CN107871055A - A kind of data analysing method and device - Google Patents
A kind of data analysing method and device Download PDFInfo
- Publication number
- CN107871055A CN107871055A CN201610854748.7A CN201610854748A CN107871055A CN 107871055 A CN107871055 A CN 107871055A CN 201610854748 A CN201610854748 A CN 201610854748A CN 107871055 A CN107871055 A CN 107871055A
- Authority
- CN
- China
- Prior art keywords
- data
- task
- data analysis
- eigenmatrix
- feature
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Fuzzy Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Stored Programmes (AREA)
Abstract
The embodiment of the invention discloses a kind of data analysing method and device, when receiving data analysis task, according to the type of the data analysis task, eigenmatrix corresponding to the type can be determined from corresponding matrix storehouse, and data to be analyzed corresponding to the data analysis task can be handled as this feature matrix, it can be seen that, eigenmatrix corresponding to being obtained from matrix storehouse comes for the processing to a data analysis task, to obtain the output characteristic of the eigenmatrix, the analysis result of the data analysis task is determined according to the output characteristic.And without be this data analysis task special configuration eigenmatrix, need to improve data analysis efficiency according to the demand of the required data analyzed and data analysis task come the time of configuration feature matrix originally so as to eliminate.
Description
Technical field
The present invention relates to data processing field, more particularly to a kind of data analysing method and device.
Background technology
With the development of data analysis technique, the importance of data is lifted therewith, and data are carried out with the analysis of data analysis
As a result often can be as some corporate decisions, the important references of development.
The process of data analysis needs to use eigenmatrix and model to handle the data of required analysis, so as to
To analysis result.The analysis result of data analysis is more accurate, and the reference value played is higher.
At present, during data analysis is carried out, in the data analysis task each time of execution, it is required for as the data
Analysis task special configuration eigenmatrix, each configuration feature matrix all can consume the plenty of time, and the efficiency of data analysis needs
Improve.
The content of the invention
In order to solve the above-mentioned technical problem, the embodiments of the invention provide a kind of data analysing method and device, Ke Yijie
The time of configuration feature matrix is saved, improves data analysis efficiency.
In a first aspect, the embodiments of the invention provide a kind of data analysing method, this method includes:
Receive data analysis task;Obtain data to be analyzed corresponding to the data analysis task;Appointed according to the data analysis
The type of business determines eigenmatrix corresponding to the type from matrix storehouse;According to the processing logic of this feature matrix to the data
Handled, to obtain the output characteristic of this feature matrix;The analysis knot of the data analysis task is determined according to the output characteristic
Fruit.
It can be seen that when receiving data analysis task, can be from corresponding matrix according to the type of the data analysis task
Eigenmatrix corresponding to the type, and the number that can be analyzed by this feature matrix the data analysis required by task are determined in storehouse
According to being handled, it is seen then that eigenmatrix corresponding to being obtained from matrix storehouse comes for the processing to a data analysis task, without
With for this data analysis task special configuration eigenmatrix, so as to eliminate the data and number that need originally according to required analysis
Carry out the time of configuration feature matrix according to the demand of analysis task, improve data analysis efficiency.
In the first possible implementation of first aspect, the matrix storehouse is included in during the analysis of historic task
The eigenmatrix configured, the historic task are completed data analysis task, the type of the historic task and the data point
The type of analysis task is identical.
It can be seen that the eigenmatrix of historic task is applied in the analysis of data analysis task, reach to eigenmatrix
The effect of reuse.Reduce to the time spent by data analysis task progress data analysis, and then improve data
The efficiency of analysis.If the eigenmatrix deposited in matrix storehouse is more comprehensive, the type of data analysis task is got over corresponding to eigenmatrix
It is more, then when receiving new data analysis task, the eigenmatrix preserved before is directly reused from matrix storehouse
The probability handled is bigger.
With reference to the implementation of the first of first aspect, in second of possible implementation, in addition to:
Determine Feature Engineering corresponding with the type from Feature Engineering storehouse according to the type, this feature engineering include from
Data analysis process of the data of analysis between obtaining output characteristic from eigenmatrix needed for obtaining;This is according to this feature matrix
Processing logic the data are handled, obtain this feature matrix output characteristic, including:Include according to this feature engineering
Data analysis process, the data are handled according to the processing logic of this feature matrix, obtain this feature matrix output
Feature.
It can be seen that can be by pre-establishing Feature Engineering storehouse, and the spy that historic task is configured during analysis
The corresponding relation that the type of sign engineering and the historic task has with its Feature Engineering configured is maintained in this feature work
It is convenient in the data analysis task received in Cheng Ku, can according to the data analysis task type directly from feature work
Feature Engineering corresponding with the type is matched in Cheng Ku, the time for configuring cumbersome data analysis step is eliminated, improves
Data analysis efficiency.
With reference to second of implementation of first aspect, in the third possible implementation, the data analysis
Journey also includes the process pre-processed to the data of required analysis, and the process of the pretreatment includes data deduplication, sampling of data
With it is data-optimized in any one or more combination.
It can be seen that the preprocessing process by recording data to be analyzed, the data analysis is being multiplexed for data analysis task
During process, remove the cumbersome pre-treatment step of manual configuration from, save the time, improve the efficiency of data analysis.
In the 4th kind of possible implementation of first aspect, this feature matrix that the matrix storehouse includes is according to spy
The feature preserved in sign storehouse configures what is obtained with processing logic, and the feature preserved in this feature storehouse is according to historical data institute structure
Build what is obtained.
It can be seen that the feature database related to this field can be constructed beforehand through the data in a field, so that
Need for corresponding to the configuration of data analysis task during eigenmatrix, spy that can be to have been had determined in direct basis feature database
Processing logic of seeking peace is configured, so as to save the time consumed of configuration feature matrix to a certain extent.
With reference to the 4th kind of implementation of first aspect, in the 5th kind of possible implementation, the historical data category
In field of telecommunications, then the feature preserved in this feature storehouse obtains according to constructed by data in the field, including:This feature storehouse
The feature of middle preservation is based on the attribute entrained by data, is built what is obtained to the data in the field of telecommunications, the attribute
Including combination any one or more in user property, position attribution, service attribute, terminal attribute and network attribute.
It can be seen that the characteristics of for field of telecommunications, pass through user property, position attribution, service attribute, terminal attribute and network
Any one or more combination in attribute, can have been constructed according to the historical data of field of telecommunications for telecommunications you on
Feature database, in order to improve the data analysis efficiency of the data analysis task of field of telecommunications.
In the 6th kind of possible implementation of first aspect, this determines the data analysis task according to the output characteristic
Analysis result, including:The model with the type matching of the data analysis task is searched from model library;According to what is found
Model is handled the output characteristic, to obtain the analysis result.
It can be seen that can by building model library in advance, and historic task is configured during analysis model, with
It is convenient to receive and the type of the historic task is stored in the model library with the corresponding relation that its model configured has
Data analysis task when, the type according to the data analysis task directly can match corresponding with the type from model library
Model, and the output characteristic of eigenmatrix can be handled according to the model, eliminates and appoint again for the data analysis
The time of business allocation models, improve data analysis efficiency.
With reference to the 6th kind of implementation of first aspect, in the 7th kind of possible implementation, the model library includes
The model configured during the analysis of historic task, the historic task are completed data analysis task, and the history is appointed
The type of business is identical with the type of the data analysis task.
It can be seen that the model that the data analysis task of same type is configured is substantially similar, the model of historic task is stored in
It may be multiplexed, saved as data analysis task allocation models in model library, during the data analysis task for handling same type
Time.
Second aspect, the embodiments of the invention provide a kind of data analysis set-up, the device includes receiving unit, obtains list
Member, determining unit and processing unit, for performing described by any implementation of above-mentioned first aspect or first aspect
Method.
The third aspect, there is provided a kind of data analytics server, including memory and processor, the memory are used to store
Program, the processor are used for configuration processor, and when described program is performed, the processor is specifically used for performing above-mentioned first
Method described by any implementation of aspect or first aspect.
Fourth aspect, there is provided a kind of computer-readable medium, the computer-readable medium are used for store program codes, institute
State program code and include the finger for being used for realizing the method described by any implementation of above-mentioned first aspect or first aspect
Order.
It can be seen from above-mentioned technical proposal when receiving data analysis task, according to the class of the data analysis task
Type, eigenmatrix corresponding to the type can be determined from corresponding matrix storehouse, and can be by this feature matrix to the data
The data of analysis are handled needed for analysis task, it is seen then that are used for from eigenmatrix corresponding to the acquisition of matrix storehouse to a number
According to the processing of analysis task, and without being this data analysis task special configuration eigenmatrix, so as to eliminate needs originally
According to the demand of the data of required analysis and data analysis task come the time of configuration feature matrix, data analysis effect is improved
Rate.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, may be used also
To obtain other accompanying drawings according to these accompanying drawings.
Fig. 1 is a kind of method flow diagram of data analysing method provided in an embodiment of the present invention;
Fig. 2 is a kind of structural representation of field of telecommunications data analysis provided in an embodiment of the present invention;
Fig. 3 is the method flow diagram of another data analysing method provided in an embodiment of the present invention;
Fig. 4 is a kind of method flow diagram of field of telecommunications data analysing method provided in an embodiment of the present invention;
Fig. 5 is the method flow diagram of another data analysing method provided in an embodiment of the present invention;
Fig. 6 is a kind of structure drawing of device of data analysis set-up provided in an embodiment of the present invention;
Fig. 7 is a kind of hardware structure diagram of data analytics server provided in an embodiment of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Whole description, it is clear that described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other under the premise of creative work is not made
Embodiment, belong to the scope of protection of the invention.
Lifted with the importance of data, it is increasing to the data analysis requirements of data.The process of data analysis needs
The data of required analysis are handled using to eigenmatrix and model, so as to obtain analysis result.It is used for although having at present
The analysis software of data analysis, but this kind of analysis software does not pay close attention to each neck primarily directed to the General Use Analysis of data
Feature possessed by data in domain.
There is the species of data in some fields more changeable, in these fields, the data analysis task institute of same type
The data that need to be analyzed may all be not quite similar.The species for having data in some fields is more fixed, such as field of telecommunications, bank's neck
Domain etc., in these fields, the data of the data analysis required by task analysis of same type may be substantially similar.It can be seen that in number
According to the more fixed field of species in, if under can the eigenmatrix that the data analysis task once handled is configured be preserved
Come, need to handle if there is a data analysis task of same type afterwards, due to the data analysis task of same type
The data of required processing are substantially similar under this field, therefore the eigenmatrix preserved before will have in this data analysis
The possibility reused in task.
Therefore, the embodiments of the invention provide a kind of data analysing method and device, when receiving data analysis task,
According to the type of the data analysis task, eigenmatrix corresponding to the type can be determined from corresponding matrix storehouse, and can
Handled with the data analyzed by this feature matrix the data analysis required by task, it is seen then that corresponding to being obtained from matrix storehouse
Eigenmatrix comes for the processing to a data analysis task, and without being this data analysis task special configuration feature square
Battle array, so as to eliminate need originally according to the data of required analysis and the demand of data analysis task come configuration feature matrix when
Between, improve data analysis efficiency.
The embodiment of the present invention is mainly used in data class and more fixed, and changes less field, such as field of telecommunications.
When handling the data analysis task in this kind of field, can be determined according to the type of data analysis task from matrix storehouse and such
Eigenmatrix corresponding to type, this feature matrix determined can be used for handling, analyze data analysis required by task analysis
Data, and from the output characteristic of this feature Output matrix data analysis required by task.
In embodiments of the present invention, data analysis task can be that specific data are carried out in order to realize a demand
Specific data analysis, if data analysis task is off line data analysis task, the data of data analysis required by task analysis
To be off-line data, the data analysis carried out is also the data analysis for off-line data.The class of one data analysis task
Type can be the concrete analysis type of data analysis, the data type of required analyze data, the algorithm used in data analysis
With any one or more combination such as used eigenmatrix type.
Data analysis task has corresponding data to be analyzed, that is, the data of data analysis required by task analysis,
In data to be analyzed corresponding to data analysis task, a data can be individually value or a record.With table structure
Exemplified by tables of data, a data can be a record in tables of data.Data can have an attribute, and feature can be with
The set of the data of same alike result.By taking the tables of data of table structure as an example, a feature can be a row or data line in table
General name, this row or data line can have identical attribute, such as a tables of data includes 4 row, and this 4 row is respectively
" name ", " age ", " sex " and " height ", each row can include at least one data with same alike result, such as
Data " Zhang San ", " Li Si " can be included in feature " name " this row and " king five ", can be wrapped in feature " age " this row
Include data " 20 years old ", " 30 years old " and " 40 years old ".For " Zhang San ", " Li Si " and " for these three data of king five ", all with table
Show the attribute of name.For " 20 years old ", " 30 years old " and " 40 years old " these three data, all there is the attribute for representing the age.It is right
For this tables of data, there can be 4 features, be respectively " name ", " age ", " sex " and " height ", each feature
Include data " 20 years old ", " 30 years old " and " 40 years old ", feature " surname including the data with same alike result, such as feature " name "
Name " includes data " Zhang San ", " Li Si " and " king five ".
One data analysis task can carry out data analysis to the data of required analysis, and the data analysis can include pair
The data processing operations such as required analyze data classification, restructuring, conversion, computing, analysis.During data analysis is carried out, need
By the data input of required analysis into eigenmatrix, will pass through needed for the computing output data analysis task of eigenmatrix
Output characteristic.In embodiments of the present invention, the input content of an eigenmatrix and output content can be characterized.One
What feature eigenmatrix, which should input and what feature can be exported, is determined by the processing logic of this eigenmatrix.
Such as in the field of telecommunications, the processing logic of eigenmatrix can be telecommunication features application programming interface (Application
Programming Interface, API).By eigenmatrix come analyzed needed for processing data analysis task data when,
Need first by data recombination, be converted to and meet the input feature vector of this feature matrix.
Next the data analysing method that the embodiment of the present invention is provided is discussed in detail.Fig. 1 provides for the embodiment of the present invention
A kind of data analysing method method flow diagram, methods described includes:
S101:Receive data analysis task.
For example, data analysis task can be the specific data in order to realize a demand and be carried out to specific data
Analysis, the data analysis task can provide the data of required analysis, home type of data analysis task of required analysis etc.
The information content.Such as make an off-network project needs and be related to 8000 business support system (Business Support
System, BSS) tables of data progress data analysis, then it can include for the data analysis task of this off-network project to this
8000 BSS tables of data carry out data analysis, to realize the task of the analysis demand of the off-network project.
" the 8000 BSS numbers that now " data analysis can will be carried out to 8000 BSS tables of data " involved by this demand
According to table " as the data analysis required by task analysis data, and then for this data analysis task provide " 8000
The type such as BSS of data and its ownership included by BSS tables of data ", carries out special data analysis.Simultaneously as the project
For what is performed under off-network environment, therefore the number to required analysis can also be included in the information content that is provided of the data analysis task
According to the characteristic condition such as offline environment analyzed, thus to " 8000 BSS tables of data " carry out data analysis can be for
The data analysis that data under offline environment are carried out.
The data analysis task that data analysis task received by the embodiment of the present invention may belong in specific area, should
Specific area can be that data class is more fixed, change less field, such as telecommunications, the bank field.
S102:Obtain data to be analyzed corresponding to the data analysis task.
For example, data to be analyzed can be carried in the data analysis task, so as to obtain the data point
Obtained together during analysis task;Data to be analyzed can also be stored in the storage location indicated by the data analysis task,
After getting the data analysis task, the data of required analysis can be extracted from indicated storage location.Divide needed for obtaining
The mode present invention of the data of analysis does not limit, and can be by SQL (Structured Query
Language, SQL) the required data analyzed of extraction.
S103:Feature square corresponding with the type is determined from matrix storehouse according to the type of the data analysis task
Battle array.
For example, the type of data analysis task can be the concrete analysis type of data analysis, required analyze data
Data type, the algorithm used in data analysis and used eigenmatrix type etc. in it is any one or more
Combination.The type of a data analysis task can be determined according to specific application scenarios or demand.
So that a required analyze data is 8000 BSS tables, off-network project data analysis task as an example, according to difference
Scene or demand, the type of the data analysis task can be the concrete analysis type of data analysis:Off-network task analysis, from
And the eigenmatrix corresponding with off-network task analysis can be determined from matrix storehouse.Or the class of the data analysis task
Type can be the data type of required analyze data:BSS tables, it is corresponding with BSS tables so as to be determined from matrix storehouse
Eigenmatrix.Or the type of the data analysis task can be the concrete analysis type of data analysis:Off-network task analysis,
And the algorithm A used in data analysis, it is relative with off-network task analysis and algorithm A so as to be determined from matrix storehouse
The eigenmatrix answered.
Involved matrix storehouse can pre-establish in the embodiment of the present invention, be mainly used to storage feature matrix,
The eigenmatrix deposited can include the eigenmatrix corresponding to different types of data analysis task, and eigenmatrix can be used
In the data analyzed needed for data analysis task corresponding to processing.Eigenmatrix in matrix storehouse has and corresponding data point
Incidence relation between the type of analysis task, so as to be determined by the type of data analysis task from matrix storehouse with being somebody's turn to do
Type has the eigenmatrix of incidence relation.The eigenmatrix determined from matrix storehouse can be used for the number received to S101
According to the processing of analysis task, so as to avoid as the process of the data analysis task special configuration eigenmatrix, eliminate originally
Need to improve data analysis according to the data of required analysis and the demand of data analysis task come the time of configuration feature matrix
Efficiency.Generally, the field belonging to the data analysis task in matrix storehouse corresponding to eigenmatrix can be connect with S101
Field belonging to the data analysis task of receipts is identical, and belongs to data class and more fix, and changes less field, such as electricity
Letter field.
Spy corresponding with the type can be determined from matrix storehouse for the S103 type by data analysis task
Matrix is levied, determines the mode of eigenmatrix from matrix storehouse the invention provides several types according to data analysis task.
The first determination mode:
This determination mode determines eigenmatrix mainly by way of matching from matrix storehouse, in this determination
In mode, eigenmatrix that matrix storehouse includes can be the eigenmatrix that is configured during the analysis of historic task.Should
Historic task is completed data analysis task.
The situation that eigenmatrix is matched from matrix storehouse is configured equivalent to from the analysis process based on historic task
Data analysis task (i.e. data analysis task received by the S101) institute being currently received is determined in the eigenmatrix come
The situation for the eigenmatrix that need to be configured.
Determined out of matrix storehouse the historic task corresponding to eigenmatrix type and S101 received by data point
The type of analysis task is identical.The eigenmatrix determined is applied to the processing to the data analysis task received by S101
In, so as to reach the effect reused to eigenmatrix.If the eigenmatrix deposited in matrix storehouse is more comprehensive, eigenmatrix
The type of corresponding data analysis task is more, then when receiving new data analysis task, is directly repeated from matrix storehouse
The probability that the eigenmatrix preserved before use is handled is bigger.
By the first determination mode, the probability that its eigenmatrix deposited in advance is multiplexed from matrix storehouse can be increased,
And without to reconfigure feature matrix to the data analysis task received by S101, reducing and the data analysis task being entered
Time spent by row data analysis, and then improve the efficiency of data analysis.
Second of determination mode:
The embodiment of the present invention, can also be true by second except that can be multiplexed the eigenmatrix matched from matrix storehouse
Determine mode and determine eigenmatrix from matrix storehouse.The eigenmatrix that now matrix storehouse includes can be protected according in feature database
The feature deposited is configured with processing logic, and the feature preserved in this feature storehouse obtains according to constructed by historical data.Adopt
It can include being directed to matching from matrix storehouse according to the type of data analysis task with the scene of second of determination mode
The scene of corresponding eigenmatrix, this scene it is to be understood that by the first determination mode from matrix storehouse matching characteristic
During matrix, the eigenmatrix in matrix storehouse and corresponding to the type without the data analysis task is found.In matrix storehouse not
Eigenmatrix corresponding to type with the data analysis task mainly can have several situations to cause.A kind of such as situation
Under, before data analysis task is received by S101 and the untreated type identical data with the data analysis task are divided
Analysis task, therefore the eigenmatrix corresponding to the type of the data analysis task is not preserved in matrix storehouse;It is such as another
In the case of, although pre-treatment, analyzed type identical data analysis task with S101 received data analysis tasks,
But the eigenmatrix of configuration is not saved in matrix storehouse.
In second of determination mode, the eigenmatrix determined out of matrix storehouse is according to the feature preserved in feature database
The eigenmatrix configured with processing logic.The feature preserved in this feature storehouse can obtain according to constructed by historical data
, the historical data is the data belonged in a field, and this field can be field of telecommunications.The historical data in one field
Can be by collecting acquisition in advance.The attribute that can be directed to entrained by historical data constructs the feature preserved in feature database.
Such as during one feature database related to field of telecommunications of structure, can be in the data collected by field of telecommunications, according to it
Each entrained attribute is clustered, so as to construct a feature database for belonging to field of telecommunications.Data institute in field of telecommunications
The attribute of carrying can include user property, position attribution, service attribute, terminal attribute and network attribute in any one or
Multinomial combination.Such as user property, the data with user property are sorted out, obtain the data of feature " user "
Set.
In addition, during construction feature storehouse, it is also possible to knowledge mapping technology can be used, belong to telecommunications to what is be collected into
The data in field such as are screened, classified at the processing operation, so as to which the data being collected into are classified into the data with different attribute
Set, forms multiple features, and then realizes the generation automation of Partial Feature in feature database.
And the processing logic preserved in feature database can be it is built-up according to the relation between different characteristic in feature database,
Such as how obtaining feature c mode according to feature a, feature b can be a kind of processing logic, it is seen then that feature and processing logic
Between there is clear and definite incidence relation.
For a field, the feature database related to this field can be constructed according to the historical data in this field.
Afterwards, receive this field need carry out data analysis data analysis task when, can utilize this feature storehouse in include
Feature and processing logic, configuration obtain eigenmatrix corresponding with the type of the data analysis task.
, can be according to the processing logic preserved in this feature storehouse when using the feature database configuration feature matrix built in advance
With feature, to generate the eigenmatrix needed for a data analysis task type.Such as a data in field of telecommunications
Analysis task, according to the type of the data analysis task, determined from feature database to realize the data analysis task, the number
According to feature a, b and c of eigenmatrix required input in analysis task, the feature d and e of required output, and determined from feature database
Go out how to be determined according to feature a, b and c to obtain feature d and e processing logic x, so as to according to feature a, b, c, d and e, and place
Reason logic x is configured to obtain the eigenmatrix of the data analysis required by task, and this feature matrix can be stored in matrix storehouse.
In addition, during according to feature database configuration feature matrix, it is also possible to can use and arrive one or more algorithms, calculate
Method can also belong to a part for processing logic.Detailed process is included in selects required spy using the processing logic determined
Levy after formed data acquisition system is arranged such as " age " one, the feature of selection is handled by the algorithm determined, then by handling
Logic utilizes feature after treatment to automatically generate the eigenmatrix F that this data analysis needs to configure, and is stored in matrix storehouse
In.Wherein, the algorithm that may be used in data analysis process can be machine learning (Machine Learning, ML),
The one or more such as time series analysis (Time series analysis), descriptive research (descriptive study)
Combination, and the algorithm that configuration feature matrix uses can be determined by the type of this data analysis task.The present invention
Described in data analysis process can include the data of the analysis needed for get to from eigenmatrix obtain output characteristic it
Between process.
With reference to accompanying drawing 2 to how to be illustrated by feature database configuration feature matrix.In scene shown in Fig. 2, pass through number
The scheme by feature database configuration feature matrix is realized according to module in analysis platform.Can construction feature storehouse in advance, specific structure
The process of building can be according to the data for belonging to field of telecommunications and/or knowledge engine (Knowledge Engine) module being collected into
Feature of the structure with different attribute, then according to the relation between different characteristic, obtain the telecommunications APIs (Telecom shown in Fig. 2
APIs) module, and then by feature and Telecom APIs module composition characteristics storehouse;Wherein, Telecom APIs modules can be with
Interactive mode explores environment (Interactive Exploratory Environment) module and carries out information exchange at any time.Due to
During configuration feature matrix, it may use and arrive algorithm, therefore algorithm (Algorithms) module is set, and should
Algorithms modules can also enter row information friendship at any time with Interactive Exploratory Environment modules
Mutually.
Received in Interactive Exploratory Environment modules from telecommunication features and derive (Telecom
Feature Derivation) send one of module belong to field of telecommunications data analysis task when, Interactive
Exploratory Environment modules can be searched first from matrix storehouse, be determined whether corresponding with the type
Eigenmatrix is present;If being not present, Interactive Exploratory Environment modules can by with feature
Information exchange between storehouse, it can determine that required processing is patrolled from the Telecom APIs modules of feature database according to the type
Volume, and required feature is chosen, and the eigenmatrix configured so as to generate this data analysis task to need, Interactive
Exploratory Environment modules can continue follow-up data analysis according to this feature matrix, wherein, from
It can be family/job site identification (Home Zone), operation that required processing logic is determined in Telecom APIs modules
Any one in system banner (OS Identification) and content identification (Content Identification) is more
Individual combination.
In addition, during required processing logic is determined, it is also possible to can be due to the class of this data analysis task
Type, one or more algorithms needed for this are determined from Algorithms modules, it is raw so as to combine the algorithm determined
Cost time data analysis needs the eigenmatrix configured.
By second of determination mode, can be constructed beforehand through the data in a field related to this field
Feature database, so as to need for eigenmatrix corresponding to the configuration of data analysis task when, can be with direct basis feature database
The feature and processing logic having had determined are configured, so as to save being consumed for configuration feature matrix to a certain extent
Time.
S104:The data are handled according to the processing logic of the eigenmatrix, to obtain the eigenmatrix
Output characteristic.
For example, after configuration obtains eigenmatrix corresponding with this data analysis task type, can utilize should
The data for the data analysis required by task analysis that processing logic contained by eigenmatrix obtains to S102 are handled accordingly
Operation, and then get the output characteristic of this feature matrix.
In the number of the data analysis required by task analysis obtained using the processing logic contained by this feature matrix to S102
During corresponding processing operation is carried out, if the input content that will enter into this feature matrix does not meet this feature matrix
The input requirements of contained processing logic, then need first to enter the data that the data analysis required by task that S102 is got is analyzed
Some data processing operations of row, such as data recombination, conversion, and then obtain the input content for meeting input requirements;It will meet
After the input content such as input feature vector of input requirements are input to this feature matrix, this feature matrix can passes through contained processing
Logic carries out computing, the output content such as output characteristic of the final output data analysis required by task to input feature vector.
S105:The analysis result of the data analysis task is determined according to the output characteristic.
For example, after the output characteristic of this data analysis required by task is determined, can be defeated to this by model
Go out feature and carry out computing, to obtain the analysis result of the data analysis task.
According to above-described embodiment as can be seen that the data analysis can be got by the data analysis task received
The data of this required analysis of task and the type of the data analysis task, so as to the class according to the data analysis task
Type determines eigenmatrix corresponding with the type from matrix storehouse, and can be by the processing logic of this feature matrix to required point
The data of analysis are handled, and obtain the output characteristic of this feature matrix, and then are realized from matrix storehouse corresponding to direct acquisition
Eigenmatrix comes for the processing to a data analysis task, and without being this data analysis task special configuration feature square
Battle array, eliminate needed according to the data of required analysis and the demand of data analysis task come the time of configuration feature matrix originally,
Improve data analysis efficiency.
, it is necessary to the data analysis step implemented during analyze data needed for basis obtains output characteristic from eigenmatrix
It is rapid relatively complicated, such as the input feature vector needed for eigenmatrix, the processing logic in eigenmatrix are converted data to input
Feature carries out the data analysis steps such as computing.By the data analyzed needed for this part from acquisition to from feature in the embodiment of the present invention
The data analysis step obtained in matrix between output characteristic is referred to as Feature Engineering.
Due to more being fixed in data class, change in less field such as field of telecommunications, Feature Engineering is in same type
It is actually similar, therefore more fixed in data class in data analysis task, changes in less field, except can
To be multiplexed outside the eigenmatrix configured in same type data analysis task, can also be multiplexed in same type data analysis task
Feature Engineering.
On the basis of embodiment corresponding to Fig. 1, for step S104, the embodiment of the present invention additionally provides one kind and is used for base
The specific method of the output characteristic of eigenmatrix is obtained in historic task, as shown in figure 3, this method includes:
S301:Feature work corresponding with the type is determined from Feature Engineering storehouse according to the type of data analysis task
Journey, the Feature Engineering include the data analysis from the data analyzed needed for acquisition to obtaining output characteristic from eigenmatrix
Process.
For example, the Feature Engineering storehouse involved by the embodiment of the present invention can be built in advance, for preserving once
Data analysis of the data through being analyzed needed for treated historic task from acquisition to obtaining output characteristic from eigenmatrix
The corresponding relation having between process and the type of the historic task and its Feature Engineering.
Feature Engineering can also include the preprocessing process to analyze data needed for data analysis task, the mistake of the pretreatment
Journey include data deduplication, sampling of data and it is data-optimized in any one or more combination.Such as data prediction can wrap
Include to reduce amount of calculation, the sampling of data carried out to required analyze data, will therefrom extract a part of data and be used for data
Analysis, or including carrying out data deduplication to required analyze data, duplicate data is removed to improve data analysis accuracy, or
Person includes carrying out required analyze data advance feature clustering, to obtain the behaviour such as input feature vector required for eigenmatrix
Make.
Due to Feature Engineering in the data analysis task of same type it is more similar, therefore can be according to data analysis task
Type directly matches the Feature Engineering for having corresponding relation with its type from the Feature Engineering storehouse of structure, so as to repeat
Corresponding data analysis process is carried out using this feature engineering.
S302:The data analysis process included according to the Feature Engineering, according to the processing logic of the eigenmatrix
The data are handled, obtain the output characteristic of the eigenmatrix.
, can be according to the type of the data analysis task from spy for example, when receiving a data analysis task
Sign engineering determines corresponding Feature Engineering in storehouse, and the Feature Engineering determined can be employed in the historic task of same type
Feature Engineering.According to this feature engineering, the data that can be directed to data analysis required by task analysis are pre-processed, although
The data analysis task and the data content that the historic task is analyzed are different, but processing step is essentially identical, therefore can
With on the basis of Feature Engineering, accordingly by the data content handled needed for each step according to the data analysis required by task
Analyze data is adjusted, you can by this feature engineer applied determined in the processing to the data analysis task, and energy
It is enough that the output characteristic of the data analysis required by task is obtained according to eigenmatrix.
4 it is further illustrated below in conjunction with the accompanying drawings, data analysis task is being obtained from data source modules, and from the data
, can be by shown in Fig. 4 if the data count amount of required analysis is excessive after the data of analysis needed for being got in analysis task
Sampling of data module extract partial data from the data of required analysis, for analysis use;Again by data prediction mould
Block carries out other pretreatment operations to the partial data extracted, as data deduplication, data-optimized etc. are any one or more
Combination.Afterwards, telecommunications API (Telecom API) modules are entered again to the data obtained after data preprocessing module is handled
Row processing operation, obtains output characteristic, and is stored in eigenmatrix as shown in Figure 4 (Feature Matrix) module, also simultaneously
It will be stored in from the data for getting required analysis to the data analysis process obtaining output characteristic from eigenmatrix shown in Fig. 4
Feature Engineering module in, as the Feature Engineering of the data analysis task, then the feature work by feature management module to acquisition
Cheng Jinhang Screening Treatments and management, can be direct so as to when obtaining same type of data analysis task from data source modules
The Feature Engineering preserved in repeatedly used features management module, to obtain the output characteristic of its eigenmatrix, and then save data point
Time spent by analysis process.
According to above-described embodiment as can be seen that can analyzed by pre-establishing Feature Engineering storehouse, and by historic task
During the corresponding pass that has with the Feature Engineering that it is configured of the type of the Feature Engineering that is configured and the historic task
System is maintained in this feature engineering storehouse, convenient in the data analysis task received, can be according to the data analysis task
Type Feature Engineering corresponding with the type is directly matched from Feature Engineering storehouse, eliminate and configure cumbersome data analysis
The time of step, improve data analysis efficiency.
, it is necessary to which the output characteristic that eigenmatrix is exported leads to during data analysis is carried out to data analysis task
Cross model and carry out computing, and then obtain the analysis result of the data analysis task.In embodiments of the present invention, model can be root
According to the common caused regular entity of set algorithm, algorithm parameter and training data, for mould used in data analysis task
Type, the input content of the model can include the output characteristic of eigenmatrix used in the data analysis task, the model
Output content can be the analysis result of the data analysis task.
It is more fixed in data class, change in less field, the model that the data analysis task of same type is configured
It is substantially similar, therefore in order to save as the time of data analysis task allocation models, in the embodiment of the present invention, there is provided preserve history
The model configured in task, the model of the data analysis task of the same type preserved in model library can be directly multiplexed, is entered
And output characteristic can be handled according to the model, obtain analysis result.
In another embodiment, as shown in figure 5, step S105 can include:
S501:The model with the type matching of the data analysis task is searched from model library.
For example, model library can be built in advance, for historic task is configured during analysis model,
And the corresponding relation having between the type of the historic task and its model configured is preserved, so as to follow-up mutually similar
The data analysis task of type its type can find corresponding model with direct basis from model library, and then realize to model
Reuse eliminate again be the data analysis task allocation models time.For a data analysis task, if energy
Model is enough determined from model library according to the type of the data analysis task, then the type of historic task corresponding to the model
Type with the data analysis task can be with identical.
S502:The output characteristic is handled according to the model found, to obtain the analysis result.
For example, as shown in figure 4, after Feature Engineering module in Fig. 4 outputs output characteristic, model construction module
Model corresponding with the type can be matched from the model prestored according to the type of data analysis task, and utilizes matching
Obtained model carries out computing to the output characteristic, so as to obtain corresponding analysis result, avoids as the data analysis task
Reconfigure the operating process of model;Meanwhile model construction module can also be by the mode input of preservation to model (Model) mould
In block, related training, Jin Erti are carried out to the model of input using prediction (Prediction) module by model (Model) module
The precise degrees of high model., can be to the model in order to increase the repeat usage of the model preserved in the model library of structure
The model preserved in storehouse carries out one or more combination operations such as model training, model evaluation, model modification.As shown in Fig. 2 it is
Ensure that interactive environment (Interactive Exploratory Environment) module explored is receiving telecommunication features
, can be according to the data when deriving the data analysis task that (Telecom Feature Derivation) module is sent
The type of analysis task matches correspondingly from model life cycle management (Model Lifecycle Management) module
Model, the model that is included of Model Lifecycle Management modules can also be utilized to establish (Build Model)
Submodule, model checking (Validate Model) submodule, model modification (Update Model) submodule and model prison
Any one or more submodules in control (Deploy Monitor Model) submodule are managed operation to model, and then
Make model more accurate, also improve the repeat usage of model.
As can be seen that can be by building model library in advance, and by historic task in analysis process according to above-described embodiment
The type of middle configured model and the historic task is stored in the mould with the corresponding relation that its model configured has
In type storehouse, convenient in the data analysis task received, the type according to the data analysis task can be directly from model library
In match model corresponding with the type, and the output characteristic of eigenmatrix can be handled according to the model, saved
Again it is the time of the data analysis task allocation models, improves data analysis efficiency.
Fig. 6 is a kind of structure drawing of device of data analysis set-up provided in an embodiment of the present invention, the data analysis set-up
600 include:Receiving unit 601, acquiring unit 602, determining unit 603 and processing unit 604, wherein:
Receiving unit 601, for receiving data analysis task.
Acquiring unit 602, for obtaining data to be analyzed corresponding to the data analysis task.
Determining unit 603, for determining the type pair from matrix storehouse according to the type of the data analysis task
The eigenmatrix answered.
Processing unit 604, the data are handled for the processing logic according to the eigenmatrix, to obtain
State the output characteristic of eigenmatrix.
Determining unit 603 is additionally operable to determine the analysis result of the data analysis task according to the output characteristic.
Optionally, the matrix storehouse is included in the eigenmatrix configured during the analysis of historic task, the history
Task is completed data analysis task, and the type of the historic task is identical with the type of the data analysis task.
Optionally, determining unit 603 is additionally operable to be determined and the type pair from Feature Engineering storehouse according to the type
The Feature Engineering answered, the Feature Engineering are included from the data analyzed needed for acquisition to obtaining output characteristic from eigenmatrix
Data analysis process;
Processing unit 604 is additionally operable to the data analysis process included according to the Feature Engineering, according to the feature square
The processing logic of battle array is handled the data, obtains the output characteristic of the eigenmatrix.
Optionally, the data analysis process also includes the process pre-processed to the data of required analysis, described pre-
The process of processing include data deduplication, sampling of data and it is data-optimized in any one or more combination.
Optionally, the eigenmatrix that the matrix storehouse includes is to be patrolled according to the feature preserved in feature database with processing
Collect and configure what is obtained, the feature preserved in the feature database obtains according to constructed by historical data.
Optionally, the historical data belongs to field of telecommunications, then the feature preserved in the feature database is according to the neck
Obtained in domain constructed by data, including:
The feature preserved in the feature database is based on the attribute entrained by data, and the data in the field of telecommunications are entered
Row structure obtains, and the attribute includes any in user property, position attribution, service attribute, terminal attribute and network attribute
One or more combinations.
Optionally, determining unit 603 is additionally operable to search the type matching with the data analysis task from model library
Model;The output characteristic is handled according to the model found, to obtain the analysis result.
Optionally, the model library is included in the model configured during the analysis of historic task, the historic task
It is identical with the type of the data analysis task for completed data analysis task, the type of the historic task.
In one embodiment, receiving unit 601 can be network interface or API, be submitted for receiving application program
Data analysis task;Acquiring unit 602, determining unit 603 and processing unit 604 can be by one or more processors Lai real
Existing, processor is specifically as follows general processor, or digital signal processor (DSP), application specific integrated circuit (ASIC),
Field programmable gate array (FPGA) or other programmable logic components.
It should be noted that realize that details may refer to the corresponding implementation of Fig. 1, Fig. 3 and Fig. 5 in embodiment corresponding to Fig. 6
The related description of example, is repeated no more here.
Fig. 7 is a kind of hardware architecture diagram of data analytics server provided in an embodiment of the present invention, and the data are divided
Analysis server 700 includes memory 701 and receiver 702, and connects respectively with the memory 701 and the receiver 702
The processor 703 connect, the memory 701 are used to store batch processing instruction, and the processor 703 is used to call the storage
The programmed instruction that device 701 stores performs following operation:
Trigger the receiver 702 and receive data analysis task;
Trigger the receiver 702 and obtain data to be analyzed corresponding to the data analysis task;
The eigenmatrix according to corresponding to the type of the data analysis task determines the type from matrix storehouse;
The data are handled according to the processing logic of the eigenmatrix, to obtain the output of the eigenmatrix
Feature;
The analysis result of the data analysis task is determined according to the output characteristic.
Alternatively, the programmed instruction that the processor 703 is additionally operable to call the memory 701 to store performs Fig. 1, Fig. 3
With Fig. 5 corresponding to other steps in embodiment.
In one embodiment, the processor 703 can be central processing unit (Central Processing Unit,
CPU), the memory 701 can be that the inside of random access memory (Random Access Memory, RAM) type is deposited
Reservoir, the receiver 702 can include General Physics interface, the physical interface can be ether (Ethernet) interface or
Asynchronous transfer mode (Asynchronous Transfer Mode, ATM) interface.The processor 703, receiver 702 and deposit
Reservoir 701 can be integrated into one or more independent circuits or hardware, such as:Application specific integrated circuit (Application
Specific Integrated Circuit, ASIC).
One of ordinary skill in the art will appreciate that:Realizing all or part of step of above method embodiment can pass through
Programmed instruction related hardware is completed, and foregoing routine can be stored in a computer read/write memory medium, and the program exists
During execution, execution the step of including above method embodiment;And foregoing storage medium can be in following media at least one
Kind:Read-only storage (English:Read-only memory, abbreviation:ROM), RAM, magnetic disc or CD etc. are various to store
The medium of program code.
It should be noted that each embodiment in this specification is described by the way of progressive, each embodiment it
Between identical similar part mutually referring to what each embodiment stressed is the difference with other embodiment.
For equipment and system embodiment, because it is substantially similar to embodiment of the method, so describe fairly simple,
The relevent part can refer to the partial explaination of embodiments of method.Equipment and system embodiment described above is only schematic
, wherein as the unit that separating component illustrates can be or may not be physically separate, be shown as unit
Part can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple networks
On unit.Some or all of module therein can be selected to realize the purpose of this embodiment scheme according to the actual needs.
Those of ordinary skill in the art are without creative efforts, you can to understand and implement.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto,
Any one skilled in the art the invention discloses technical scope in, the change or replacement that can readily occur in,
It should all be included within the scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims
It is defined.
Claims (16)
1. a kind of data analysing method, it is characterised in that methods described includes:
Receive data analysis task;
Obtain data to be analyzed corresponding to the data analysis task;
The eigenmatrix according to corresponding to the type of the data analysis task determines the type from matrix storehouse;
The data are handled according to the processing logic of the eigenmatrix, to obtain the output of eigenmatrix spy
Sign;
The analysis result of the data analysis task is determined according to the output characteristic.
2. according to the method for claim 1, it is characterised in that the matrix storehouse is included in during the analysis of historic task
The eigenmatrix configured, the historic task are completed data analysis task, the type of the historic task with it is described
The type of data analysis task is identical.
3. according to the method for claim 2, it is characterised in that also include:
Feature Engineering corresponding with the type is determined from Feature Engineering storehouse according to the type, the Feature Engineering includes
From data analysis process of the data analyzed needed for acquisition to obtaining output characteristic from eigenmatrix;
The processing logic according to the eigenmatrix is handled the data, and the output for obtaining the eigenmatrix is special
Sign, including:
The data analysis process included according to the Feature Engineering, according to the processing logic of the eigenmatrix to the data
Handled, obtain the output characteristic of the eigenmatrix.
4. according to the method for claim 3, it is characterised in that the data analysis process also includes the number to required analysis
According to the process pre-processed, the process of the pretreatment includes data deduplication, sampling of data and data-optimized middle any one
Or multinomial combination.
5. according to the method for claim 1, it is characterised in that the eigenmatrix that the matrix storehouse includes is basis
The feature preserved in feature database configures what is obtained with processing logic, and the feature preserved in the feature database is according to historical data
It is constructed to obtain.
6. according to the method for claim 5, it is characterised in that the historical data belongs to field of telecommunications, then the feature
The feature preserved in storehouse obtains according to constructed by data in the field, including:
The feature preserved in the feature database is based on the attribute entrained by data, and structure is carried out to the data in the field of telecommunications
Build what is obtained, the attribute includes any one in user property, position attribution, service attribute, terminal attribute and network attribute
Or multinomial combination.
7. according to the method for claim 1, it is characterised in that described that the data analysis is determined according to the output characteristic
The analysis result of task, including:
The model with the type matching of the data analysis task is searched from model library;
The output characteristic is handled according to the model found, to obtain the analysis result.
8. according to the method for claim 7, it is characterised in that the model library is included in during the analysis of historic task
The model configured, the historic task are completed data analysis task, the type of the historic task and the data
The type of analysis task is identical.
9. a kind of data analysis set-up, it is characterised in that described device includes receiving unit, acquiring unit, determining unit and place
Manage unit:
The receiving unit, for receiving data analysis task;
The acquiring unit, for obtaining data to be analyzed corresponding to the data analysis task;
The determining unit, corresponding to determining the type from matrix storehouse according to the type of the data analysis task
Eigenmatrix;
The processing unit, the data are handled for the processing logic according to the eigenmatrix, with described in acquisition
The output characteristic of eigenmatrix;
The determining unit is additionally operable to determine the analysis result of the data analysis task according to the output characteristic.
10. device according to claim 9, it is characterised in that the matrix storehouse is included in the analysis process of historic task
Middle configured eigenmatrix, the historic task are completed data analysis task, the type of the historic task and institute
The type for stating data analysis task is identical.
11. device according to claim 10, it is characterised in that the determining unit is additionally operable to according to the type from spy
Feature Engineering corresponding with the type is determined in sign engineering storehouse, the Feature Engineering is included from the data analyzed needed for acquisition
Data analysis process between obtaining output characteristic from eigenmatrix;
The processing unit is additionally operable to the data analysis process included according to the Feature Engineering, according to the eigenmatrix
Processing logic is handled the data, obtains the output characteristic of the eigenmatrix.
12. device according to claim 11, it is characterised in that the data analysis process also includes to required analysis
The process that data are pre-processed, the process of the pretreatment include data deduplication, sampling of data and it is data-optimized in it is any one
Item or multinomial combination.
13. device according to claim 9, it is characterised in that the eigenmatrix that the matrix storehouse includes is root
Configure what is obtained according to the feature preserved in feature database and processing logic, the feature preserved in the feature database is according to history number
Obtained according to constructed.
14. device according to claim 13, it is characterised in that the historical data belongs to field of telecommunications, then the spy
The feature preserved in sign storehouse obtains according to constructed by data in the field, including:
The feature preserved in the feature database is based on the attribute entrained by data, and structure is carried out to the data in the field of telecommunications
Build what is obtained, the attribute includes any one in user property, position attribution, service attribute, terminal attribute and network attribute
Or multinomial combination.
15. device according to claim 9, it is characterised in that the determining unit be additionally operable to from model library search with
The model of the type matching of the data analysis task;The output characteristic is handled according to the model found, with
Obtain the analysis result.
16. device according to claim 15, it is characterised in that the model library is included in the analysis process of historic task
Middle configured model, the historic task are completed data analysis task, the type of the historic task and the number
It is identical according to the type of analysis task.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210254763.3A CN114791927A (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
CN201610854748.7A CN107871055B (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610854748.7A CN107871055B (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210254763.3A Division CN114791927A (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107871055A true CN107871055A (en) | 2018-04-03 |
CN107871055B CN107871055B (en) | 2022-03-29 |
Family
ID=61750877
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210254763.3A Pending CN114791927A (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
CN201610854748.7A Active CN107871055B (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210254763.3A Pending CN114791927A (en) | 2016-09-27 | 2016-09-27 | Data analysis method and device |
Country Status (1)
Country | Link |
---|---|
CN (2) | CN114791927A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109299178A (en) * | 2018-09-30 | 2019-02-01 | 北京九章云极科技有限公司 | A kind of application method and data analysis system |
CN109740774A (en) * | 2019-02-28 | 2019-05-10 | 中国公路工程咨询集团有限公司 | The modification method and electronic equipment in maintenance of surface measure library |
CN110163380A (en) * | 2018-04-28 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Data analysing method, model training method, device, equipment and storage medium |
CN110275880A (en) * | 2019-05-21 | 2019-09-24 | 阿里巴巴集团控股有限公司 | Data analysing method, device, server and readable storage medium storing program for executing |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101021897A (en) * | 2006-12-27 | 2007-08-22 | 中山大学 | Two-dimensional linear discrimination human face analysis identificating method based on interblock correlation |
CN101324923A (en) * | 2008-08-05 | 2008-12-17 | 北京中星微电子有限公司 | Method and apparatus for extracting human face recognition characteristic |
CN101409643A (en) * | 2008-11-24 | 2009-04-15 | 北京中创信测科技股份有限公司 | Method, apparatus and corresponding system for analyzing telecom network modeling |
CN101685458A (en) * | 2008-09-27 | 2010-03-31 | 华为技术有限公司 | Recommendation method and system based on collaborative filtering |
CN102184250A (en) * | 2011-05-24 | 2011-09-14 | 东华大学 | Garment fabric sample retrieving method based on colored image matching |
US20130191186A1 (en) * | 2012-01-24 | 2013-07-25 | International Business Machines Corporation | System, method and computer program for capturing relationships between business outcomes, persons and technical assets |
CN103268362A (en) * | 2013-06-08 | 2013-08-28 | 国家电网公司 | Auxiliary design method of virtual terminals on basis of general template and key character matching |
CN104184589A (en) * | 2014-08-26 | 2014-12-03 | 重庆邮电大学 | Identity authentication method, terminal device and system |
-
2016
- 2016-09-27 CN CN202210254763.3A patent/CN114791927A/en active Pending
- 2016-09-27 CN CN201610854748.7A patent/CN107871055B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101021897A (en) * | 2006-12-27 | 2007-08-22 | 中山大学 | Two-dimensional linear discrimination human face analysis identificating method based on interblock correlation |
CN101324923A (en) * | 2008-08-05 | 2008-12-17 | 北京中星微电子有限公司 | Method and apparatus for extracting human face recognition characteristic |
CN101685458A (en) * | 2008-09-27 | 2010-03-31 | 华为技术有限公司 | Recommendation method and system based on collaborative filtering |
CN101409643A (en) * | 2008-11-24 | 2009-04-15 | 北京中创信测科技股份有限公司 | Method, apparatus and corresponding system for analyzing telecom network modeling |
CN102184250A (en) * | 2011-05-24 | 2011-09-14 | 东华大学 | Garment fabric sample retrieving method based on colored image matching |
US20130191186A1 (en) * | 2012-01-24 | 2013-07-25 | International Business Machines Corporation | System, method and computer program for capturing relationships between business outcomes, persons and technical assets |
CN103268362A (en) * | 2013-06-08 | 2013-08-28 | 国家电网公司 | Auxiliary design method of virtual terminals on basis of general template and key character matching |
CN104184589A (en) * | 2014-08-26 | 2014-12-03 | 重庆邮电大学 | Identity authentication method, terminal device and system |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110163380A (en) * | 2018-04-28 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Data analysing method, model training method, device, equipment and storage medium |
CN110163380B (en) * | 2018-04-28 | 2023-07-07 | 腾讯科技(深圳)有限公司 | Data analysis method, model training method, device, equipment and storage medium |
CN109299178A (en) * | 2018-09-30 | 2019-02-01 | 北京九章云极科技有限公司 | A kind of application method and data analysis system |
CN109299178B (en) * | 2018-09-30 | 2020-01-14 | 北京九章云极科技有限公司 | Model application method and data analysis system |
CN109740774A (en) * | 2019-02-28 | 2019-05-10 | 中国公路工程咨询集团有限公司 | The modification method and electronic equipment in maintenance of surface measure library |
CN109740774B (en) * | 2019-02-28 | 2021-07-30 | 中国公路工程咨询集团有限公司 | Correction method of pavement maintenance measure library and electronic equipment |
CN110275880A (en) * | 2019-05-21 | 2019-09-24 | 阿里巴巴集团控股有限公司 | Data analysing method, device, server and readable storage medium storing program for executing |
Also Published As
Publication number | Publication date |
---|---|
CN107871055B (en) | 2022-03-29 |
CN114791927A (en) | 2022-07-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Bi et al. | A big data clustering algorithm for mitigating the risk of customer churn | |
CN107766511A (en) | Intelligent answer method, terminal and storage medium | |
CN106682097A (en) | Method and device for processing log data | |
CN108259371A (en) | A kind of network flow data analysis method and device based on stream process | |
CN110334274A (en) | Information-pushing method, device, computer equipment and storage medium | |
CN110598070B (en) | Application type identification method and device, server and storage medium | |
CN106022708A (en) | Method for predicting employee resignation | |
CN106982150A (en) | A kind of mobile Internet user behavior analysis method based on Hadoop | |
CN102521411A (en) | Formula and formula data processing device and formula and formula data processing method | |
CN104796300B (en) | A kind of packet feature extracting method and device | |
CN104423968B (en) | It designs the method for service logic, execute its server and storage medium | |
CN107871055A (en) | A kind of data analysing method and device | |
CN109033277A (en) | Class brain system, method, equipment and storage medium based on machine learning | |
CN106991577A (en) | A kind of method and device for determining targeted customer | |
CN110147657A (en) | A kind of user right configuration method and device | |
CN107908796A (en) | E-Government duplicate checking method, apparatus and computer-readable recording medium | |
CN111815432A (en) | Financial service risk prediction method and device | |
CN111931809A (en) | Data processing method and device, storage medium and electronic equipment | |
CN110334119A (en) | A kind of data correlation processing method, device, equipment and medium | |
CN107748739A (en) | A kind of extracting method and relevant apparatus of short message text masterplate | |
CN109828790A (en) | A kind of data processing method and system based on Shen prestige isomery many-core processor | |
CN110378739B (en) | Data traffic matching method and device | |
CN111178701A (en) | Risk control method and device based on feature derivation technology and electronic equipment | |
CN110472246A (en) | Work order classification method, device and storage medium | |
Śniegula et al. | Study of machine learning methods for customer churn prediction in telecommunication company |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |