CN109800277A - A kind of machine learning platform and the data model optimization method based on the platform - Google Patents
A kind of machine learning platform and the data model optimization method based on the platform Download PDFInfo
- Publication number
- CN109800277A CN109800277A CN201811554215.2A CN201811554215A CN109800277A CN 109800277 A CN109800277 A CN 109800277A CN 201811554215 A CN201811554215 A CN 201811554215A CN 109800277 A CN109800277 A CN 109800277A
- Authority
- CN
- China
- Prior art keywords
- model
- data
- algorithm
- module
- machine learning
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of machine learning platform and the data model optimization method based on the platform, it is related to data model assessment technology field.The present invention includes: model construction module, data acquisition module, data processing module, model evaluation module and model optimization module；The model construction module includes model construction unit and model release unit；The model construction unit is built for model；The model release unit issues the data model built.The present invention passes through model construction module rapid build data model, source data is obtained using data acquisition module and proposes work to data model after handling by data processing module, guarantee to enter the validity of the data of data model, reduce model evaluation error, model evaluation module is by the confusion matrix and model accuracy of output model to data model evaluation, the convenient optimization to data model.
Description
Technical field
The invention belongs to data model assessment technology fields, more particularly to a kind of machine learning platform and are based on the platform
Data model optimization method.
Background technique
With the commonly used and every field of big data, the processing of data is more and more important to acquisition of information.In data
During processing, often the effect of different data model treatment is different, and different types of data is risen using different data models
The effect arrived is also far from each other.This requires apply to assess with different types of data to different data model；This mistake
Journey is very complicated, and needs to be verified with mass data.If complicated using manual operation process, it is not easy to observe, and
Heavy workload is extremely difficult to due effect.This just needs to establish machine learning platform, for building various data models, simultaneously
The machine learning platform passes through big data cluster extraction source data；And it by source data verify data model and establishes good
Model evaluation method constructs excellent data model and data model evaluation method.
This invention address that researching and developing a kind of machine learning platform and the data model optimization method based on the platform, pass through source
Data verification data model and good model evaluation method is established, constructs excellent data model and data model evaluation side
Method solves the problems, such as complicated existing data model building, verifying heavy workload and not can be carried out good model verifying.
Summary of the invention
The purpose of the present invention is to provide a kind of machine learning platform and the data model optimization method based on the platform, lead to
Model construction module rapid build data model is crossed, source data is obtained using data acquisition module and by data processing module
Work is proposed after reason to data model, and data model assessment is carried out by model evaluation module, finally uses model optimization module pair
Data model optimization, obtains good data model, solves complicated existing data model building, verifying heavy workload and not
The problem of can be carried out the verifying of good model.
In order to solve the above technical problems, the present invention is achieved by the following technical solutions:
The present invention be a kind of machine learning platform, comprising: model construction module, data acquisition module, data processing module,
Model evaluation module and model optimization module；
The model construction module includes model construction unit and model release unit；The model construction unit is used for
Model is built；The model release unit issues the data model built；
The data acquisition module obtains source data from the mysql database of hdfs big data cluster and is write as hive table；
The data processing module further includes data pre-processing unit and data filtering units；The data pre-processing unit for pair
Source data sampling, ratio are split, type is converted and Missing Data Filling；The data filtering units for filter it is extra record and
Field；
The model evaluation module further includes algorithm picks unit and assessment unit；The algorithm picks unit is according to mould
The assessment algorithm that type Feature Selection adapts to；The assessment unit is according to the assessment algorithm of selection to data model evaluation；
The model optimization module is more excellent by precise information source, adjustment data processing method, the positive negative sample of adjustment, selection
Algorithm and adjustment algorithm parameter are to model optimization.
Preferably, the data processing module is also used to pretreated data being filled into the data model corresponding
Position；The model evaluation module is by predicting the data in data model and assessing.
Preferably, several assessment algorithms are also stored in the mysql database；The assessment algorithm specifically includes logic and returns
Reduction method, Decision Tree Algorithm, NB Algorithm, random forest sorting algorithm, gradient promoted tree classification algorithm,
Kmeans algorithm, gradient boosted tree regression algorithm and decision tree regression algorithm.
Preferably, the source data sampling includes stochastical sampling and stratified sampling again；The ratio fractionation is by source data
It splits into for model training and is verified for model.
Data model optimization method based on machine learning platform, comprises the following processes:
Reading data: the data acquisition module selects tables of data from data source；
Data Mining: the data processing module is by checking that source data and visualized graphs heuristic data are distributed feelings
Condition；
Data processing: the data processing module distinguishes source data Field Sanitization, type conversion, Numerical Index, sample
And sample balance；
Algorithms of Selecting: the model evaluation module is according to the Feature Selection prediction algorithm of data model；
Model evaluation: the model evaluation module is by the confusion matrix and model accuracy of output model to data model
Assessment；
Model optimization: the model optimization module passes through precise information source, adjustment data processing method, the positive and negative sample of adjustment
Originally, the more excellent algorithm of selection and adjustment algorithm parameter are to model optimization.
Preferably, the Field Sanitization is for filtering out the field and serious loss unrelated with the data model to be established
Field；The type conversion is for being converted into double type for critical field, to adapt to enter model training；
The Numerical Index is used to establish convenient for identification and the index searched different type variable；The sample, which is distinguished, to be used
In the data separation that will be chosen at model training data and model verify data；Sample balance is for will be in training sample
Sample balance is conquered, guarantees the unbiasedness of data model.
The invention has the following advantages:
1, the present invention obtains source data simultaneously using data acquisition module by model construction module rapid build data model
Work is proposed after handling by data processing module to data model, wherein data processing module turns source data Field Sanitization, type
Change, Numerical Index, sample distinguish and sample balance, guarantee enter data model data validity, reduce model comment
Estimate error, model evaluation module, to data model evaluation, facilitates logarithm by the confusion matrix and model accuracy of output model
According to the optimization of model.
2, the present invention by model optimization module by precise information source, adjustment data processing method, the positive negative sample of adjustment,
It selects more excellent algorithm and adjustment algorithm parameter to model optimization, improves the confidence level and accuracy of data model.
Certainly, it implements any of the products of the present invention and does not necessarily require achieving all the advantages described above at the same time.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, will be described below to embodiment required
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for ability
For the those of ordinary skill of domain, without creative efforts, it can also be obtained according to these attached drawings other attached
Figure.
Fig. 1 is a kind of structural schematic diagram of machine learning platform of the invention；
Fig. 2 is the flow chart of the data model optimization method of the invention based on machine learning platform.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts all other
Embodiment shall fall within the protection scope of the present invention.
Refering to Figure 1, the present invention is a kind of machine learning platform, comprising: model construction module, data acquisition mould
Block, data processing module, model evaluation module and model optimization module；
Model construction module includes model construction unit and model release unit；Model construction unit is taken for model
It builds；Model release unit issues the data model built；
Data acquisition module obtains source data from the mysql database of hdfs big data cluster and is write as hive table；It reads
The hdfs file on big data cluster that is connected simultaneously reads mysql database；After flow of task is executed, the number of component generation
According to being written in hive table；If the table name of input is not present, then the table can be automatically created, if existing, when which executes
The table can be prompted existing, data will not be written.Data processing module further includes data pre-processing unit and data filtering list
Member；Data pre-processing unit is used for source data sampling, ratio are split, type is converted and Missing Data Filling；Wherein, type turns
The field type of table of changing commanders changes into another type, and support is converted to double, int, string three types；Missing Data Filling
For null value or a specified value are replaced with maximum value, minimum value, mean value or a customized value.It can pass through
The configured list of a given missing values fills the missing values for inputting table with specified value to realize；
Data filtering units are for filtering extra record and field；Wherein, record filtering, which refers to, expresses data according to filtering
Formula is screened；Field Sanitization, which refers to, deletes field extra in table.
Model evaluation module further includes algorithm picks unit and assessment unit；Algorithm picks unit is selected according to the aspect of model
Take the assessment algorithm of adaptation；Assessment unit is according to the assessment algorithm of selection to data model evaluation；
Model optimization module passes through precise information source, adjustment data processing method, the positive negative sample of adjustment, the more excellent algorithm of selection
And adjustment algorithm parameter is to model optimization.
Wherein, data processing module is also used to for pretreated data to be filled into the corresponding position of data model；Model
Evaluation module is by predicting the data in data model and assessing.
Wherein, several assessment algorithms are also stored in mysql database；Assessment algorithm specifically includes logistic regression algorithm, determines
Plan tree classification algorithm, NB Algorithm, random forest sorting algorithm, gradient promote tree classification algorithm, Kmeans algorithm, ladder
Spend boosted tree regression algorithm and decision tree regression algorithm.
Wherein, source data sampling includes stochastical sampling and stratified sampling again；Sampled data is generated in a random basis, supports to press
Number sampling, proportional sampling two ways；Data set by field value layered extraction certain proportion or certain data with
Press proof sheet.It is sampled by number, number of samples 100, then according to grouping column field value, each value extracts 100 datas.By than
Example sampling: the data for the extraction 50% that grouping column field value is 0, the data for the extraction 80% that field value is 1, more fields
Value, and so on.Ratio fractionation is to split into source data to be used to model training and be used for model verify；To the data of input
It is split in proportion, exports two parts of data respectively；Primary contract: the ratio that training dataset accounts for data source is defaulted as 0.8, model
Enclose is between 0 to 1.
It please refers to shown in Fig. 2, the data model optimization method based on machine learning platform comprises the following processes:
Reading data: data acquisition module selects tables of data from data source；
Data Mining: data processing module is by checking source data and visualized graphs heuristic data distribution situation；
Data processing: data processing module to source data Field Sanitization, type conversion, Numerical Index, sample distinguish and
Sample balance；
Algorithms of Selecting: model evaluation module is according to the Feature Selection prediction algorithm of data model；
Model evaluation: model evaluation module comments data model by the confusion matrix and model accuracy of output model
Estimate；
Model optimization: model optimization module passes through precise information source, adjustment data processing method, the positive negative sample of adjustment, choosing
More excellent algorithm and adjustment algorithm parameter are selected to model optimization.
Wherein, Field Sanitization is used to filter out the word of the field and serious loss unrelated with the data model to be established
Section；Type is converted for critical field to be converted into double type, to adapt to enter model training；
Numerical Index is used to establish convenient for identification and the index searched different type variable；Sample is distinguished for that will choose
Data separation at model training data and model verify data；Sample is balanced for putting down the sample of conquering in training sample
Weighing apparatus, guarantees the unbiasedness of data model.
The present invention in actual use, uses model construction unit member data model first；Model release unit
The data model built is issued.Data acquisition module selects tables of data from data source；It is specifically to read to be connected
Hdfs file on big data cluster simultaneously reads the source data in mysql database.
For data processing module by checking source data and visualized graphs heuristic data distribution situation, specific includes complete
Table statistics: the maximum value of each field, minimum value, average value, standard deviation, duplicate removal record number, missing record number, summary journal are checked
Number；Histogram: static fields value is in each section distribution situation；Cake chart: statistical classification field, every class value record sum and
Accounting.
Data processing module is distinguished to source data Field Sanitization, type conversion, Numerical Index, sample and sample balances,
Specifically are as follows: Field Sanitization possess filter out to model unrelated field (such as: User ID), missing values be more than 70% field
Deng；Critical field is converted to double type by type conversion, to fit into model training；And for classifying type variable, need
Create Numerical Index；The sample that sample subregion can generally choose 80% carries out model training, and 20% sample carries out model verifying；
In order to guarantee the unbiasedness of model, allowed in training sample as far as possible, positive and negative sample size 1:1 is balanced.Stratified sampling can be used, to just
Negative sample is sampled respectively, adjustment difference sampling proportion, so that positive and negative sample size is in 1:1 ratio.
Model evaluation module belongs to classifying type problem, institute according to the Feature Selection prediction algorithm of data model, customer churn
With selection sort type algorithm, this example selection logistic regression algorithm is demonstrated.Model evaluation module is obscured by output model
Matrix and model accuracy are to data model evaluation, and Primary Reference index is the accuracy of model, by checking as a result, exportable
The confusion matrix and accuracy model evaluation index of model.
Model optimization module passes through precise information source, adjustment data processing method, the positive negative sample of adjustment, the more excellent algorithm of selection
And adjustment algorithm parameter is to model optimization；The purpose of model optimization is the accuracy in order to improve model, can be from following five
A aspect is started with: 1, data source, according to business experience, finds out the key influence factor for influencing customer churn as far as possible；2, at data
Reason mode: including missing values processing, Field Sanitization etc.；3, positive and negative sample proportion is adjusted；4, different model algorithms is selected；5, it adjusts
The parameter of each algorithm in integral mould.
It is worth noting that, included each unit is only drawn according to function logic in the above system embodiment
Point, but be not limited to the above division, as long as corresponding functions can be realized；In addition, each functional unit is specific
Title is also only for convenience of distinguishing each other, the protection scope being not intended to restrict the invention.
In addition, those of ordinary skill in the art will appreciate that realizing all or part of the steps in the various embodiments described above method
It is that relevant hardware can be instructed to complete by program, corresponding program can store to be situated between in a computer-readable storage
In matter.
Present invention disclosed above preferred embodiment is only intended to help to illustrate the present invention.There is no detailed for preferred embodiment
All details are described, are not limited the invention to the specific embodiments described.Obviously, according to the content of this specification,
It can make many modifications and variations.These embodiments are chosen and specifically described to this specification, is in order to better explain the present invention
Principle and practical application, so that skilled artisan be enable to better understand and utilize the present invention.The present invention is only
It is limited by claims and its full scope and equivalent.
Claims (6)
1. a kind of machine learning platform characterized by comprising model construction module, data acquisition module, data processing mould
Block, model evaluation module and model optimization module；
The model construction module includes model construction unit and model release unit；The model construction unit is used for model
Build；The model release unit issues the data model built；
The data acquisition module obtains source data from the mysql database of hdfs big data cluster and is write as hive table；It is described
Data processing module further includes data pre-processing unit and data filtering units；The data pre-processing unit is used for source number
According to sampling, ratio are split, type is converted and Missing Data Filling；The data filtering units are for filtering extra record and word
Section；
The model evaluation module further includes algorithm picks unit and assessment unit；The algorithm picks unit is according to model spy
Sign chooses the assessment algorithm adapted to；The assessment unit is according to the assessment algorithm of selection to data model evaluation；
The model optimization module passes through precise information source, adjustment data processing method, the positive negative sample of adjustment, the more excellent algorithm of selection
And adjustment algorithm parameter is to model optimization.
2. a kind of machine learning platform according to claim 1, which is characterized in that the data processing module be also used to by
Pretreated data are filled into the corresponding position of the data model；The model evaluation module passes through in data model
Data prediction and assessment.
3. a kind of machine learning platform according to claim 1, which is characterized in that also stored in the mysql database
Several assessment algorithms；The assessment algorithm specifically include logistic regression algorithm, Decision Tree Algorithm, NB Algorithm,
Random forest sorting algorithm, gradient promote tree classification algorithm, Kmeans algorithm, gradient boosted tree regression algorithm and decision tree and return
Reduction method.
4. a kind of machine learning platform according to claim 1, which is characterized in that the source data sampling includes random again
Sampling and stratified sampling；The ratio fractionation is to split into source data to be used to model training and be used for model verify.
5. the data model optimization method based on machine learning platform as described in claim 1-4 is any one, which is characterized in that
It comprises the following processes:
Reading data: the data acquisition module selects tables of data from data source；
Data Mining: the data processing module is by checking source data and visualized graphs heuristic data distribution situation；
Data processing: the data processing module to source data Field Sanitization, type conversion, Numerical Index, sample distinguish and
Sample balance；
Algorithms of Selecting: the model evaluation module is according to the Feature Selection prediction algorithm of data model；
Model evaluation: the model evaluation module comments data model by the confusion matrix and model accuracy of output model
Estimate；
Model optimization: the model optimization module passes through precise information source, adjustment data processing method, the positive negative sample of adjustment, choosing
More excellent algorithm and adjustment algorithm parameter are selected to model optimization.
6. the data model optimization method according to claim 2 based on machine learning platform, which is characterized in that
The Field Sanitization is used to filter out the field of the field and serious loss unrelated with the data model to be established；It is described
Type is converted for critical field to be converted into double type, to adapt to enter model training；
The Numerical Index is used to establish convenient for identification and the index searched different type variable；The sample differentiation is used for will
The data separation of selection is at model training data and model verify data；Sample balance is for by conquering in training sample
Sample balance, guarantees the unbiasedness of data model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811554215.2A CN109800277A (en) | 2018-12-18 | 2018-12-18 | A kind of machine learning platform and the data model optimization method based on the platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811554215.2A CN109800277A (en) | 2018-12-18 | 2018-12-18 | A kind of machine learning platform and the data model optimization method based on the platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109800277A true CN109800277A (en) | 2019-05-24 |
Family
ID=66557126
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811554215.2A Pending CN109800277A (en) | 2018-12-18 | 2018-12-18 | A kind of machine learning platform and the data model optimization method based on the platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109800277A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991658A (en) * | 2019-11-28 | 2020-04-10 | 重庆紫光华山智安科技有限公司 | Model training method and device, electronic equipment and computer readable storage medium |
| CN111177495A (en) * | 2019-12-05 | 2020-05-19 | 北京永洪商智科技有限公司 | Method for intelligently identifying data content and generating corresponding industry report |
| CN111340241A (en) * | 2020-05-15 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Data processing method, system and device |
| CN111523798A (en) * | 2020-04-21 | 2020-08-11 | 武汉市奥拓智能科技有限公司 | Automatic modeling method, device and system and electronic equipment thereof |
| CN112527321A (en) * | 2020-12-29 | 2021-03-19 | 平安银行股份有限公司 | Deep learning-based application online method, system, device and medium |
| CN113516638A (en) * | 2021-06-25 | 2021-10-19 | 中南大学 | Neural network internal feature importance visualization analysis and feature migration method |
| CN114066110A (en) * | 2020-07-29 | 2022-02-18 | 太极计算机股份有限公司 | System for providing machine learning service for user |
| CN115168916A (en) * | 2022-07-26 | 2022-10-11 | 北京大数据先进技术研究院 | Digital object credible evidence storing method and system for mobile terminal application |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106250987A (en) * | 2016-07-22 | 2016-12-21 | 无锡华云数据技术服务有限公司 | A kind of machine learning method, device and big data platform |
| CN106779087A (en) * | 2016-11-30 | 2017-05-31 | 福建亿榕信息技术有限公司 | A kind of general-purpose machinery learning data analysis platform |
| CN106779084A (en) * | 2016-09-30 | 2017-05-31 | 广州英康唯尔互联网服务有限公司 | Machine learning system and method |
| CN107103050A (en) * | 2017-03-31 | 2017-08-29 | 海通安恒（大连）大数据科技有限公司 | A kind of big data Modeling Platform and method |
| CN107229976A (en) * | 2017-06-08 | 2017-10-03 | 郑州云海信息技术有限公司 | A kind of distributed machines learning system based on spark |
| WO2018063488A1 (en) * | 2016-09-30 | 2018-04-05 | Doppler Labs, Inc. | Context aware hearing optimization engine |
-
2018
- 2018-12-18 CN CN201811554215.2A patent/CN109800277A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106250987A (en) * | 2016-07-22 | 2016-12-21 | 无锡华云数据技术服务有限公司 | A kind of machine learning method, device and big data platform |
| CN106779084A (en) * | 2016-09-30 | 2017-05-31 | 广州英康唯尔互联网服务有限公司 | Machine learning system and method |
| WO2018063488A1 (en) * | 2016-09-30 | 2018-04-05 | Doppler Labs, Inc. | Context aware hearing optimization engine |
| CN106779087A (en) * | 2016-11-30 | 2017-05-31 | 福建亿榕信息技术有限公司 | A kind of general-purpose machinery learning data analysis platform |
| CN107103050A (en) * | 2017-03-31 | 2017-08-29 | 海通安恒（大连）大数据科技有限公司 | A kind of big data Modeling Platform and method |
| CN107229976A (en) * | 2017-06-08 | 2017-10-03 | 郑州云海信息技术有限公司 | A kind of distributed machines learning system based on spark |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991658A (en) * | 2019-11-28 | 2020-04-10 | 重庆紫光华山智安科技有限公司 | Model training method and device, electronic equipment and computer readable storage medium |
| CN111177495A (en) * | 2019-12-05 | 2020-05-19 | 北京永洪商智科技有限公司 | Method for intelligently identifying data content and generating corresponding industry report |
| CN111523798A (en) * | 2020-04-21 | 2020-08-11 | 武汉市奥拓智能科技有限公司 | Automatic modeling method, device and system and electronic equipment thereof |
| CN111523798B (en) * | 2020-04-21 | 2023-09-01 | 武汉市奥拓智能科技有限公司 | Automatic modeling method, device, system and electronic equipment thereof |
| CN111340241A (en) * | 2020-05-15 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Data processing method, system and device |
| CN114066110A (en) * | 2020-07-29 | 2022-02-18 | 太极计算机股份有限公司 | System for providing machine learning service for user |
| CN112527321A (en) * | 2020-12-29 | 2021-03-19 | 平安银行股份有限公司 | Deep learning-based application online method, system, device and medium |
| CN112527321B (en) * | 2020-12-29 | 2022-05-27 | 平安银行股份有限公司 | Deep learning-based application online method, system, device and medium |
| CN113516638A (en) * | 2021-06-25 | 2021-10-19 | 中南大学 | Neural network internal feature importance visualization analysis and feature migration method |
| CN115168916A (en) * | 2022-07-26 | 2022-10-11 | 北京大数据先进技术研究院 | Digital object credible evidence storing method and system for mobile terminal application |
| CN115168916B (en) * | 2022-07-26 | 2023-01-13 | 北京大数据先进技术研究院 | Digital object credible evidence storing method and system for mobile terminal application |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109800277A (en) | A kind of machine learning platform and the data model optimization method based on the platform | |
| US11429878B2 (en) | Cognitive recommendations for data preparation | |
| US11327935B2 (en) | Intelligent data quality | |
| US11422986B1 (en) | One-click database migration with automatic selection of a database | |
| JP7125900B2 (en) | A multidimensional recursive learning process and system used to discover complex dyadic or multiple counterparty relationships | |
| US11093521B2 (en) | Just-in-time data quality assessment for best record creation | |
| CN112036153B (en) | Work order error correction method and device, computer readable storage medium and computer equipment | |
| CN108960269B (en) | Feature acquisition method and device for data set and computing equipment | |
| US9286361B2 (en) | Extract-transform-load processor controller | |
| CN106776982A (en) | A kind of social media sentiment analysis method of use machine learning | |
| EP2827240A1 (en) | Method for generating control-code by a control-code-diagram | |
| CN109299178A (en) | A kind of application method and data analysis system | |
| CN107240011B (en) | Index configuration processing method and system | |
| CN112000656A (en) | Intelligent data cleaning method and device based on metadata | |
| CN110109902A (en) | A kind of electric business platform recommender system based on integrated learning approach | |
| CN113554350A (en) | Activity evaluation method and apparatus, electronic device and computer readable storage medium | |
| CN111210321B (en) | Risk early warning method and system based on contract management | |
| US7853442B2 (en) | Printed circuit board design instruction support method and device | |
| CN106228439A (en) | A kind of financial data management method and system | |
| CN105912452A (en) | Automated data analysis method and device | |
| EP3489838A1 (en) | Method and apparatus for determining an association | |
| WO2020024342A1 (en) | Data parsing method and device | |
| US10025838B2 (en) | Extract transform load input suggestion | |
| CN114610590A (en) | Method, device and equipment for determining operation time length and storage medium | |
| KR100888422B1 (en) | Testing system of user interface for financial contract and testing method using thereof |
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: 20190524 |