CN110942180B - Industrial design matching service side prediction method based on xgboost algorithm - Google Patents
Industrial design matching service side prediction method based on xgboost algorithm Download PDFInfo
- Publication number
- CN110942180B CN110942180B CN201911103850.3A CN201911103850A CN110942180B CN 110942180 B CN110942180 B CN 110942180B CN 201911103850 A CN201911103850 A CN 201911103850A CN 110942180 B CN110942180 B CN 110942180B
- Authority
- CN
- China
- Prior art keywords
- prediction model
- prediction
- matching
- data
- industrial design
- 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.)
- Active
Links
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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q10/067—Enterprise or organisation modelling
-
- 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
Abstract
The invention relates to an industrial design matching service side prediction method based on an xgboost algorithm, which comprises the steps of obtaining data; data processing; establishing a prediction model of an Xgboost algorithm; the predictive model of the Xgboost algorithm is checked. The industrial design matching service side prediction system predicts whether a service side is suitable for matching with the requirements of industrial design users or not through a first prediction model, and solves the problem of data distribution with extremely unbalanced data caused by selecting one service side from a plurality of service sides; and then the second prediction model is adopted to further predict the final result after matching to give an accurate prediction result, and the dual-model prediction model framework of the industrial design matching service side prediction system can effectively balance the feature importance to avoid overfitting, so that the prediction accuracy is improved, the manual collection of data or the visit of an industrial design service side is not required, and the comparison and selection are performed, so that the time and the labor are consumed.
Description
Technical Field
The invention relates to the technical field of industrial design platforms, in particular to an industrial design matching service side prediction method based on an xgboost algorithm.
Background
The industrial design refers to the design of hardware products and comprises the contents of mechanical design, modeling design, circuit design, packaging design, mold design, hand board design and the like.
At present, most of production enterprises need to find a proper third-party industrial design service party to design new products for the new products, but China is a huge production enterprise for large production, but the actual industrial design enterprises are few, unbalance of the two is not communicated with information, and the selection of the production enterprises is more limited, so that the production enterprises have few innovation and do not break through.
The general enterprises need to carry out industrial design products to select different industrial design service parties, and the existing enterprises or the demand parties select the service parties of the industrial design through human judgment, and the judgment mode has the following problems:
1. it is difficult to judge the design ability of the service side, the industrial design is generic to the design aesthetics, but the design level of the service side for different products is different, and the difference is influenced by experience, originality and product research level, and is generally difficult to judge before the service side does not design a specific product, and the general requirement side of the industrial design can only be judged by relying on the background or past content demonstration of the service side as a reference, wherein the artificial communication influence factor is large.
2. The service side has small selection range and is limited by information and space, the service side contacted by the demand side only covers a limited number of service sides in the nearby area, and needs to be in butt joint communication with each service side, so that the time and the labor are consumed.
Therefore, the industrial design matching service party prediction method based on the xgboost algorithm is provided, and a demand party can quickly select the most suitable service party from a large number of service parties according to specific industrial design demands.
Disclosure of Invention
The invention provides an xgboost algorithm-based industrial design matching service party prediction method, which is characterized in that a double-model framework is formed by a first prediction model and a second prediction model based on the xgboost algorithm, and the industrial design requirement party is comprehensively judged to be matched with a proper service party.
The technical scheme of the invention is as follows:
an industrial design matching service side prediction method based on an xgboost algorithm comprises the following steps:
s1, acquiring data:
acquiring history matching result data from a database, wherein data of history demand side requirements, service side and transaction matching form training data;
s2, data processing:
preprocessing the data required by the historical demand party, the service party and the transaction matching to obtain a training data set containing a plurality of preprocessed data sets;
s3, establishing a prediction model of an Xgboost algorithm:
the training data set is used as training data of a prediction model, the training data of the prediction model comprises a test set and a training set, and the training set is utilized to train the prediction model of the Xgboost algorithm, so that the prediction model of the Xgboost algorithm is obtained; the prediction model is divided into a first prediction model and a second prediction model;
s4, checking a prediction model of the Xgboost algorithm:
establishing a checking module, wherein the checking module adopts a confusion matrix, a classification error rate curve, a auc curve, a cross verification method or feature importance to detect a prediction matching result output by a prediction model of the Xgboost algorithm and judges whether the checking result is larger than or equal to a preset target accuracy;
s5, predicting and matching the type of the requirement of the demand party to design a service party:
and obtaining design data of the service side from the database, extracting characteristics of the data, and inputting the characteristics of the corresponding data into a prediction model of the Xgboost algorithm to obtain the industrial design service side matched with the requirements of the requirement side.
Preferably, the first prediction model and the second prediction model are constructed according to the training data set containing different data and combining different amounts of training data.
Preferably, the first prediction model is a dual classification model for predicting whether the industrial design service side matches the demand of the user of the demand side.
Preferably, the second prediction model is a multi-classification model for predicting the outcome of a matching industrial design service.
Preferably, the results of the predictive match industry design service include at least the probability of project success, project failure, no response to a match, an interruption after a match, a rejection of a match, and a termination of a project.
Preferably, the checking module is configured to check a prediction result output by the first prediction model or the second prediction model.
Preferably, the training data set includes at least historical runner demand data, historical transaction data, historical matching data, industrial design cases of the server, background information of the server, and background information of the runner.
Preferably, the specific step of the inspection module for inspecting the prediction result of the prediction model includes the following steps:
4a: the results output by the first prediction model and the second prediction model are respectively checked through the checking module, and whether the checking result is larger than or equal to a preset target accuracy rate is judged; if the accuracy is greater than or equal to the preset target accuracy, executing the step 4b; if the accuracy is smaller than the preset target accuracy, executing the step 4c;
4b: comprehensively checking the prediction results output by the first prediction model and the second prediction model through the checking module, outputting the results and storing the prediction models;
4c: and (2) respectively adjusting the training data of the first prediction model and the second prediction model through training an adjusting model, and executing the step (S2).
Preferably, the training tuning model is used for tuning training data in the first prediction model and the second prediction model.
Based on the industrial design matching service side prediction method, the invention also provides a prediction method for calling the prediction model matching service side, which comprises the following steps:
s01, acquiring data: acquiring demand data of a consumer industrial design user;
s02, calling a prediction model: inputting the demand data in the step S01 into the first prediction model for matching prediction, judging a service side of the matching prediction of the first prediction model, and judging whether the predicted service side meets the demand of a user of the industrial design of the demand side; if yes, executing step S03; if not, executing step S04;
s03, inputting the demand data in the step S01 into the second prediction model for prediction, and outputting a matched prediction result;
s04, directly outputting the prediction result of the first prediction model matching as non-matching.
The beneficial effects of the invention are as follows: compared with the prior art, the embodiment of the invention has the following advantages:
(1) And forming a double-model framework through the first prediction model and the second prediction model, and comprehensively judging that the industrial design user matches a proper service party.
(2) Whether the service side is suitable for matching with the requirements of the industrial design user is predicted and known through the first prediction model, so that the problem of extremely unbalanced data distribution caused by selecting one service side from a plurality of service sides is solved; and then, a second prediction model is adopted to further predict a final result after matching to give an accurate prediction result, and the dual-model prediction model framework of the industrial design matching service side prediction system can effectively balance the feature importance to avoid overfitting, so that the prediction accuracy is improved.
(3) The industrial design service side meeting the design requirement of the demand side can be predicted by the industrial design matching service side prediction method, and the problems of time consumption and labor consumption caused by manual data collection or visiting of the industrial design service side and comparison and selection are solved.
(4) The prediction model in the industrial design matching service side prediction method can learn and judge a large amount of industrial design comprehensive information, can avoid artificial judgment deviation caused by insufficient information acquisition, and is more accurate for matching the demand side with the industrial design service side, thereby providing convenience for the selection of industrial design consumers.
Description of the drawings:
fig. 1 is a frame diagram of an industrial design matching service side prediction method based on an xgboost algorithm in an embodiment of the invention.
Fig. 2 is a frame diagram of a prediction model in the industrial design matching service side prediction method based on the xgboost algorithm in the embodiment of the invention.
FIG. 3 is a flowchart illustrating a step of checking a prediction result of a prediction model by using the xgboost algorithm-based industrial design matching service side prediction method according to an embodiment of the present invention.
Fig. 4 is a flowchart illustrating a method for calling a prediction model to match a server according to an embodiment of the present invention.
Detailed Description
In order to make the technical scheme and technical effects of the invention more clear, the invention is further described below with reference to specific embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the embodiments of the present invention, the meaning of "plurality" is two or more, unless explicitly defined otherwise.
The industrial design matching service side prediction method based on the xgboost algorithm provided by the embodiment of the invention can be applied to service platforms of service type industries such as industrial design industry, intellectual property industry, insurance industry, real estate industry and the like. The invention will now be further explained by way of example with reference to the accompanying drawings in which a suitable industrial design service party is adapted to the design requirements for the requirements party.
Embodiment one:
as shown in fig. 1 and fig. 2, an embodiment of the present invention provides an xgboost algorithm-based industrial design matching service side prediction method, which includes the following steps:
s1, acquiring data:
acquiring history matching result data from a database, wherein data of history demand side requirements, service side and transaction matching form training data;
s2, data processing:
preprocessing the data required by the historical demand party, the service party and the transaction matching to obtain a training data set containing a plurality of preprocessed data sets;
s3, establishing a prediction model of an Xgboost algorithm:
the training data set is used as training data of a prediction model, the training data of the prediction model comprises a test set and a training set, and the training set is utilized to train the prediction model of the Xgboost algorithm, so that the prediction model of the Xgboost algorithm is obtained;
s4, checking a prediction model of the Xgboost algorithm:
establishing a checking module, wherein the checking module adopts a confusion matrix, a classification error rate curve, a auc curve, a cross verification method or feature importance to detect a prediction matching result output by a prediction model of the Xgboost algorithm and judges whether the checking result is larger than or equal to a preset target accuracy; the prediction model is divided into a first prediction model and a second prediction model;
s5, predicting and matching the type of the requirement of the demand party to design a service party:
and obtaining design data of the service side from the database, extracting characteristics of the data, and inputting the characteristics of the corresponding data into a prediction model of the Xgboost algorithm to obtain the industrial design service side matched with the requirements of the requirement side.
Optionally, in step S1 of the embodiment of the present application, the main source of the historical data in the database is a website, the data on the website is self-owned data accumulated through the historical service, and the data mainly includes historical demand data, historical transaction data, historical matching data, industrial design cases of the service side, background information of the demand side, and the like.
Optionally, in step S2 of the embodiment of the present application, the data processing mainly adopts a preprocessing mode, where the preprocessing mode includes a missing value complementary 0, an abnormal data removing and a single-hot method 0-1 digitizing. Specifically, the missing value 0-supplementing mode is used for supplementing the missing value of the data in the training data set with 0. The abnormal data eliminating mode is to eliminate the data below or above the limit in the training data set (the data parameters are quantized to be in the range of 0-2 and the values above or below 0). The 0-1 numerical mode of the independent heating method is an extraction processing mode of partial data features (such as product style, product attribute, control mode, main material, processing technology and the like of a design scheme) in the training data set, for example, the attribute values are subjected to 0-1 numerical mode by using the independent heating method according to information of industrial design cases of a demand information and a service side, dimension calculation of each feature and demand similarity of the demand side are constructed into feature dimensions, so that a feature dimension set is formed, and the feature dimension set is combined with the training data set to obtain a training data set of the prediction model.
Optionally, in the step S2 of the embodiment of the present application, the preprocessing mode used may further use average value and non-null term counting for the numerical class features in the training data set, and also perform interval day calculation and average interval day calculation for the time type data, so as to obtain the extracted feature data. In this embodiment, the feature extraction unit mainly performs mean value and non-null item count on the numerical value class features in the original data in the database to form new features, performs interval number calculation and average interval number calculation on the time type data to form new features, and implements preprocessing on the data.
It should be noted that, the description is given by taking the historical demand data of the demand party or the industrial design case of the service party as the first embodiment.
The content of the demand party demand or the industrial design case is composed of text labels, wherein the labels comprise a product style, the attribute of the product, a control mode, main materials, a processing technology and the like, for example, the text labels of the product style comprise Japanese Korean, european, american, chinese and middle east, the text labels of the areas are coded into 5 dimensions by using single heat, the labels are marked as 1, and the labels are marked as 0; the text labels are respectively classified into product design information and market background information, and the label dimensions of the product design information and the market background information are respectively and independently formed into vector matrixes; and then, calculating the similarity between each requirement and the industrial design case of the service side by adopting a cosine similarity calculation method according to the product design information and the market background information. The cosine similarity is evaluated by calculating the cosine value of the included angle of the two vectors, and the cosine value between the two vectors can be obtained by using the Euclidean dot product formula: a.b= | a| b cos θ.
a and b are two attribute vectors, the rest chord similarity theta is given by dot product and vector length, and the calculation formula is:
the similarity range given by the cosine value between the two vectors is-1; wherein, -1 means that the directions of the two vectors are exactly opposite, 1 means that the directions are exactly the same, 0 means that the vectors are independent, and a value between-1 and 0 and 1 means that the similarity features between the industrial design cases of the demander and the service side are similar, and the similarity calculates corresponding product design information and market background information according to the class respectively, so that more detail information can be provided to help promote the accuracy of the prediction result of the prediction model.
Historical transaction data is described as embodiment two.
The historical transaction data mainly comprises objects of two parties of a transaction, an amount and transaction time. The data processing module takes the characteristic of the expansion of the transaction time in the historical transaction data as the average time interval between the current time interval and the transaction, takes the characteristic of the expansion of the amount in the historical transaction data as the sum of the average transaction amount and the transaction amount, and provides more information than the original characteristic.
The history matching data is described as embodiment three.
The history matching data mainly comprises the demand singular number of the demand party and the number of times of the deal of the demand party, the matching number of times of the service party and the number of times of the deal of the service party, and transaction data corresponding to each class is counted according to the class derivation of the industrial design class (class is the industrial design class of the platform, such as household appliances-kitchen small household appliances-electric kettles), so that more information is provided than the original characteristics.
The description will be made with the background information of the service side as the fourth embodiment.
The background information of the service side mainly comprises the specialty class of the service side (the product class of which the service is anti-adept in design), the establishment time of the service side, the number of companies of the service side, the number of prizes obtained by the service side, the transaction amount of the service side in exchange, the transaction number of the service side in exchange, the service score of the service side in exchange, whether the service side enters a residence platform, the company address of the service side and the like. The data processing module constructs the above background information into more features according to the number, the score, the number and the like of each category statistics, and provides more detail information.
The description will be made with the background information of the demand side as embodiment five.
The background information of the demand party mainly comprises the enterprise situation of the demand party, the enterprise type of the demand party, the number of companies of the demand party, the establishment time of the demand party, the total amount of the demand party, the average amount of the demand party, and the like, and the background information processing data of the demand party is obtained by processing the independent heat codes of the enterprise type, the enterprise situation and the number of companies.
The training data set of the present application includes at least historical runner demand data, historical transaction data, historical matching data, industrial design cases of the runner, background information of the runner, and background information of the runner.
As shown in fig. 2, fig. 2 is a frame diagram of a prediction model in an xgboost algorithm-based industrial design matching service side prediction method according to an embodiment of the present invention. The first prediction model and the second prediction model are built according to the test data packet containing different data and combining different numbers of training data sets.
It should be noted that, the first prediction model may be formed by acquiring training data sets of 100 project cases from a database, and the second prediction model may be formed by acquiring training data sets of 200 project cases or more than 100 project cases from the database; or dividing different training data according to the training results of the training data set, and dividing the training data into the first prediction model and the second prediction model according to the different training data.
The first prediction model is a double-classification model and is used for predicting whether an industrial design service side matches the requirement of a user of a requirement side or not, and the output result of the first prediction model is the probability of matching and unmatched. Specifically, the first prediction model predicts whether to match the two classes of the service side, and the second prediction model predicts that the matching service side is successful, and the rest is unsuccessful data because of one successful matching each time, so that the training data of the first prediction model contains a plurality of unsuccessful data of the matching service side.
The second prediction model is a multi-classification model and is used for predicting and matching the result of the industrial design service side, and the second prediction model outputs the result of predicting and matching the industrial design service side and comprises the probability of project success, project failure, no response of matching, interruption after matching, refusal of matching and project termination.
The inspection module is used for inspecting the prediction results output by the first prediction model or the second prediction model. Specifically, the inspection module inspects the result output by the first prediction model or the second prediction model by adopting a machine-learned confusion matrix, a classification error rate curve, a auc curve, a learning curve of a k-fold cross validation method, a feature importance and other methods, and outputs the probability of matching each item.
Compared with the prior art, the industrial design matching service side prediction method based on the xgboost algorithm comprehensively judges that an industrial design user matches a proper service side through a double-model framework formed by the first prediction model and the second prediction model; whether the service side is suitable for matching with the requirements of the industrial design user is predicted and known through the first prediction model, and the problem of data distribution with extremely unbalanced data caused by selecting one service side from a plurality of service sides is solved.
Example two
As shown in fig. 3, fig. 3 is a flowchart illustrating a step of checking a prediction result of a prediction model according to an embodiment of the present invention. The industrial design matching service side prediction method based on the xgboost algorithm provided by the embodiment of the application, wherein the specific steps of the detection module for detecting the prediction result of the prediction model include:
4a: the results output by the first prediction model and the second prediction model are respectively checked through the checking module, and whether the checking result is larger than or equal to a preset target accuracy rate is judged; if the accuracy is greater than or equal to the preset target accuracy, executing the step 4b; if the accuracy is smaller than the preset target accuracy, executing the step 4c;
4b: comprehensively checking the prediction results output by the first prediction model and the second prediction model through the checking module, outputting the results and storing the prediction models;
4c: and (2) respectively adjusting the training data of the first prediction model and the second prediction model through training an adjusting model, and executing the step (S2).
It should be noted that, the target of the preset target accuracy rate matching is 95%. The industrial design is enabled to be matched with the result output by the prediction model of the service side prediction method with high accuracy through training and checking the first prediction model and the second prediction model in a circulating mode.
Optionally, the training tuning model is configured to perform data tuning on the first prediction model by adopting a network search configuration mode according to the test result. Specifically, the first prediction model can adopt a network search configuration L1 regularization and L2 regularization mode to perform coarse classification on a pair of data, so that the ratio of sub-tree resampling to node resampling is reduced on the premise of maximum depth of characteristic parameters in the first prediction model, and the first prediction model can be accurately classified; the method and the device avoid dimension reduction of the data (the dimension reduction mode of the data parameters can reduce the information quantity of the training data set to cause that the transition depends on the characteristic with large information quantity), and also avoid the occurrence of the conditions of over fitting and obvious deviation of the undersampling or oversampling processing mode of adopting characteristic sampling to the data, which result in low matching accuracy.
Optionally, the training tuning model adjusts the feature parameters of the second prediction model by adding cases or expanding feature data. Specifically, historical transaction data is added to the test data packet in the second prediction model to expand the data in the training data set, so that the accuracy of the predicted output result is improved. In this embodiment, the parameters are adjusted by network searching, so that the second prediction model fully considers factors such as cases, duration, specialty and the like, and more accurately predicts the probability result of the matching service side.
It should be noted that, since each requirement is directed to a separate product, that is, a unique and suitable service party needs to be matched from a large number of service parties, an extremely unbalanced training data set is generated, and the training data set is large in volume. Therefore, the design of the training tuning model improves the accuracy of the output results of the first prediction model and the second prediction model, and improves the efficiency and accuracy of the matching of the first prediction model for the first prediction model.
As shown in fig. 4, an embodiment of the present application provides a prediction method for calling a prediction model to match a service side, including the following steps:
s01, acquiring data: acquiring demand data of a consumer industrial design user;
s02, calling a prediction model: inputting the demand data in the step S01 into the first prediction model for matching prediction, judging a service side of the matching prediction of the first prediction model, and judging whether the predicted service side meets the demand of a user of the industrial design of the demand side; if yes, executing step S03; if not, executing step S04;
s03, inputting the demand data in the step S01 into the second prediction model for prediction, and outputting a matched prediction result;
s04, directly outputting the prediction result of the first prediction model matching as non-matching.
Acquiring data of an industrial design demand party on the industrial design matching service party prediction method based on the xgboost algorithm, and predicting on the prediction module to obtain the following result:
TABLE 1
TABLE 2
Wherein, table 1 is the result output by the first prediction model, and table 2 is the result output by the second prediction model.
As can be seen from the data in tables 1 and 2, the industrial design demander is predicted according to the industrial design matching service side prediction method based on the xgboost algorithm, the a service side and the B service side output by the first prediction model meet the requirements of the industrial design demander, the C service side does not meet the requirements of the industrial design demander, and the second prediction model tests the a service side and the B service side for the requirements according to the industrial design demander, so that the a service side meets the industrial design demander.
Compared with the prior art, the prediction method for calling the prediction model to match the service side provided by the embodiment of the application gives an accurate prediction result through the second prediction model and further predicting the final result after matching.
The foregoing is a further detailed description of the invention in connection with the preferred embodiments, and it is not intended that the invention be limited to the specific embodiments described. For those skilled in the art, the architecture of the invention can be flexible and changeable without departing from the concept of the invention, and serial products can be derived. But a few simple derivatives or substitutions should be construed as falling within the scope of the invention as defined by the appended claims.
Claims (7)
1. An industrial design matching service side prediction method based on an xgboost algorithm is characterized by comprising the following steps of:
s1, acquiring data:
acquiring history matching result data from a database, wherein data of history demand side requirements, service side and transaction matching form training data;
s2, data processing:
preprocessing the data required by the historical demand party, the service party and the transaction matching to obtain a training data set containing a plurality of preprocessed data sets;
s3, establishing a prediction model of an Xgboost algorithm:
the training data set is used as training data of a prediction model, the training data of the prediction model comprises a test set and a training set, and the training set is utilized to train the prediction model of the Xgboost algorithm, so that the prediction model of the Xgboost algorithm is obtained; the prediction model is divided into a first prediction model and a second prediction model;
s4, checking a prediction model of the Xgboost algorithm:
establishing a checking module, wherein the checking module adopts a confusion matrix, a classification error rate curve, a auc curve, a cross verification method or feature importance to detect a prediction matching result output by a prediction model of the Xgboost algorithm and judges whether the checking result is larger than or equal to a preset target accuracy;
s5, predicting and matching the type of the requirement of the demand party to design a service party:
acquiring design data of a service party from a database, extracting characteristics of the data, and inputting the characteristics of the corresponding data into a prediction model of the Xgboost algorithm to obtain an industrial design service party matched with the requirements of a demand party;
the first prediction model is a double-classification model and is used for predicting whether an industrial design service side matches the requirement of a user of a requirement side; the second prediction model is a multi-classification model and is used for predicting and matching the result of the industrial design service side; the specific steps of the inspection module for inspecting the prediction result of the prediction model include:
4a: the results output by the first prediction model and the second prediction model are respectively checked through the checking module, and whether the checking result is larger than or equal to a preset target accuracy rate is judged; if the accuracy is greater than or equal to the preset target accuracy, executing the step 4b; if the accuracy is smaller than the preset target accuracy, executing the step 4c;
4b: comprehensively checking the prediction results output by the first prediction model and the second prediction model through the checking module, outputting the results and storing the prediction models;
4c: and (2) respectively adjusting the training data of the first prediction model and the second prediction model through training an adjusting model, and executing the step (S2).
2. The xgboost algorithm-based industrial design matching service side prediction method according to claim 1, wherein the first prediction model and the second prediction model are built according to the training data set containing different data and combining different amounts of training data.
3. The xgboost algorithm-based industrial design matching service side prediction method according to claim 1, wherein the result of predicting matching an industrial design service side at least includes the probability of project success, project failure, no response to matching, interruption after matching, refusal of matching, and project termination.
4. The xgboost algorithm-based industrial design matching service side prediction method according to claim 2, wherein the checking module is configured to check a prediction result output by the first prediction model or the second prediction model.
5. The xgboost algorithm-based industrial design matching service party prediction method according to claim 1, wherein the training data set includes at least historical demand party demand data, historical transaction data, historical matching data, industrial design cases of the service party, background information of the service party, and background information of the demand party.
6. The xgboost algorithm-based industrial design matching service side prediction method according to claim 1, wherein the training tuning model is used for tuning training data in the first prediction model and the second prediction model.
7. An industrial design matching service side prediction method based on an xgboost algorithm according to any one of claims 1 to 6, characterized in that the prediction method of calling the prediction model matching service side comprises the following steps:
s01, acquiring data: acquiring demand data of a consumer industrial design user;
s02, calling a prediction model: inputting the demand data in the step S01 into the first prediction model for matching prediction, judging a service side of the matching prediction of the first prediction model, and judging whether the predicted service side meets the demand of a user of the industrial design of the demand side; if yes, executing step S03; if not, executing step S04;
s03, inputting the demand data in the step S01 into the second prediction model for prediction, and outputting a matched prediction result;
s04, directly outputting the prediction result of the first prediction model matching as non-matching.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911103850.3A CN110942180B (en) | 2019-11-12 | 2019-11-12 | Industrial design matching service side prediction method based on xgboost algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911103850.3A CN110942180B (en) | 2019-11-12 | 2019-11-12 | Industrial design matching service side prediction method based on xgboost algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110942180A CN110942180A (en) | 2020-03-31 |
| CN110942180B true CN110942180B (en) | 2023-07-04 |
Family
ID=69906703
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911103850.3A Active CN110942180B (en) | 2019-11-12 | 2019-11-12 | Industrial design matching service side prediction method based on xgboost algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110942180B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114049505B (en) * | 2021-10-11 | 2022-08-23 | 数采小博科技发展有限公司 | Method, device, equipment and medium for matching and identifying commodities |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109145031A (en) * | 2018-08-20 | 2019-01-04 | 国网安徽省电力有限公司合肥供电公司 | A kind of multi-source data multidimensional reconstructing method of service-oriented market access demand |
| CN109739844A (en) * | 2018-12-26 | 2019-05-10 | 西安电子科技大学 | Data classification method based on decaying weight |
| CN110111886A (en) * | 2019-05-16 | 2019-08-09 | 闻康集团股份有限公司 | A kind of intelligent interrogation system and method based on XGBoost disease forecasting |
| CN110378739A (en) * | 2019-07-23 | 2019-10-25 | 中国联合网络通信集团有限公司 | A kind of data traffic matching process and device |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104765642B (en) * | 2015-03-24 | 2017-11-10 | 长沙理工大学 | Virtual machine deployment method and system based on dynamic prediction model under cloud environment |
| CN106022599A (en) * | 2016-05-18 | 2016-10-12 | 德稻全球创新网络（北京）有限公司 | Industrial design talent level evaluation method and system |
| CN106951633A (en) * | 2017-03-02 | 2017-07-14 | 郑州轻工业学院 | Industrial design service platform |
| CN107274010A (en) * | 2017-06-01 | 2017-10-20 | 西北工业大学 | Towards the multiple target creative design evaluation method of industrial design cloud service platform |
| CN109829105A (en) * | 2019-01-18 | 2019-05-31 | 王惟惟 | It is a kind of can application block chain technology artificial intelligence create auxiliary system |
-
2019
- 2019-11-12 CN CN201911103850.3A patent/CN110942180B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109145031A (en) * | 2018-08-20 | 2019-01-04 | 国网安徽省电力有限公司合肥供电公司 | A kind of multi-source data multidimensional reconstructing method of service-oriented market access demand |
| CN109739844A (en) * | 2018-12-26 | 2019-05-10 | 西安电子科技大学 | Data classification method based on decaying weight |
| CN110111886A (en) * | 2019-05-16 | 2019-08-09 | 闻康集团股份有限公司 | A kind of intelligent interrogation system and method based on XGBoost disease forecasting |
| CN110378739A (en) * | 2019-07-23 | 2019-10-25 | 中国联合网络通信集团有限公司 | A kind of data traffic matching process and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110942180A (en) | 2020-03-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106296195A (en) | A kind of Risk Identification Method and device | |
| CN113011973B (en) | Method and equipment for financial transaction supervision model based on intelligent contract data lake | |
| CN109598517B (en) | Commodity clearance processing, object processing and category prediction method and device thereof | |
| CN105069072A (en) | Emotional analysis based mixed user scoring information recommendation method and apparatus | |
| CN109376766B (en) | Portrait prediction classification method, device and equipment | |
| CN111461164B (en) | Sample data set capacity expansion method and model training method | |
| CN109388697A (en) | Handle the method, apparatus and electric terminal of question and answer corpus | |
| CN103955714A (en) | Navy detection model construction method and system and navy detection method | |
| CN110134777A (en) | Problem De-weight method, device, electronic equipment and computer readable storage medium | |
| CN110727855A (en) | Personalized recommendation method based on improved factorization machine | |
| CN110009045A (en) | The recognition methods of internet-of-things terminal and device | |
| CN110929764A (en) | Picture auditing method and device, electronic equipment and storage medium | |
| CN114358657A (en) | Post recommendation method and device based on model fusion | |
| CN109886206B (en) | Three-dimensional object identification method and equipment | |
| CN114861050A (en) | Feature fusion recommendation method and system based on neural network | |
| CN110942180B (en) | Industrial design matching service side prediction method based on xgboost algorithm | |
| EP4273750A1 (en) | Data processing method and apparatus, computing device, and test simplification device | |
| CN106952060A (en) | Logistic Scheduling method and device | |
| CN116340643B (en) | Object recommendation adjustment method and device, storage medium and electronic equipment | |
| CN111143533A (en) | Customer service method and system based on user behavior data | |
| CN111667339B (en) | Defamation malicious user detection method based on improved recurrent neural network | |
| CN115082165A (en) | Product innovation design scheme recommendation method and system based on similarity relation recognition | |
| CN111651512A (en) | Multisource heterogeneous commodity feature weight solving method and device based on semi-supervised learning | |
| CN116977021B (en) | Automatic pushing method for system butt joint based on big data | |
| CN116821512B (en) | Recommendation model training method and device, recommendation method and device |
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 |