CN110659937A - Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm - Google Patents
Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm Download PDFInfo
- Publication number
- CN110659937A CN110659937A CN201910889785.5A CN201910889785A CN110659937A CN 110659937 A CN110659937 A CN 110659937A CN 201910889785 A CN201910889785 A CN 201910889785A CN 110659937 A CN110659937 A CN 110659937A
- Authority
- CN
- China
- Prior art keywords
- supplier
- gradient
- tree
- bid
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 claims abstract description 33
- 238000011156 evaluation Methods 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000010276 construction Methods 0.000 claims abstract description 10
- 238000012216 screening Methods 0.000 claims abstract description 9
- 230000002159 abnormal effect Effects 0.000 claims abstract description 4
- 238000004140 cleaning Methods 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims abstract description 4
- 230000003993 interaction Effects 0.000 claims abstract description 3
- 230000008569 process Effects 0.000 claims description 25
- 230000006872 improvement Effects 0.000 claims description 22
- 230000006399 behavior Effects 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 230000004044 response Effects 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 3
- 230000007812 deficiency Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000013138 pruning Methods 0.000 claims description 3
- 238000013139 quantization Methods 0.000 claims 6
- 230000000694 effects Effects 0.000 abstract description 3
- 238000007418 data mining Methods 0.000 abstract description 2
- 230000002452 interceptive effect Effects 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0282—Rating or review of business operators or products
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of prediction methods in data mining, and particularly relates to a gradient-boosted tree-based improved supplier quantitative scoring prediction algorithm, which is characterized by comprising the following steps of: s1: the data module is used for acquiring data required by quantitative grading of a supplier, and performing cleaning, abnormal value processing and missing value processing; s2: the feature engineering module carries out feature construction and feature intelligent screening; s3: the training module comprises a weak learner, a regularization function and gradient lifting training; s4: the prediction module uses the trained model to carry out testing; s5: the evaluation module comprises a service evaluation module and an algorithm evaluation module; s6: the online application module quantitatively scores the suppliers by using the trained model. The method utilizes the gradient lifting tree to intelligently screen the features, calculates the weight for the features in a nonlinear mode, saves the cost of judging the importance of the features by adopting a salesman with business experience, takes the interaction of the features into consideration in the algorithm, and has obvious effect.
Description
Technical Field
The invention belongs to the technical field of prediction methods in data mining, and particularly relates to an improved supplier quantitative scoring prediction algorithm based on a gradient lifting tree.
Background
At present, in a bidding link of a mining enterprise, the mining enterprise searches related information from a network by the personal experience of a buyer or subjectively selects a provider due to interest relations when searching for the provider for the first time. Therefore, an inappropriate supplier is often selected, and the process of selecting the supplier takes a long time, which increases the cost of the enterprise. In addition, because the enterprise purchases various materials daily and purchases frequently, the method is not suitable for the bidding process which is long in time consumption and high in cost. In addition, mining enterprises tend to pay more attention to quality problems when purchasing professional equipment, and therefore tend to be more prone to suppliers with over-experience and high satisfaction. And when purchasing daily life needs of enterprises, more emphasis is placed on price factors.
Based on the background, it is a key problem to be solved urgently how to refine the basic requirements of the enterprise and to judge various factors influencing the enterprise benefits to collect specific indexes of different types of product evaluations of the suppliers, how to perform standardized processing on the indexes and give proper weight to the indexes, and further to perform quantitative analysis and accurate evaluation on the suppliers by adopting a scientific calculation method, and to select the suppliers with higher values.
Disclosure of Invention
The invention aims to provide a gradient lifting tree improved supplier quantitative scoring prediction algorithm, which is characterized in that a gradient lifting tree is constructed through the learning of sample input data, a model is trained, and the error between a predicted value and an actual value is calculated, so that a supplier meeting requirements is selected.
The purpose of the invention is realized by the following technical scheme:
the invention discloses a gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm, which is characterized by comprising the following steps of:
s1: the data module is used for acquiring data required by quantitative grading of a supplier and performing some cleaning, abnormal value processing and missing value processing;
s2: the characteristic engineering module comprises a characteristic construction module and a characteristic intelligent screening module;
s3: the training module comprises a weak learner, a regularization function and gradient lifting training;
s4: the prediction module uses the trained model to carry out testing;
s5: the evaluation module comprises a service evaluation module and an algorithm evaluation module;
s6: the online application module quantitatively scores the suppliers by using the trained model.
The specific process of the step S1 is as follows:
s11: data acquisition: the data is derived from a supplier attribute table and a supplier bid process behavior table;
s12: and (3) label construction: in the historical data, a supplier with a bid-winning behavior marks 1, and a supplier without the bid-winning behavior marks 0;
s13: data cleansing and outlier processing.
S14: and (5) processing missing values.
The specific process of the step S2 is as follows:
s21: the feature construction comprises two types, wherein the first type is the feature of the supplier and the second type is the behavior feature of the interaction between the supplier and the tenderer;
is characterized by comprising the following steps: supplier service type, proportion of bid frequency to main service type, total number of bids of main service type, bid frequency, bid-winning frequency, ratio of bid-winning amount to service type, bid-winning rate, proportion of bid-winning frequency to main service type, total amount of bid-winning, interest-offering degree, effective bid rate, response frequency, invitation frequency, response rate and overdue frequency of bid-winning service fee.
S22: intelligent screening of characteristics: finding out the locally optimal characteristics by using an information gain algorithm, wherein the information gain algorithm is as follows:
g(D,A)=H(D)-H(DIA)
the specific process of the step S3 is as follows:
s31: constructing a regression tree: using regression tree generation and regression tree pruning to select the optimal feature under the current condition as a division rule, namely the locally optimal feature;
s32: gradient iteration: weak learners (regression trees) are used to make up for the deficiencies of all trees that have been previously built; the residual error of the gradient lifting tree established by the gradient lifting improvement algorithm uses the mean square error: l (Y, f (x)) ═ Y-f (x))2The optimization goal of the gradient lifting improvement algorithm is to minimize the loss function J ═ ΣiL(Yi-F(Xi));
S33: overfitting was prevented using the Shrinkage algorithm, which is as follows:
the specific process of the step S4 is as follows:
and predicting the supplier bid score through the trained model and the new characteristic, wherein in order to distribute the score output by the gradient promotion improvement algorithm between (0, 1), the prediction score normalization formula is as follows:
the specific process of the step S5 is as follows:
the evaluation index uses AUC, and the feature label combination of the gradient lifting improvement algorithm and the inherent parameters of the gradient lifting improvement algorithm are debugged by observing the change of the AUC in the training set; the gradient boosting tree parameters include: learning rate, weak classifier number, sample subset proportion, branch minimum sample size, leaf node minimum sample size, tree maximum depth, maximum leaf node size, tree classification number, maximum feature subset size, loss function, and random seed parameter.
The specific process of the step S6 is as follows:
the gradient promotion improvement algorithm is used for integrating a plurality of characteristic factors for each supplier bid by the enterprise in the past, and calculating scores for sorting.
The invention has the advantages that:
the improved supplier quantitative scoring prediction algorithm based on the gradient lifting tree avoids the condition that a tenderer subjectively selects suppliers according to own business experience, and has overlarge limitation; when the gradient lifting tree improvement algorithm calculates the quality score of a supplier, the residual error of the former tree is learned, and the residual error is closer to the real quality of the supplier, and the automatic mode of machine learning has great significance for establishing an intelligent model system of the bid and bid technology in the promotion of the acquisition and marketing business.
Drawings
FIG. 1 is a flow chart of supplier quantitative score prediction according to the present invention.
FIG. 2 is a diagram of a missing data value processing process according to the present invention.
FIG. 3 is a flow chart illustrating a supplier forecasting method according to the present invention.
Detailed Description
The following further describes the embodiments of the present invention with reference to the drawings.
As shown in fig. 1, the gradient-boosted tree-based improved supplier quantitative score prediction method of the present invention includes a data module, a feature engineering module, a training module, a prediction module, an evaluation module, and an online application module, and includes the following steps:
s1: the data module is used for acquiring data in the supplier registration and bidding processes for quantitative grading of suppliers and performing some cleaning, abnormal value processing and missing value processing;
s11: data acquisition: the data is derived from the data of the supplier registration process and the bidding process, one is a supplier attribute table, and the other is a supplier bidding process behavior table;
s12: and (3) label construction: in the historical data, a supplier with a bid-winning behavior marks 1, and a supplier without the bid-winning behavior marks 0;
s13: data cleansing and outlier processing.
S14: missing value handling is shown in figure 2.
S2: carrying out feature construction and feature intelligent screening on the acquired data;
s21: the feature construction comprises two types, wherein the first type is the feature of the supplier and the second type is the interactive feature of the supplier and the tenderer; in the present invention, the supplier quantitative score table is shown in table 1:
TABLE 1
Feature 1: a type of service of the offer;
feature 2: the proportion of the bidding times to the type of the service;
feature 3: the total number of bidding times of the business type;
feature 4: the number of bids;
feature 5: the number of winning the bid;
feature 6: the amount of the winning bid accounts for the service type ratio;
feature 7: the winning rate;
feature 8: the winning number accounts for the proportion of the service type;
feature 9: a total amount of the bid;
feature 10: the extent of the offer;
feature 11: effective bidding rate;
feature 12: the number of responses;
feature 13: the number of invitations;
feature 14: a response rate;
feature 15: the overdue times of the winning bid service fee.
S22: intelligent screening of characteristics: finding out locally optimal characteristics by using an information gain algorithm;
an internal node is used for representing a feature or attribute in a gradient lifting tree improvement algorithm, a leaf node represents a class, and the criterion of feature selection is to find out the locally optimal feature, namely, after classification is carried out according to the feature, a data set can be separated as much as possible; the invention uses the information gain algorithm to find out the local optimal characteristics, and the larger the information gain is, the more important the characteristics are;
the information gain algorithm is as follows:
g(D,A)=H(D)-H(DIA)
where the information entropy represents the expectation of the information quantity, the formula is as follows:
s3: training the optimized features by using a training module, wherein the training module comprises a weak learner, a regularization function and a gradient lifting training module;
s31: constructing a regression tree: selecting the optimal characteristic under the current condition as a division rule by using a regression tree generation and regression tree pruning method, namely determining the locally optimal characteristic;
s32: gradient iteration: the gradient lifting is the core of the gradient lifting improvement algorithm, each tree is a residual of the sum of all previous tree conclusions, and the residual is an accumulated amount of true values obtained after prediction. Gradient boosting is used here where at each step of the gradient boosting improvement algorithm, a weak learner (regression tree) is present to make up for the deficiencies of all trees that have been previously established;
the residual error of the gradient lifting tree established by the gradient lifting improvement algorithm is the mean square error: l (Y, f (x)) ═ Y-f (x))2The optimization goal of the gradient lifting improvement algorithm is to minimize the loss function J ═ ΣiL(Yi-F(Xi) ); when the minimum value is found, F (X) is calculatedi) As a parameter, the partial derivative is calculated for the loss function,so when the mean square error is taken as the loss function, the residual error can be obtained to be the negative direction of the gradientThe relationship between gradient descent and prediction can be expressed as follows:
F(Xi):=F(Xi)+h(Xi)
F(Xi):=F(Xi)+Yii-F(Xi)
establishing the next regression tree to reduce the residual error is equivalent to establishing the minimized gradient of the next decision tree, and the same is true when the minimum gradient is updated to the whole forest;
s33: using the Shrinkage method to prevent over-trained fitting;
the krinkage method is represented as follows:
wherein the factor v (0< v <1) is a weight per tree for controlling the learning rate of the boosting process;
fm-1(x) Is the predicted result of the first m trees;
S4: testing the trained model by using a prediction module;
the specific prediction process of the present invention is shown in fig. 3, and predicts the bid-attracting score of the supplier through the trained model and the new feature, and in order to distribute the score output by the gradient boost improvement algorithm between (0, 1), the normalization formula of the prediction score is as follows:
s5: optimizing the prediction effect by using an evaluation module, wherein the evaluation module comprises a service evaluation module and an algorithm evaluation module;
AUC is used for evaluating the index, and the feature label combination and parameters of the gradient lifting improvement algorithm are debugged by observing the change of the index AUC. The gradient boosting tree parameters include: learning rate, weak classifier number, sample subset proportion, branch minimum sample size, leaf node minimum sample size, tree maximum depth, maximum leaf node size, tree classification number, maximum feature subset size, loss function, and random seed parameter.
S6: an online application module: quantitatively scoring the suppliers by using the trained models;
the gradient promotion improvement algorithm is used for integrating various characteristic factors and calculating scores for each supplier bid by the previous enterprise to sort, so that the labor cost of a buyer is saved, the characteristic factors can be adjusted according to different purchasing requirements in the sorting result, and a model is trained, so that the goal of automatically screening bidders by the enterprise in the purchasing process is achieved.
The improved supplier quantitative scoring prediction algorithm based on the gradient lifting tree avoids the situation that a tenderer subjectively selects suppliers according to own business experience, has limitation, considers the multi-dimensionality of the suppliers and the interactive characteristics in the tendering and bidding behavior process of the suppliers on data, utilizes the intelligent screening characteristics of the gradient lifting tree to calculate the weight for the characteristics in a nonlinear mode, saves the cost of judging the importance of the characteristics by the buyer by using the business experience, considers the interactive effect of the characteristics and has obvious effect; when the gradient lifting tree improvement algorithm calculates the quality score of a supplier, the residual error of the former tree is learned, and the residual error is closer to the real quality of the supplier, and the automatic mode of machine learning has great significance for establishing an intelligent model system of the bid and bid technology in the promotion of the acquisition and marketing business.
Claims (7)
1. A gradient-boosting-tree-based improved supplier quantitative score prediction algorithm is characterized by comprising the following steps of:
s1: the data module is used for acquiring data required by quantitative grading of a supplier and performing some cleaning, abnormal value processing and missing value processing;
s2: the feature engineering module carries out feature construction and feature intelligent screening;
s3: the training module comprises a weak learner, a regularization function and gradient lifting training;
s4: the prediction module uses the trained model to carry out testing;
s5: the evaluation module comprises a service evaluation module and an algorithm evaluation module;
s6: the online application module quantitatively scores the suppliers by using the trained model.
2. The gradient-boosting-tree-based improved supplier quantization scoring prediction algorithm as claimed in claim 1, wherein the specific process of the step S1 is as follows:
s11: data acquisition: the data is derived from a supplier attribute table and a supplier bid process behavior table;
s12: and (3) label construction: in the historical data, a supplier with a bid-winning behavior marks 1, and a supplier without the bid-winning behavior marks 0;
s13: data cleansing and outlier processing.
S14: and (5) processing missing values.
3. The gradient-boosting-tree-based improved supplier quantization scoring prediction algorithm as claimed in claim 1, wherein the specific process of the step S2 is as follows:
s21: the feature construction comprises two types, wherein the first type is the feature of the supplier and the second type is the behavior feature of the interaction between the supplier and the tenderer;
is characterized by comprising the following steps: supplier service type, proportion of bid frequency to the service type, total number of bids of the service type, bid frequency, bid-winning frequency, ratio of bid-winning amount to service type, bid-winning rate, proportion of bid-winning frequency to the service type, total amount of bid-winning, interest-offering degree, effective bid rate, response frequency, invitation frequency, response rate and overdue frequency of bid-winning service fee;
s22: intelligent screening of characteristics: finding out the locally optimal characteristics by using an information gain algorithm, wherein the information gain algorithm is as follows:
g(D,A)=H(D)-H(DIA)
4. the gradient-boosting-tree-based improved supplier quantization scoring prediction algorithm as claimed in claim 1, wherein the specific process of the step S3 is as follows:
s31: constructing a regression tree: using regression tree generation and regression tree pruning to select the optimal feature under the current condition as a division rule, namely the locally optimal feature;
s32: gradient iteration: weak learners (regression trees) are used to make up for the deficiencies of all trees that have been previously built; the residual error of the gradient lifting tree established by the gradient lifting improvement algorithm uses the mean square error: l (Y, f (x)) ═ Y-f (x))2The optimization goal of the gradient lifting improvement algorithm is to minimize the loss function J ═ ΣiL(Yi-F(Xi));
S33: overfitting was prevented using the Shrinkage algorithm, which is as follows:
5. the gradient-boosting-tree-based improved supplier quantization scoring prediction algorithm as claimed in claim 1, wherein the specific process of the step S4 is as follows:
and predicting the supplier bid score through the trained model and the new characteristic, wherein in order to distribute the score output by the gradient promotion improvement algorithm between (0, 1), the prediction score normalization formula is as follows:
6. the gradient-boosting-tree-based improved supplier quantization scoring prediction algorithm as claimed in claim 1, wherein the specific process of the step S5 is as follows:
the evaluation index uses AUC, and the feature label combination of the gradient lifting improvement algorithm and the inherent parameters of the gradient lifting improvement algorithm are debugged by observing the change of the AUC in the training set; the gradient boosting tree parameters include: learning rate, weak classifier number, sample subset proportion, branch minimum sample size, leaf node minimum sample size, tree maximum depth, maximum leaf node size, tree classification number, maximum feature subset size, loss function, and random seed parameter.
7. The gradient-boosting-tree-based improved supplier quantization scoring prediction algorithm as claimed in claim 1, wherein the specific process of the step S6 is as follows:
the gradient promotion improvement algorithm is used for integrating a plurality of characteristic factors for each supplier bid by the enterprise in the past, and calculating scores for sorting.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910889785.5A CN110659937A (en) | 2019-09-20 | 2019-09-20 | Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910889785.5A CN110659937A (en) | 2019-09-20 | 2019-09-20 | Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110659937A true CN110659937A (en) | 2020-01-07 |
Family
ID=69037409
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910889785.5A Pending CN110659937A (en) | 2019-09-20 | 2019-09-20 | Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110659937A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111259953A (en) * | 2020-01-15 | 2020-06-09 | 云南电网有限责任公司电力科学研究院 | Equipment defect time prediction method based on capacitive equipment defect data |
CN111652468A (en) * | 2020-04-27 | 2020-09-11 | 平安医疗健康管理股份有限公司 | Business process generation method and device, storage medium and computer equipment |
CN112561293A (en) * | 2020-12-08 | 2021-03-26 | 爱信诺征信有限公司 | System and electronic equipment for evaluating suppliers by buyers |
CN112766981A (en) * | 2020-12-30 | 2021-05-07 | 国网英大国际控股集团有限公司 | Business circle tree construction method and system based on machine learning |
CN114707699A (en) * | 2022-03-04 | 2022-07-05 | 易视腾科技股份有限公司 | Network television order unsubscribe prediction method based on machine learning |
CN115221652A (en) * | 2022-07-16 | 2022-10-21 | 天津布尔科技有限公司 | Engine intercooler fault prediction method and device and storage medium |
CN118350691A (en) * | 2024-03-28 | 2024-07-16 | 北京清华同得科技有限公司 | Evaluation method and system for regional intelligent introduction enterprises |
-
2019
- 2019-09-20 CN CN201910889785.5A patent/CN110659937A/en active Pending
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111259953A (en) * | 2020-01-15 | 2020-06-09 | 云南电网有限责任公司电力科学研究院 | Equipment defect time prediction method based on capacitive equipment defect data |
CN111259953B (en) * | 2020-01-15 | 2023-10-20 | 云南电网有限责任公司电力科学研究院 | Equipment defect time prediction method based on capacitive equipment defect data |
CN111652468A (en) * | 2020-04-27 | 2020-09-11 | 平安医疗健康管理股份有限公司 | Business process generation method and device, storage medium and computer equipment |
CN112561293A (en) * | 2020-12-08 | 2021-03-26 | 爱信诺征信有限公司 | System and electronic equipment for evaluating suppliers by buyers |
CN112766981A (en) * | 2020-12-30 | 2021-05-07 | 国网英大国际控股集团有限公司 | Business circle tree construction method and system based on machine learning |
CN114707699A (en) * | 2022-03-04 | 2022-07-05 | 易视腾科技股份有限公司 | Network television order unsubscribe prediction method based on machine learning |
CN115221652A (en) * | 2022-07-16 | 2022-10-21 | 天津布尔科技有限公司 | Engine intercooler fault prediction method and device and storage medium |
CN118350691A (en) * | 2024-03-28 | 2024-07-16 | 北京清华同得科技有限公司 | Evaluation method and system for regional intelligent introduction enterprises |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110659937A (en) | Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm | |
Cheung et al. | A multi-perspective knowledge-based system for customer service management | |
CN104700152B (en) | A kind of tobacco Method for Sales Forecast method of fusion season sales information and search behavior information | |
CN108388974A (en) | Top-tier customer Optimum Identification Method and device based on random forest and decision tree | |
CN108280541A (en) | Customer service strategies formulating method, device based on random forest and decision tree | |
CN1413332A (en) | Methods and system for automated inferred valuation of credit scoring | |
CN108830492B (en) | Method for determining spot-check merchants based on big data | |
CN106408325A (en) | User consumption behavior prediction analysis method based on user payment information and system | |
CN112149352B (en) | Prediction method for marketing activity clicking by combining GBDT automatic characteristic engineering | |
CN112561568A (en) | Target customer prediction method, device and storage medium | |
CN116911962B (en) | Article selecting device and method based on data model | |
CN108182597A (en) | A kind of clicking rate predictor method based on decision tree and logistic regression | |
CN109993380A (en) | A kind of information processing method, device and computer readable storage medium | |
CN113674013A (en) | Advertisement bidding adjustment method and system based on merchant self-defined rules | |
CN111178957A (en) | Method for early warning sudden increase of electric quantity of electricity consumption customer | |
CN114372848A (en) | Tobacco industry intelligent marketing system based on machine learning | |
CN118037401A (en) | Knowledge graph-based agricultural product electronic commerce recommendation system | |
CN116663909A (en) | Provider risk identification data processing method and device | |
CN111861679A (en) | Commodity recommendation method based on artificial intelligence | |
CN111222914A (en) | Accurate auxiliary type delivery system for goods based on e-commerce platform | |
CN114862506A (en) | Financial product recommendation method based on deep reinforcement learning | |
CN114004691A (en) | Line scoring method, device, equipment and storage medium based on fusion algorithm | |
CN117807452A (en) | Ordering method, device, equipment and storage medium based on target matching | |
CN112767114A (en) | Enterprise diversified decision method and device, electronic equipment and storage medium | |
CN116503001A (en) | Method, device, equipment and medium for generating purchase order |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200107 |