling price of the sold beats of each recording node;
extracting a plurality of second image features of each sold beat product according to the second image information;
calculating M similarity of each first image feature and each second image feature;
Calculating a first classification degree of the target beats according to the M similarities and the selling prices of the sold beats based on a predetermined first formula;
wherein the first formula comprises:
wherein C is z A first classification degree of the target beat z, A zq Representing the similarity between the target beat z and the sold beat Q, Q q Shows the selling price of the sold beats q, P zi,qj The similarity between the ith first image feature of the target racket and the jth second image feature of the sold racket q is represented, and lambda represents a selling price fluctuation parameter, and 0 < lambda < 1.
5. The integrated learning model-based beat classification method of any one of claims 1 to 4, wherein the integrated learning model system further comprises a standard blockchain in which third image information and standard selling prices of a first number of standard beats are stored;
before the acquiring the first image information of the target beat from the target information node, the method further includes:
acquiring third image information and selling price of a second number of standard beats from the standard blockchain, wherein the second number is not more than the first number;
calculating a third image characteristic of each standard racket, and calculating a third classification degree of each standard racket according to the third image characteristic, wherein the third classification degree is used for indicating the theoretical classification value of the standard racket;
And training the integrated learning model according to the third image characteristics and the third classification degree of the second number of standard beats to obtain a trained integrated learning model.
6. The integrated learning model-based beat classification method of claim 5, wherein the integrated learning model comprises a first prediction layer, a second prediction layer, and a third prediction layer, wherein models included in the first prediction layer, the second prediction layer, and the third prediction layer are not identical;
training the integrated learning model according to the third image features and the third classification degree of the second number of standard beats to obtain a trained integrated learning model, including:
dividing the second number of standard beats into a first training set and a first verification set, wherein the first training set comprises image features and third classification degrees of X standard beats, and the first verification set comprises image features and third classification degrees of Y standard beats;
training the first prediction layer according to the first training set and the first verification set to obtain a trained first prediction layer;
dividing the output result of the first prediction layer into a second training set and a second verification set, and training the second prediction layer according to the second training set and the second verification set to obtain a trained second prediction layer;
Dividing the output result of the second prediction layer into a third training set and a third verification set, and training the third prediction layer according to the third training set and the third verification set to obtain a trained third prediction layer;
and constructing the trained integrated learning model according to the trained first prediction layer, the trained second prediction layer and the trained third prediction layer.
7. The integrated learning model-based beat classification method of claim 5, wherein the formula for calculating the third classification degree of each standard beat is:
wherein C is y A third classification degree of the standard racket y, B yp The similarity of the standard racket y and other standard racket p is represented, lambda represents the selling price fluctuation parameter, and lambda is more than 0 and less than 1.
8. The integrated learning model system comprises an information block chain, wherein each information node of the information block chain correspondingly stores image information of an uploaded beat; the device comprises:
the first acquisition module is used for acquiring first image information of a target beat from a target information node, extracting a plurality of first image features of the target beat according to the first image information, wherein the target information node is any information node of the information block chain;
A first calculating module, configured to calculate a first classification degree of the target beat product according to the plurality of first image features, where the first classification degree is used to indicate an actual classification value of the target beat product;
the second calculation module is used for predicting a second classification degree of the target beat according to the plurality of first image features based on the trained integrated learning model, and the second classification degree is used for indicating the predicted classification value of the target beat;
and the classification module is used for classifying the target beats according to the first classification degree and the second classification degree.
9. An AI classification system comprising a memory and a processor, the memory storing a computer program executable on the processor, wherein the processor, when executing the computer program, performs the steps of the integrated learning model-based beat classification method of any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the integrated learning model-based beat classification method of any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310664036.9A CN116580249B (en) | 2023-06-06 | 2023-06-06 | Method, system and storage medium for classifying beats based on ensemble learning model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310664036.9A CN116580249B (en) | 2023-06-06 | 2023-06-06 | Method, system and storage medium for classifying beats based on ensemble learning model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116580249A true CN116580249A (en) | 2023-08-11 |
CN116580249B CN116580249B (en) | 2024-02-20 |
Family
ID=87539688
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310664036.9A Active CN116580249B (en) | 2023-06-06 | 2023-06-06 | Method, system and storage medium for classifying beats based on ensemble learning model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116580249B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117113199A (en) * | 2023-10-23 | 2023-11-24 | 浙江星汉信息技术股份有限公司 | File security management system and method based on artificial intelligence |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103984914A (en) * | 2013-02-07 | 2014-08-13 | 东芝泰格有限公司 | Information processing apparatus and control method thereof |
CN109993552A (en) * | 2017-12-29 | 2019-07-09 | 北京奇虎科技有限公司 | A kind of PMP price acquisition methods and device |
CN111931852A (en) * | 2020-08-12 | 2020-11-13 | 中国工商银行股份有限公司 | Target object pricing method and device |
CN112685589A (en) * | 2020-12-24 | 2021-04-20 | 深圳市对庄科技有限公司 | Jadeite hierarchical valuation assessment method, system and application based on artificial intelligence |
CN113344645A (en) * | 2021-07-07 | 2021-09-03 | 中国工商银行股份有限公司 | House price prediction method and device and electronic equipment |
CN113379432A (en) * | 2021-06-30 | 2021-09-10 | 焦点科技股份有限公司 | Sales system customer matching method based on machine learning |
CN114580533A (en) * | 2022-03-04 | 2022-06-03 | 腾讯科技(深圳)有限公司 | Method, apparatus, device, medium, and program product for training feature extraction model |
CN114663167A (en) * | 2022-04-25 | 2022-06-24 | 未鲲(上海)科技服务有限公司 | Evaluation method and device based on shooting information, electronic equipment and storage medium |
CN115147660A (en) * | 2022-07-25 | 2022-10-04 | 平安科技(深圳)有限公司 | Image classification method, device and equipment based on incremental learning and storage medium |
CN115641179A (en) * | 2021-07-19 | 2023-01-24 | 中移动金融科技有限公司 | Information pushing method and device and electronic equipment |
US20230030431A1 (en) * | 2021-04-13 | 2023-02-02 | Beijing Baidu Netcom Science Technology Co., Ltd. | Method and apparatus for extracting feature, device, and storage medium |
-
2023
- 2023-06-06 CN CN202310664036.9A patent/CN116580249B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103984914A (en) * | 2013-02-07 | 2014-08-13 | 东芝泰格有限公司 | Information processing apparatus and control method thereof |
CN109993552A (en) * | 2017-12-29 | 2019-07-09 | 北京奇虎科技有限公司 | A kind of PMP price acquisition methods and device |
CN111931852A (en) * | 2020-08-12 | 2020-11-13 | 中国工商银行股份有限公司 | Target object pricing method and device |
CN112685589A (en) * | 2020-12-24 | 2021-04-20 | 深圳市对庄科技有限公司 | Jadeite hierarchical valuation assessment method, system and application based on artificial intelligence |
US20230030431A1 (en) * | 2021-04-13 | 2023-02-02 | Beijing Baidu Netcom Science Technology Co., Ltd. | Method and apparatus for extracting feature, device, and storage medium |
CN113379432A (en) * | 2021-06-30 | 2021-09-10 | 焦点科技股份有限公司 | Sales system customer matching method based on machine learning |
CN113344645A (en) * | 2021-07-07 | 2021-09-03 | 中国工商银行股份有限公司 | House price prediction method and device and electronic equipment |
CN115641179A (en) * | 2021-07-19 | 2023-01-24 | 中移动金融科技有限公司 | Information pushing method and device and electronic equipment |
CN114580533A (en) * | 2022-03-04 | 2022-06-03 | 腾讯科技(深圳)有限公司 | Method, apparatus, device, medium, and program product for training feature extraction model |
CN114663167A (en) * | 2022-04-25 | 2022-06-24 | 未鲲(上海)科技服务有限公司 | Evaluation method and device based on shooting information, electronic equipment and storage medium |
CN115147660A (en) * | 2022-07-25 | 2022-10-04 | 平安科技(深圳)有限公司 | Image classification method, device and equipment based on incremental learning and storage medium |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117113199A (en) * | 2023-10-23 | 2023-11-24 | 浙江星汉信息技术股份有限公司 | File security management system and method based on artificial intelligence |
Also Published As
Publication number | Publication date |
---|---|
CN116580249B (en) | 2024-02-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3028184B1 (en) | Method and system for searching images | |
US8724902B2 (en) | Processing image data | |
JP5848833B2 (en) | Method and system for comparing images | |
CN111340144B (en) | Risk sample detection method and device, electronic equipment and storage medium | |
WO2020238229A1 (en) | Transaction feature generation model training method and devices, and transaction feature generation method and devices | |
CN116580249B (en) | Method, system and storage medium for classifying beats based on ensemble learning model | |
CN110750658B (en) | Recommendation method of media resource, server and computer readable storage medium | |
CN111666275B (en) | Data processing method and device, electronic equipment and storage medium | |
JP2015504215A5 (en) | ||
CN111340226B (en) | Training and testing method, device and equipment for quantitative neural network model | |
CN108734587A (en) | The recommendation method and terminal device of financial product | |
CN112241494A (en) | Key information pushing method and device based on user behavior data | |
CN111401339A (en) | Method and device for identifying age of person in face image and electronic equipment | |
He et al. | Pointinst3d: Segmenting 3d instances by points | |
CN115563568A (en) | Abnormal data detection method and device, electronic device and storage medium | |
CN111475541A (en) | Data decision method and device, electronic equipment and storage medium | |
CN111401989A (en) | Tool management method, system, medium and intelligent device based on machine learning | |
CN114663167A (en) | Evaluation method and device based on shooting information, electronic equipment and storage medium | |
Zhang et al. | Fast algorithm for 2D fragment assembly based on partial EMD | |
CN112348590A (en) | Method and device for determining value of article, electronic equipment and storage medium | |
TWM561277U (en) | Computing apparatus for processing price images of a financial instrument | |
CN116012086A (en) | Commodity price estimating method, commodity price estimating device, electronic equipment and storage medium | |
CN113743533A (en) | Picture clustering method and device and storage medium | |
WO2023173801A1 (en) | Blockchain digital resource processing method and apparatus, and computer device and storage medium | |
CN113763080A (en) | Method and device for determining recommended article, electronic equipment and storage 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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |