CN108648023A - A kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree - Google Patents
A kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree Download PDFInfo
- Publication number
- CN108648023A CN108648023A CN201810485114.8A CN201810485114A CN108648023A CN 108648023 A CN108648023 A CN 108648023A CN 201810485114 A CN201810485114 A CN 201810485114A CN 108648023 A CN108648023 A CN 108648023A
- Authority
- CN
- China
- Prior art keywords
- data
- businessman
- model
- mean value
- passenger flow
- 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
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
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (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 present invention relates to a kind of businessman's passenger flow forecast methods of fusion history mean value and boosted tree, which is characterized in that includes the following steps：The complete behavioral data of the businessman of certain time period is pre-processed；To passing through pretreated data construction feature；Passenger flow forecast model is built based on history mean value and boosted tree；Carry out passenger flow forecast.The present invention proposes internet businessman's passenger flow forecast model that history mean value is merged with boosted tree.The essence of the model is to promote tree-model and history mean value model, according to the weight coefficient calculated by calculation formula, the weighted sum that merges according to a certain percentage.The present invention not only allows for how improving the precision of prediction of model, and also contemplates the dependence of the prediction and time of the volume of the flow of passengers, and is made that comparative analysis to the prediction result of different models.
Description
Technical field
The present invention relates to it is a kind of fusion history mean value and boosted tree passenger flow forecast model, belong to Intelligent Information Processing and
Machine learning field.
Background technology
The development of Location based service makes the transaction data sharp increase of internet businessman " on-line off-line ".Comparison tradition
Retail trade, the marketing of internet businessman gives more concerns to customer consumption, introduction, customer service in product details page
Service, easily mobile payment etc. is devoted to bring better consumption experience for user.For example, certain business intelligence clothes
Business platform can provide sales forecast for each businessman.Based on prediction result, businessman can establish trusting relationship with user, attract
To more loyalties user and optimizing management decision, reduce cost, improve user experience.
Existing sales forecast technology is generally by historical data, and simply the progress of usage time weighting sequence method is pre-
It surveys.But in real life, the consumer behavior of user suffers from the influence of the factors such as festivals or holidays, weather, at this point, existing skill
Art can not predict the volume of the flow of passengers of businessman in time, and precision of prediction may be caused unsatisfactory, and the volume of the flow of passengers predicted is in very great Cheng
Deviate the practical volume of the flow of passengers of businessman on degree.
Invention content
The object of the present invention is to provide a kind of methods that more can predict to precision the volume of the flow of passengers.
In order to achieve the above object, the technical solution of the present invention is to provide a kind of quotient of fusion history mean value and boosted tree
Family's passenger flow forecast method, which is characterized in that include the following steps：
Step 1 pre-processes the complete behavioral data of the businessman of certain time period, and businessman is complete, and behavior data packet includes quotient
Family's characteristic, user's payment behavior data and user browsing behavior data；
Step 2, to passing through pretreated data construction feature, increase festivals or holidays data and weather characteristics data；
Step 3 builds passenger flow forecast model based on history mean value and boosted tree, includes the following steps：
Step 301 builds 2 learning models to XGBoost and GBDT respectively, to the depth of 2 learning models adjustment trees,
The parameter of learning rate and iterations, determine XGBoost learning models learning rate and tree depth capacity when, introduce
Built-in cv functions in XGBoost learning models；
Step 302, the data obtained using step 2 are trained XGBoost learning models and GBDT learning models, if
It is fixed to predict day, the average volume of the flow of passengers, sales volume increment before calculating prediction day to some day；.
Step 4, the volume of the flow of passengers trained using the correlation matrix of the history sales volume of section in those years as step 3 are pre-
The input for surveying model, by the Model Fusion of the sales volume of the following certain time period and XGBoost learning models and GBDT learning models
Weight coefficient Credit as output：
In formula,It is the average sales volume of section in those years；FuslastIt is the sales volume of section in those years, as a result,
The average sales volume for the section in those years that XGBoost learning models, GBDT learning models and history mean value model are obtained and
Sales volume value substitutes into weight coefficient Credit formula, finds out corresponding weight coefficient respectively, finally, 2 groups that training is obtained
The Different Results of XGBoost learning models and 2 groups of GBDT learning models are corresponding by what is found out respectively to history mean value model respectively
Weight coefficient ratio fusion, obtain the volume of the flow of passengers for predicting the following certain time period.
Preferably, pretreatment described in step 1 includes the following steps：
The preceding 7 days data of businessman's opening and sales volume interrupt 3 days front and back in step 101, the rejecting complete behavioral data of businessman
Data, remaining data is divided into training set and test set；
Duplicate data in step 102, removal training set and test set, using rule-based method to training set and survey
Examination concentrates the data after duplicate removal to be normalized, to eliminate single user in the short time make a big purchase in large quantities and caused by it is abnormal
Data；
For due to special timing node and caused by abnormal data and unpredicted fluctuation and caused by it is abnormal
Data are rejected using model pre-training method, that is, use poor fitting algorithm to carry out pre-training to passenger flow forecast model, remove number
The data for being 10% and 25% according to middle residual error.
Preferably, the step 2 includes the following steps：
Step 201, the weather data for acquiring national each province and city；
Weather conditions simple conversion for Precipitation Index and is become a fine day index two indices, and generates human comfort by step 202
Spend an important feature of the index as passenger flow forecast model training；
Step 203, the festivals or holidays data for acquiring current slot will be labeled as 0 working day, and weekend is labeled as 1, vacation mark
Note is 2.
The present invention proposes internet businessman's passenger flow forecast model that history mean value is merged with boosted tree.The sheet of the model
Matter is to promote tree-model to merge according to a certain percentage according to the weight coefficient calculated by calculation formula with history mean value model
Weighted sum.The present invention not only allows for how improving the precision of prediction of model, and also contemplate the volume of the flow of passengers prediction and when
Between dependence, and comparative analysis is made that the prediction result of different models.
Description of the drawings
Fig. 1 is history mean value and boosted tree Fusion Model prognostic chart：
Fig. 2 is time series Weight Regression Model prognostic chart.
Specific implementation mode
Present invention will be further explained below with reference to specific examples.It should be understood that these embodiments are merely to illustrate the present invention
Rather than it limits the scope of the invention.In addition, it should also be understood that, after reading the content taught by the present invention, people in the art
Member can make various changes or modifications the present invention, and such equivalent forms equally fall within the application the appended claims and limited
Range.
The present invention provides a kind of businessman's passenger flow forecast methods of fusion history mean value and boosted tree, including following step
Suddenly：
Step 1：The complete behavioral data of businessman is pre-processed
The data that the present invention uses come from Tianchi big data platform, include the quotient on July 1st, 1 to 31 days October next year altogether
The complete behavioral data of family.Wherein include " businessman feature " data, " user's payment behavior " data and " user browsing behavior " data.
Due to directly not only will produce error using initial data training pattern, a large amount of computing resource can be also expended.Therefore, to original
The processing such as exceptional value present in data set rejected, duplicate removal, normalization.On the one hand, since businessman is from platform is entered to pin
There are certain startup times for the amount of selling increase, and are likely to occur the phenomenon that certain section of time sales volume interrupts, therefore, before businessman's opening
7 days data and sales volumes interrupt front and back 3 days data not as training data；On the other hand, due to existing in initial data
The case where single user makes a big purchase in large quantities in short time uses rule-based to eliminate influence of this abnormal consumption to prediction
Method initial data is normalized.In addition, there is also some special timing nodes and unpredicted in initial data
Fluctuation：Such as large-scale festivals or holidays (such as Mid-autumn Festival, National Day), stop doing business, single user is big when businessman carries out advertising campaign
The case where amount purchase.For these rule-based reluctant exceptional values of method, present invention employs model pre-training sides
Method.That is, using poor fitting algorithm to passenger flow forecast model pre-training first, it is 10% and 25% to remove residual error in initial data
Data.Since prediction target is the day sales volume of businessman, the data after pretreatment for training are the quotient counted by the hour
The total sales volume of family.
Step 2: to passing through pretreated data construction feature.
To improve the accuracy of model prediction, the present invention acquires the weather data and festivals or holidays day destiny in national each province and city
According to the supplement as initial data.In the data such as gas epidemic disaster, the air pressure additionally acquired, rule of thumb, by weather conditions letter
Single-turn is changed to Precipitation Index and becomes a fine day index two indices, since human body is for the non-linear relationship of the impression of meteorologic parameter, therefore
Generate an important spy of the Body Comfort Index (Comfort Index of Human Body, SSD) as model training
Sign.Finally, model training and the feature and label predicting to use are as shown in table 1.
The feature that 1 model training of table is used with prediction
Step 3: building passenger flow forecast model based on history mean value and boosted tree.
To obtain the high passenger flow forecast model of accuracy, present invention employs the training methods in two stages.For the first time
In the training in stage, XGBoost (eXtreme Gradient Boost) and GBDT (Gradient Boosting have been used
Decision Tree) model.The parameter of model training is as shown in table 2 and table 3.Each model used respectively 2 groups of parameters into
Row training, obtains 4 models in total.
The different parameters of 2 XGBoost algorithms of table
| XGBoost | No. 1 | No. 2 |
| Object function | Linear regression model (LRM) | Linear regression model (LRM) |
| The depth capacity of tree | 3 | 5 |
| Learning rate | 0.1 | 0.03 |
| Boosted tree number | 500 | 1600 |
| L1 regularization term parameters | 0 | 1 |
| L2 regularization term parameters | 1 | 0 |
The different parameters of 3 GBDT algorithms of table
| GBDT | The depth capacity of tree | Learning rate | Boosted tree number | Training oversampling ratio |
| No. 1 | 3 | 0.1 | 500 | 0.95 |
| No. 2 | 5 | 0.1 | 500 | 0.95 |
The present invention adjusts the parameter of depth, learning rate and the iterations set in XGBoost and GBDT algorithms,
In No. 1 model of XGBoost algorithms, under normal circumstances, the value of learning rate is defaulted as 0.1, and the depth capacity set is defaulted as 3.
But for different problems, ideal learning rate can sometimes fluctuate between some specific interval ranges.The depth of tree
It is bigger, then it is higher to the fitting degree of data.Therefore, the present invention No. 2 models for determining XGBoost algorithms learning rate and
When the depth capacity of tree, cv functions built-in in XGBoost algorithms are introduced into, cv functions are tested in each round iteration using intersection
Card, according to the adjustment of algorithm parameter, and returns to ideal decision tree quantity, therefore, is more accurately calculated by cv functions, will
The learning rate of No. 2 models is adjusted to 0.03, and the depth capacity of tree is 5.The training of second stage has used history mean value model.History
Mean value model finds out 21 days sales volume average value before prediction day, obtains daily average pin first on the basis of predicting day
Amount；Secondly, as unit of week, the median and average value of sales volume weekly are counted, sales volume weekly is obtained by linear fit
Increment.
Step 4: carrying out multi-model Weighted Fusion to trained learner, businessman's volume of the flow of passengers is predicted；In the past 21
The correlation matrix of it history sales volume is as input；By the mould of following two weeks sales volumes and history mean value model and first stage
The weight coefficient of type fusion is as output.The integration percentage of mean value model is up to 0.75.The weight coefficient Credit meters of fusion
Calculate such as formula：
In formula,It is average sales volume past three weeks, FuslastFor sales volume past three weeks.As a result, by XGBoost,
(history mean value model is one kind on the basis of predicting day, and some day is arrived before finding out prediction day for GBDT and history mean value model
The information such as the average volume of the flow of passengers, sales volume increment, then using weight coefficient as the ratio of fusion, reach the passenger flow of following 14 days of prediction
Amount) obtained average sales volume past three weeks and sales volume value, it substitutes into weight coefficient formula respectively, corresponding weight can be found out
Coefficient is：0.47,0.34,0.19.Finally, the Different Results of the 2 groups of XGBoost models and 2 groups of GBDT that training are obtained are distinguished
The ratio fusion for pressing 0.47,0.34,0.19 respectively with history mean value model, obtains the volume of the flow of passengers of following 14 days of prediction.
By optimization algorithm parameter, modeling result is predicted using test set sample, the operation result and essence of algorithm
Degree test is as shown in table 4.
4 history mean value of table and boosted tree Fusion Model accuracy test
Performance Evaluation has been carried out to the model of proposition using XGBoost customized evaluation functions in experiment.Call evaluation
When function, the predicted value on verification collection and verification collection is passed to as function parameter, returns to the assessed value of a floating point type
fevalerror.The value of fevalerror is bigger, and model prediction accuracy is lower.Conversely, the value of fevalerror is smaller, model is pre-
It is higher to survey precision.The result shows that increasing with training set sample size, operation time increases, and fevalerror values gradually subtract
It is small, it is but gradually increased in precision.The Fusion Model of history mean value and boosted tree has that precision of prediction is higher, arithmetic speed as a result,
Faster advantage.
Since time series reflects feature of the entity attribute in time sequencing, it is thereby achieved that time series weights
Regression algorithm, analyze 2 kinds of algorithms prediction result after, obtain Fig. 1 and preceding 500 internet businessmans shown in Fig. 2 future 14
It volume of the flow of passengers development trend.Wherein, horizontal axis is the ID number of businessman, and the longitudinal axis then indicates the predicted value of the volume of the flow of passengers.Analyze the volume of the flow of passengers
Known to development trend：
1) maximum to the percentage contribution of model with the relevant variable of browse action, it is interacted most this is because browsing is user
Major way, abundant information degree are far above other feature；
2) commodity that part businessman may be managed are had higher rating, and the return rate of customer makes the volume of the flow of passengers of part businessman steady
Step rises.
3) the most total volume of the flow of passengers of businessman's fortnight has breached 5000, has been even up to about 25000 grade on a small quantity
Not.This is particularly likely that caused by certain recent advertising campaign of businessman.For example it is distributed by platform different degrees of preferential
Certificate, cash red packet buy the activities such as completely certain amount of money is preferential.But the migration efficiency for how adjusting oneself is attracted to more passenger flows
Amount seems most important.
Claims (3)
1. a kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree, which is characterized in that include the following steps：
Step 1 pre-processes the complete behavioral data of the businessman of certain time period, and the complete behavior data packet of businessman includes businessman spy
Levy data, user's payment behavior data and user browsing behavior data；
Step 2, to passing through pretreated data construction feature, increase festivals or holidays data and weather characteristics data；
Step 3 builds passenger flow forecast model based on history mean value and boosted tree, includes the following steps：
Step 301 builds 2 learning models to XGBoost and GBDT respectively, depth, study to 2 learning model adjustment trees
The parameter of rate and iterations, determine XGBoost learning models learning rate and tree depth capacity when, introduce
Built-in cv functions in XGBoost learning models；
Step 302, the data obtained using step 2 are trained XGBoost learning models and GBDT learning models, and setting is pre-
It surveys day, the average volume of the flow of passengers, sales volume increment before calculating prediction day to some day；.
Step 4, the passenger flow forecast mould that the correlation matrix of the history sales volume of section in those years has been trained as step 3
The input of type, by the power of the sales volume of the following certain time period and XGBoost learning models and the Model Fusion of GBDT learning models
Weight coefficient Credit is as output：
In formula,It is the average sales volume of section in those years；FuslastIt is the sales volume of section in those years, as a result, will
The average sales volume and pin for the section in those years that XGBoost learning models, GBDT learning models and history mean value model obtain
Magnitude substitutes into weight coefficient Credit formula, finds out corresponding weight coefficient respectively, finally, 2 groups that training is obtained
The Different Results of XGBoost learning models and 2 groups of GBDT learning models are corresponding by what is found out respectively to history mean value model respectively
Weight coefficient ratio fusion, obtain the volume of the flow of passengers for predicting the following certain time period.
2. businessman's passenger flow forecast method of a kind of fusion history mean value and boosted tree according to claim 1, feature
It is, pretreatment described in step 1 includes the following steps：
Step 101 rejects the number that the preceding 7 days data of businessman's opening and sales volume in the complete behavioral data of businessman interrupt front and back 3 days
According to remaining data is divided into training set and test set；
Duplicate data in step 102, removal training set and test set, using rule-based method to training set and test set
Data after middle duplicate removal are normalized, to eliminate single user in the short time make a big purchase in large quantities and caused by abnormal number
According to；
For due to special timing node and caused by abnormal data and unpredicted fluctuation and caused by abnormal data,
It is rejected using model pre-training method, that is, uses poor fitting algorithm to carry out pre-training to passenger flow forecast model, in clearing data
The data that residual error is 10% and 25%.
3. businessman's passenger flow forecast method of a kind of fusion history mean value and boosted tree according to claim 1, feature
It is, the step 2 includes the following steps：
Step 201, the weather data for acquiring national each province and city；
Weather conditions simple conversion for Precipitation Index and is become a fine day index two indices, and generates human comfort and refer to by step 202
An important feature of the number as passenger flow forecast model training；
Step 203, the festivals or holidays data for acquiring current slot, will be labeled as 0 working day, weekend is labeled as 1, and vacation is labeled as
2。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810485114.8A CN108648023A (en) | 2018-05-18 | 2018-05-18 | A kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810485114.8A CN108648023A (en) | 2018-05-18 | 2018-05-18 | A kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108648023A true CN108648023A (en) | 2018-10-12 |
Family
ID=63756926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810485114.8A Pending CN108648023A (en) | 2018-05-18 | 2018-05-18 | A kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108648023A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109272365A (en) * | 2018-10-15 | 2019-01-25 | 广东工业大学 | A kind of flow of restaurant guests prediction technique, device and equipment based on data mining |
| CN109658141A (en) * | 2018-12-12 | 2019-04-19 | 中智关爱通（上海）科技股份有限公司 | Merchandise sales prediction technique, server and storage medium |
| CN109697630A (en) * | 2018-11-21 | 2019-04-30 | 浙江大学城市学院 | A kind of businessman's volume of the flow of passengers multiplicity and prediction technique based on sparse regression |
| CN109727073A (en) * | 2018-12-29 | 2019-05-07 | 携程旅游网络技术（上海）有限公司 | Flowing of access control method, system, electronic equipment and storage medium |
| CN109784979A (en) * | 2018-12-19 | 2019-05-21 | 重庆邮电大学 | A kind of supply chain needing forecasting method of big data driving |
| CN109871702A (en) * | 2019-02-18 | 2019-06-11 | 深圳前海微众银行股份有限公司 | Federal model training method, system, equipment and computer readable storage medium |
| CN111148118A (en) * | 2019-12-18 | 2020-05-12 | 福建省南鸿通讯科技有限公司 | Flow prediction and carrier turn-off method and system based on time sequence |
| CN112163704A (en) * | 2020-09-29 | 2021-01-01 | 筑客网络技术（上海）有限公司 | High-quality supplier prediction method for building material tender platform |
| CN112165412A (en) * | 2020-09-24 | 2021-01-01 | 中盈优创资讯科技有限公司 | Flow prediction method based on time series model |
| CN112308351A (en) * | 2019-07-26 | 2021-02-02 | 上海漕泾热电有限责任公司 | Heat supply balance prediction system and method for cogeneration power plant |
| CN112396248A (en) * | 2020-11-30 | 2021-02-23 | 中国民航信息网络股份有限公司 | Method and system for processing flight booking value based on mean value model |
| CN113657652A (en) * | 2021-07-31 | 2021-11-16 | 腾讯科技（深圳）有限公司 | Flow quantity prediction method, device, equipment and readable storage medium |
| CN117557300A (en) * | 2024-01-12 | 2024-02-13 | 湖南大学 | Method and system for deducing business liveness based on energy consumption data of main equipment |
-
2018
- 2018-05-18 CN CN201810485114.8A patent/CN108648023A/en active Pending
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109272365A (en) * | 2018-10-15 | 2019-01-25 | 广东工业大学 | A kind of flow of restaurant guests prediction technique, device and equipment based on data mining |
| CN109697630A (en) * | 2018-11-21 | 2019-04-30 | 浙江大学城市学院 | A kind of businessman's volume of the flow of passengers multiplicity and prediction technique based on sparse regression |
| CN109697630B (en) * | 2018-11-21 | 2020-08-04 | 浙江大学城市学院 | Sparse regression-based merchant passenger flow volume multi-factor analysis and prediction method |
| CN109658141A (en) * | 2018-12-12 | 2019-04-19 | 中智关爱通（上海）科技股份有限公司 | Merchandise sales prediction technique, server and storage medium |
| CN109784979A (en) * | 2018-12-19 | 2019-05-21 | 重庆邮电大学 | A kind of supply chain needing forecasting method of big data driving |
| CN109727073A (en) * | 2018-12-29 | 2019-05-07 | 携程旅游网络技术（上海）有限公司 | Flowing of access control method, system, electronic equipment and storage medium |
| CN109727073B (en) * | 2018-12-29 | 2021-03-05 | 携程旅游网络技术（上海）有限公司 | Access flow control method, system, electronic device and storage medium |
| CN109871702A (en) * | 2019-02-18 | 2019-06-11 | 深圳前海微众银行股份有限公司 | Federal model training method, system, equipment and computer readable storage medium |
| CN112308351A (en) * | 2019-07-26 | 2021-02-02 | 上海漕泾热电有限责任公司 | Heat supply balance prediction system and method for cogeneration power plant |
| CN112308351B (en) * | 2019-07-26 | 2022-09-09 | 上海漕泾热电有限责任公司 | Heat supply balance prediction system and method for cogeneration power plant |
| CN111148118A (en) * | 2019-12-18 | 2020-05-12 | 福建省南鸿通讯科技有限公司 | Flow prediction and carrier turn-off method and system based on time sequence |
| CN111148118B (en) * | 2019-12-18 | 2023-03-14 | 福建省南鸿通讯科技有限公司 | Flow prediction and carrier wave turn-off method and system based on time sequence |
| CN112165412A (en) * | 2020-09-24 | 2021-01-01 | 中盈优创资讯科技有限公司 | Flow prediction method based on time series model |
| CN112163704A (en) * | 2020-09-29 | 2021-01-01 | 筑客网络技术（上海）有限公司 | High-quality supplier prediction method for building material tender platform |
| CN112163704B (en) * | 2020-09-29 | 2021-05-14 | 筑客网络技术（上海）有限公司 | High-quality supplier prediction method for building material tender platform |
| CN112396248A (en) * | 2020-11-30 | 2021-02-23 | 中国民航信息网络股份有限公司 | Method and system for processing flight booking value based on mean value model |
| CN113657652A (en) * | 2021-07-31 | 2021-11-16 | 腾讯科技（深圳）有限公司 | Flow quantity prediction method, device, equipment and readable storage medium |
| CN113657652B (en) * | 2021-07-31 | 2023-06-20 | 腾讯科技（深圳）有限公司 | Method, device, equipment and readable storage medium for predicting flow quantity |
| CN117557300A (en) * | 2024-01-12 | 2024-02-13 | 湖南大学 | Method and system for deducing business liveness based on energy consumption data of main equipment |
| CN117557300B (en) * | 2024-01-12 | 2024-04-05 | 湖南大学 | Method and system for deducing business liveness based on energy consumption data of main equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108648023A (en) | A kind of businessman's passenger flow forecast method of fusion history mean value and boosted tree | |
| Li et al. | Risk prediction in financial management of listed companies based on optimized BP neural network under digital economy | |
| CN108846520B (en) | Loan overdue prediction method, loan overdue prediction device and computer-readable storage medium | |
| CN109255506B (en) | Internet financial user loan overdue prediction method based on big data | |
| CN110599336B (en) | Financial product purchase prediction method and system | |
| CN106529729A (en) | Method and system for forecasting default of credit card user based on BP_Adaboost model | |
| CN112561598A (en) | Customer loss prediction and retrieval method and system based on customer portrait | |
| CN107578270A (en) | A kind of construction method, device and the computing device of financial label | |
| CN110415111A (en) | Merge the method for logistic regression credit examination & approval with expert features based on user data | |
| US11210673B2 (en) | Transaction feature generation | |
| CN109285038A (en) | Method for predicting bank customer loss through BP neural network | |
| CN104321794A (en) | A system and method using multi-dimensional rating to determine an entity's future commercial viability | |
| CN108389069A (en) | Top-tier customer recognition methods based on random forest and logistic regression and device | |
| CN108364191A (en) | Top-tier customer Optimum Identification Method and device based on random forest and logistic regression | |
| CN106294882A (en) | Data digging method and device | |
| CN115829683A (en) | Power integration commodity recommendation method and system based on inverse reward learning optimization | |
| CN116187808A (en) | Electric power package recommendation method based on virtual power plant user-package label portrait | |
| CN117522479A (en) | Accurate Internet advertisement delivery method and system | |
| Rofik et al. | The Optimization of Credit Scoring Model Using Stacking Ensemble Learning and Oversampling Techniques | |
| CN111967973B (en) | Bank customer data processing method and device | |
| CN116502898A (en) | Enterprise risk portrait generation method and device based on neural network | |
| CN113298179B (en) | Customs commodity abnormal price detection method and device | |
| CN109635289A (en) | Entry classification method and audit information abstracting method | |
| CN114066602A (en) | Financial industry risk control method and device | |
| CN114493724A (en) | Multi-task learning-based bidding keyword effect estimation model construction method |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181012 |
|
| WD01 | Invention patent application deemed withdrawn after publication |