CN111401941A - Vehicle sales prediction method based on XGboost recommendation algorithm - Google Patents
Vehicle sales prediction method based on XGboost recommendation algorithm Download PDFInfo
- Publication number
- CN111401941A CN111401941A CN202010151726.0A CN202010151726A CN111401941A CN 111401941 A CN111401941 A CN 111401941A CN 202010151726 A CN202010151726 A CN 202010151726A CN 111401941 A CN111401941 A CN 111401941A
- Authority
- CN
- China
- Prior art keywords
- data
- vehicle
- xgboost
- prediction
- sales
- 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
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
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Game Theory and Decision Science (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a vehicle sales prediction method based on an XGboost recommendation algorithm, which comprises the steps of firstly, utilizing '0' value filling, mean value filling, mode filling, XGboost filling and the like to realize accurate filling of data; then, based on the filled data, by utilizing a sliding window technology of vehicle yield and sales volume and combining a one _ hot coding technology, extracting a characteristic value of the vehicle information; and finally, inputting the extracted vehicle information features into a high-precision vehicle prediction model XFVS based on an XGboost algorithm to realize the accurate prediction of the vehicle sales volume. According to the method, the purpose of improving the prediction precision is achieved by filling missing data and extracting important features according to the historical sales data of the vehicle.
Description
Technical Field
The invention belongs to the technical field of sales forecasting, relates to a vehicle sales forecasting method, and particularly relates to a vehicle sales forecasting method based on an XGboost recommendation algorithm.
Background
Accurate passenger car sales forecasts are critical to passenger car enterprises and governments. For passenger car enterprises, research and development and production processes of passenger cars need relatively large research and development cost, time cost and inventory cost, and when the passenger car enterprises use consumers to purchase the passenger cars, accurate vehicle sales prediction is carried out on sales data such as vehicle attributes, sales prices, sales dates and the like according to vehicle selective purchasing, so that the key is to make reasonable production plans, adjust production, control cost and reduce loss. For government departments, the market development of the passenger cars can be mastered by utilizing the sales volume prediction of the passenger cars, the capacity can be monitored, and the development policy of the passenger car industry can be adjusted. In view of the wide application prospect and the huge economic value of accurate vehicle sales forecast, the vehicle sales forecast has become the leading direction of attention of automobile production enterprises, government departments and academic researchers in recent years.
The existing vehicle sales prediction methods analyze factors influencing vehicle sales from economic indexes such as total domestic production value, income level and disposable income and then predict the sales, but the vehicle indexes such as brands, vehicle types, discharge capacity, price range of successful delivery, power, fuel types, vehicle sizes and the like are not considered from the needs of consumers, so that the dependence of vehicle prediction on single economic indexes is too strong, the fluctuation of predicted values is too large, the prediction precision is low and the like, and the accurate vehicle sales prediction cannot be improved.
Disclosure of Invention
In order to solve the technical problems, the invention provides a vehicle sales prediction method based on an XGboost recommendation algorithm.
The technical scheme adopted by the invention is as follows: a vehicle sales prediction method based on an XGboost recommendation algorithm is characterized by comprising the following steps:
step 1: acquiring an original data set;
step 2: processing an original data set, including mean value filling, XGboost filling and '0' value filling operation, and finally outputting a complete data set;
and step 3: extracting features of the output complete data set;
acquiring the category characteristics of the vehicle by adopting one _ hot codes, acquiring the basic characteristics of the vehicle by adopting the generated unique configuration, and acquiring the time sequence characteristics of the vehicle by adopting a time sequence difference-based sliding window method to form a characteristic set for model training and prediction;
and 5: extracting the characteristic data Xt={Xt1,Xt2,Xt3,...,XtTInputting the data into an XGboost-based model to realize the prediction of vehicle sales;
the XGboost model is as follows:
wherein the content of the first and second substances,is a prediction value based on a time series t, Xt={Xt1,Xt2,Xt3,...,XtTIs the extracted time-series based feature set, ρ is the model's parameter set, Remp (f) is the empirical risk;
the prediction functions f for linear regression prediction and non-linear regression are:
F(x,ρ)=ax+b (1)
F(x,ρ)=(a.ψ(x))+b (2)
for data in which the data is linear in the input space, linear regression prediction is performed using equation 1; for the data which is not linear in the input space, the data is mapped to the high-dimensional feature space through the kernel function to execute linear regression in the high-dimensional feature space, so that the empirical risk is minimized, and the prediction accuracy of the vehicle sales volume is improved.
According to the method, a vehicle sales data set is utilized, data filling such as mean filling, 0 value filling and XGboost filling is carried out on the data set to obtain complete data, a time sequence difference sliding window method based on sales is adopted, a one _ hot coding technology is combined to extract characteristic values of vehicle information, and the extracted characteristics are input into a prediction algorithm based on an XGboost model to realize high-precision vehicle sales prediction. The method can improve the vehicle sales prediction precision, reduce the production cost of enterprises, and provide theoretical support for government departments to grasp the development of the passenger vehicle market, monitor the productivity and adjust the development policy of the passenger vehicle industry.
Drawings
FIG. 1 is a flow chart of a sliding window method based on timing difference implemented by the present invention.
FIG. 2 is a flow chart of a vehicle sales prediction model according to an embodiment of the present invention.
FIG. 3 is a flow chart of a model parameter adjustment process according to an embodiment of the invention.
Detailed description of the invention
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.
Referring to fig. 1, the vehicle sales prediction method based on the XGBoost recommendation algorithm provided by the present invention includes the following steps:
step 1: acquiring an original data set from a Tianchi tournament platform of Aliyun;
step 2: processing an original data set, including mean value filling, XGboost filling and '0' value filling operation, and finally outputting a complete data set;
in this embodiment, mean value filling is performed by first defining R ═ Pj,Cj,Bj,CLjDenotes a group, j belongs to N denotes the serial number of the group, PjIndicating sales province ID, CjIndicating a sale city ID, BjIndicating a sale vehicle brand ID, C LjData representing the vehicle type ID of the sales vehicle, wherein the fields in R are the same are represented as the same group; in advance ofLine mean filling, data in the same group are as followsObtaining a mean value of the required fill data, wherein AvgiMeans within the group, SumiIndicating the sum of the entries within the group that need to be filled with data, numiIs the number of entries in the group.
In the XGboost filling in the embodiment, the known basic data is input into the XGboost algorithm for predicting the data of the same province, the same automobile brand and the same automobile type in the same group, and then the predicted value is filled into the missing data for XGboost filling. In this embodiment, "0" value is filled, mainly aiming at some fields which cannot be filled, but if the field is deleted, the field which affects the extraction features of the model and affects the prediction result is filled to "0" by adopting a "0" value filling mode.
And step 3: extracting features of the output complete data set;
acquiring the category characteristics of the vehicle by adopting one _ hot codes, acquiring the basic characteristics of the vehicle by adopting the generated unique configuration, and acquiring the time sequence characteristics of the vehicle by adopting a time sequence difference-based sliding window method to form a characteristic set for model training and prediction;
in the embodiment, the category features of the vehicle are obtained by adopting one _ hot codes, original one-dimensional features are converted into multi-dimensional features, the dimension depends on the number of different feature values of the original features, so that original category variables are changed into binary vector representation, only one index value in the binary vector is 1, and values at other positions are 0, and the category features of the vehicle are extracted.
In this embodiment, the configuration attributes of partial data existing in the data set are not unique, which greatly affects the feature extraction. For example, the vehicle model, the brand, and the like are the same, but there are a plurality of different powers, and since the prediction system predicts the sales volume of a single brand and vehicle model, it does not predict the power, and it is necessary to perform an attribute-unique operation on the vehicle power. According to formula 3, the total sales in the group is removed from the sales of the single record to obtain the weight, and then the attributes in the group are multiplied by the weight by using formula 4, added together and combined into one record. And after the unique configuration is finished, corresponding replacement filling is carried out, the power of items with the same data, such as the same vehicle type, the same brand and the like, is normalized, the influence of non-unique configuration attributes on the vehicle sales prediction precision is eliminated, and therefore a foundation is laid for improving the vehicle sales prediction precision.
Where Sale represents the passenger car sales and w represents the weight of the sales in the group.
pitem=Sum(item*w) (4)
Where item represents an entry in the group, w represents the weight of the piece of data in the group, PitemRepresenting the value obtained after the attribute is normalized. The method comprises the steps of obtaining the category characteristics, the basic characteristics and the time sequence characteristics of the vehicle through one _ hot coding, generating unique configuration, time sequence difference sliding window based methods and other characteristic extraction methods, and providing data bases for next-stage model training.
Referring to fig. 2, in the embodiment, the time sequence characteristics of the vehicle are obtained by using a time sequence difference-based sliding window method, and since the data of the production volume and the sales volume of the passenger vehicle have time sequence and periodicity characteristics, the production volume and the sales volume in the current month have a certain relation with the sales volume in the previous months, such as the same growth trend, and the production volume and the sales volume in the current year and the same year have greater similarity. When the time sequence difference sliding window method is used for extracting the characteristics, the size of a time sequence difference window is assumed to be N, the same group of the pin quantity data of the previous N months is filled in the window with the size of N, the characteristics of the pin quantity of the N months are extracted, the window is continuously slid forwards, the tail data in the window are subtracted, new data are added to the head of the window, and the pin quantity characteristics of the corresponding N months are extracted to serve as the time sequence characteristics. The method utilizes the time sequence characteristics of the vehicle sales data, and continuously extracts the characteristics of the vehicle by using a time sequence difference sliding window method, thereby improving the prediction accuracy of the vehicle.
in the model training and predicting stage, the extracted feature set is divided into a training set and a verification set according to the distribution condition of time series and data, the feature set is input into a linear regression model, a GBDT, XGboost and other tree structure models and an L STM neural network model target model, corresponding logarithm difference square roots are calculated according to a formula 5, the models are trained and parameter-adjusted continuously to improve the predicting precision of the models, and the XGboost model with higher predicting precision is selected as a predicting core module of the system to form a complete high-precision vehicle sales predicting system.
Where the calculated log-difference square root is represented, n is the total number of test samples, piIs the predicted value of the ith test sample, aiIs the true value of the ith test sample.
Please refer to fig. 3, which is a process of training and adjusting parameters of a high-precision vehicle sales prediction system, where the debugging and model training of system parameters are a process of gradual iteration, and in this embodiment, parameters in the finally selected XGBoost are set and continuously debugged, then a logarithm difference square root of the iteration is obtained, the magnitude of the logarithm difference square root is compared, and the comparison is performed through a history value, so as to obtain parameters of the final system. The specific parameter debugging process comprises the following steps: firstly, initializing the learning rate of a model and the number of decision trees, then optimizing specific parameters of the decision trees, such as max _ depth, min _ child _ weight, gamma, subsample (model learning rate, tree maximum depth, node number, sample sampling rate and regularization setting), and the like, and after the optimization of the specific parameters is completed, optimizing regularization parameters of XGboost to reduce the complexity of the model, thereby improving the performance of the model. Finally, the learning rate is readjusted to determine the desired combination of parameters.
And 5: extracting the characteristic data Xt={Xt1,Xt2,Xt3,...,XtTInputting the data into an XGboost-based model to realize the prediction of vehicle sales;
the XGboost model is as follows:
wherein the content of the first and second substances,is a prediction value based on a time series t, Xt={Xt1,Xt2,Xt3,...,XtTIs the extracted time-series based feature set, ρ is the model's parameter set, Remp (f) is the empirical risk;
the prediction functions f for linear regression prediction and non-linear regression are:
F(x,ρ)=ax+b (1)
F(x,ρ)=(a.ψ(x))+b (2)
for data in which the data is linear in the input space, linear regression prediction is performed using equation 1; for the data which is not linear in the input space, the data is mapped to the high-dimensional feature space through the kernel function to execute linear regression in the high-dimensional feature space, so that the empirical risk is minimized, and the prediction accuracy of the vehicle sales volume is improved.
The invention provides a prediction algorithm based on an XGboost model, which is mainly used for extracting data with time sequence characteristics and a high-precision vehicle sales prediction system based on the XGboost by considering data characteristics which have larger influence on vehicle sales and combining the data characteristics based on a time sequence difference sliding window method. The result obtained by the method has higher precision and lower calculation cost.
It should be understood that parts of the specification not set forth in detail are well within the prior art.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (9)
1. A vehicle sales prediction method based on an XGboost recommendation algorithm is characterized by comprising the following steps:
step 1: acquiring an original data set;
step 2: processing an original data set, including mean value filling, XGboost filling and '0' value filling operation, and finally outputting a complete data set;
and step 3: extracting features of the output complete data set;
acquiring the category characteristics of the vehicle by adopting one _ hot codes, acquiring the basic characteristics of the vehicle by adopting the generated unique configuration, and acquiring the time sequence characteristics of the vehicle by adopting a time sequence difference-based sliding window method to form a characteristic set for model training and prediction;
step 4, extracting corresponding data features according to specific requirements, inputting the data features into a linear regression model, a GBDT, an XGboost and an L STM neural network model, and selecting the best model XGboost prediction algorithm from the four prediction models according to prediction accuracy and performance to serve as a high-precision vehicle sales prediction model;
and 5: extracting the characteristic data Xt={Xt1,Xt2,Xt3,...,XtTInputting the data into an XGboost-based model to realize the prediction of vehicle sales;
the XGboost model is as follows:
wherein the content of the first and second substances,is a prediction value based on a time series t, Xt={Xt1,Xt2,Xt3,...,XtTIs the extracted time-series based feature set, ρ is the model's parameter set, Remp (f) is the empirical risk;
the prediction functions f for linear regression prediction and non-linear regression are:
F(x,ρ)=ax+b (1)
F(x,ρ)=(a.ψ(x))+b (2)
for data in which the data is linear in the input space, linear regression prediction is performed using equation 1; for the data which is not linear in the input space, the data is mapped to the high-dimensional feature space through the kernel function to execute linear regression in the high-dimensional feature space, so that the empirical risk is minimized, and the prediction accuracy of the vehicle sales volume is improved.
2. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: in step 2, mean filling is performed by first defining R ═ { P ═ Pj,Cj,Bj,CLjDenotes a group, j belongs to N denotes the serial number of the group, PjIndicating sales province ID, CjIndicating a sale city ID, BjIndicating a sale vehicle brand ID, C LjData representing the vehicle type ID of the sales vehicle, wherein the fields in R are the same are represented as the same group; when mean filling is performed, the data in the same group are as followsObtaining a mean value of the required fill data, wherein AvgiMeans within the group, SumiIndicating the sum of the entries within the group that need to be filled with data, numiIs the number of entries in the group.
3. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: and 2, the XGboost filling is performed, namely for the data of the same province, the same automobile brand and the same automobile type in the same group, the known basic data is input into an XGboost algorithm for prediction, and then the predicted value is filled into missing data for XGboost filling.
4. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: the '0' value is filled in the step 2, mainly aiming at some fields which cannot be filled, but if the field is deleted, the field which influences the extraction characteristics of the model and influences the prediction result is filled to be '0' by adopting a '0' value filling mode.
5. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: in step 3, the category features of the vehicle are obtained by adopting one _ hot codes, original one-dimensional features are converted into multi-dimensional features, the dimension depends on the number of different feature values of the original features, so that original category variables are changed into binary vector representation, only one index value in the binary vector is 1, and the values of other positions are 0, and the category features of the vehicle are extracted.
6. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: in the step 3, generating unique configuration to obtain basic characteristics of the vehicle, firstly, removing total sales in the group from sales of a single record by using a formula 3 to obtain a weight, then multiplying each attribute in the group by the weight by using a formula 4, adding the attributes together and combining the attributes into a record; after the unique configuration is completed, corresponding replacement filling is carried out, and the power of the items with the same vehicle type and brand data is normalized;
where, salt represents the sales volume of the passenger car recorded individually, sum (salt) represents the total sales volume in the group, and w represents the weight of the sales volume in the group;
pitem=Sum(item*w) (4)
where item represents an entry in the group, w represents the weight of the piece of data in the group, PitemRepresents the value obtained after the attribute is normalized uniquely, and Sum (item w) represents the total value of the pieces of data multiplied by the weight occupied within the group.
7. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: and 3, acquiring the time sequence characteristics of the vehicle by adopting a time sequence difference-based sliding window method, supposing that the size of a time sequence difference window is N, filling the same group of the sales data of the previous N months into the window with the size of N, extracting the characteristics of the sales of the N months, continuously sliding the window forwards, subtracting the tail data in the window, adding new data at the head of the window, and extracting the sales characteristics of the corresponding N months as the time sequence characteristics.
8. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 1, wherein: in step 4, continuously training and debugging parameters in the four models, then obtaining a logarithm difference square root of each iterative debugging, determining model parameters according to the logarithm difference square root, finally selecting an XGboost model with the smallest logarithm difference square root as a model with the best prediction precision, and taking the model as a vehicle sales prediction model;
wherein the logarithm difference square root is:
where the calculated log-difference square root is represented, n is the total number of test samples, piIs the predicted value of the ith test sample, aiIs the true value of the ith test sample.
9. The XGboost recommendation algorithm-based vehicle sales prediction method of claim 8, wherein the parameter debugging process is: firstly, initializing the learning rate of a model and the number of decision trees, then optimizing the lifting parameters of the decision trees, and after the optimization of specific parameters is completed, optimizing the regularization parameters of XGboost to reduce the complexity of the model, thereby improving the expression of the model; finally, the learning rate is readjusted to determine the desired combination of parameters.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010151726.0A CN111401941A (en) | 2020-03-06 | 2020-03-06 | Vehicle sales prediction method based on XGboost recommendation algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010151726.0A CN111401941A (en) | 2020-03-06 | 2020-03-06 | Vehicle sales prediction method based on XGboost recommendation algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111401941A true CN111401941A (en) | 2020-07-10 |
Family
ID=71434107
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010151726.0A Pending CN111401941A (en) | 2020-03-06 | 2020-03-06 | Vehicle sales prediction method based on XGboost recommendation algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111401941A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111861667A (en) * | 2020-07-21 | 2020-10-30 | 上海仙豆智能机器人有限公司 | Vehicle recommendation method and device, electronic equipment and storage medium |
CN112613636A (en) * | 2020-11-25 | 2021-04-06 | 国网江苏省电力有限公司江阴市供电分公司 | Short-term prediction method for XGboost wind speed based on feature engineering |
CN113449002A (en) * | 2021-06-28 | 2021-09-28 | 平安银行股份有限公司 | Vehicle recommendation method and device, electronic equipment and storage medium |
CN115185996A (en) * | 2022-07-19 | 2022-10-14 | 广州凯园软件科技有限公司 | Education training comprehensive evaluation system based on point system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106600044A (en) * | 2016-12-06 | 2017-04-26 | 北京中交兴路信息科技有限公司 | Method and apparatus for determining vehicle sales quantity prediction model |
CN107705160A (en) * | 2017-07-12 | 2018-02-16 | 大连理工大学 | A kind of combination econometrics and heuristic intelligent automobile Method for Sales Forecast method and system |
CN110009405A (en) * | 2019-03-22 | 2019-07-12 | 广州威尔森信息科技有限公司 | A kind of passenger car sales volume simulating and predicting method based on network generalized extreme value model |
CN110599234A (en) * | 2019-08-13 | 2019-12-20 | 广东工业大学 | Product sales prediction method |
-
2020
- 2020-03-06 CN CN202010151726.0A patent/CN111401941A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106600044A (en) * | 2016-12-06 | 2017-04-26 | 北京中交兴路信息科技有限公司 | Method and apparatus for determining vehicle sales quantity prediction model |
CN107705160A (en) * | 2017-07-12 | 2018-02-16 | 大连理工大学 | A kind of combination econometrics and heuristic intelligent automobile Method for Sales Forecast method and system |
CN110009405A (en) * | 2019-03-22 | 2019-07-12 | 广州威尔森信息科技有限公司 | A kind of passenger car sales volume simulating and predicting method based on network generalized extreme value model |
CN110599234A (en) * | 2019-08-13 | 2019-12-20 | 广东工业大学 | Product sales prediction method |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111861667A (en) * | 2020-07-21 | 2020-10-30 | 上海仙豆智能机器人有限公司 | Vehicle recommendation method and device, electronic equipment and storage medium |
CN112613636A (en) * | 2020-11-25 | 2021-04-06 | 国网江苏省电力有限公司江阴市供电分公司 | Short-term prediction method for XGboost wind speed based on feature engineering |
CN113449002A (en) * | 2021-06-28 | 2021-09-28 | 平安银行股份有限公司 | Vehicle recommendation method and device, electronic equipment and storage medium |
CN115185996A (en) * | 2022-07-19 | 2022-10-14 | 广州凯园软件科技有限公司 | Education training comprehensive evaluation system based on point system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111401941A (en) | Vehicle sales prediction method based on XGboost recommendation algorithm | |
CN113962364B (en) | Multi-factor power load prediction method based on deep learning | |
CN101551884B (en) | A fast CVR electric load forecast method for large samples | |
CN107424070A (en) | A kind of loan user credit ranking method and system based on machine learning | |
CN111832839B (en) | Energy consumption prediction method based on sufficient incremental learning | |
CN108710905A (en) | One kind being based on the united spare part quantitative forecasting technique of multi-model and system | |
CN111476677A (en) | Big data-based electricity consumption type electricity sales quantity analysis and prediction method and system | |
CN115907611B (en) | Fitting inventory control method based on fitting market value | |
CN114792156A (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
CN110689162A (en) | Bus load prediction method, device and system based on user side classification | |
CN115470962A (en) | LightGBM-based enterprise confidence loss risk prediction model construction method | |
CN114492191A (en) | Heat station equipment residual life evaluation method based on DBN-SVR | |
CN115759415A (en) | Power consumption demand prediction method based on LSTM-SVR | |
CN110738363A (en) | photovoltaic power generation power prediction model and construction method and application thereof | |
CN112241785B (en) | Book interview method based on deep reinforcement learning | |
CN113705188A (en) | Intelligent evaluation method for customs import and export commodity specification declaration | |
Xu et al. | An improved credit card users default prediction model based on RIPPER | |
CN112288187A (en) | Big data-based electricity sales amount prediction method | |
CN116503158A (en) | Enterprise bankruptcy risk early warning method, system and device based on data driving | |
CN105741143A (en) | Load characteristic and cluster analysis based electric power commodity pricing model establishment method | |
CN116542701A (en) | Carbon price prediction method and system based on CNN-LSTM combination model | |
CN114626898A (en) | Sales forecasting method, tool, system, equipment and storage medium | |
CN115687788A (en) | Intelligent business opportunity recommendation method and system | |
CN114998048A (en) | Electric quantity change factor analysis method and device, computer equipment and storage medium | |
CN113111588A (en) | NO of gas turbineXEmission concentration prediction 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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200710 |