CN103544299B - A kind of construction method of business intelligence cloud computing system - Google Patents
A kind of construction method of business intelligence cloud computing system Download PDFInfo
- Publication number
- CN103544299B CN103544299B CN201310530032.8A CN201310530032A CN103544299B CN 103544299 B CN103544299 B CN 103544299B CN 201310530032 A CN201310530032 A CN 201310530032A CN 103544299 B CN103544299 B CN 103544299B
- Authority
- CN
- China
- Prior art keywords
- data
- variable
- algorithm
- module
- user
- 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.)
- Active
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
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention is the system constituting method of a kind of business intelligence cloud computing.System mainly includes " data review module ", " variable analysis module " and " algorithm general program module "；User enters Web Server or application program by browser and connects APP Server, selection algorithm, submits data to by prescribed form；The data that user is submitted to by " data review module " check；The algorithm that " variable analysis module " selects according to user and the requirement to data form, the data submitting user to are analyzed determining variable parameter；The automatic founding mathematical models of variable parameter determined by " algorithm general program module " basis calculates.Use the business intelligence cloud computing system that the present invention builds, only it is to be understood that what the algorithm of data mining can do, submit data on request to, system just can calculate by founding mathematical models automatically, can also be directly embedded in application program and carry out data mining, realize the seamless connection of data mining and application program, it is simple to apply and popularize.
Description
Technical field
The present invention relates to the construction method of a kind of business intelligence cloud computing system, belong to big data, business
Intelligence, data mining and field of cloud calculation.
Background technology
The value of big data is the knowledge containing in data, and how from data, Extracting Knowledge is big
Data, the core of business intelligence.Though having the systems such as SAS, SPSS, MATLAB to count at present
According to excavation, but there is layman and be difficult with, be not easy to embed the problem such as application program of user,
Not only need user to grasp data mining mathematical theory, also want input variable to describe and certain mathematical table
Reach formula, the language (such as: R language etc.) that even GPRS is special.
Summary of the invention
For solving the problems referred to above, the present invention proposes the construction method of a kind of business intelligence cloud computing system,
The system built by the present invention, it is not necessary to user grasps data mining theories, inputting mathematical expression formula, change
Amount describes, without the language that user learning is special, as long as user knows that the algorithm that system comprises can do
What, selection algorithm, by regulation submit to data, system just can automatically analyze variable, set up mathematical modulo
Type calculates, it is simple to the universal and application of data mining technology, and can be easy to embed application
Program, it is achieved data mining and the seamless connection of application program.
It is an object of the invention to be achieved through the following technical solutions: a kind of business intelligence cloud computing system
Construction method, Internet or LAN sets up a Web Server or APP Server, its
It is characterised by:
System mainly includes " data review module ", " variable analysis module " and " algorithm general program
Module "；
" data review module " is used for checking data, and the algorithm selected according to user and algorithm are to data
The requirement of form, the data that user is submitted to, if meet the data form that algorithm specifies and check；
" variable analysis module " determines variable for analytical data, the algorithm selected according to user and calculation
The data that user is submitted to by method by the requirement of data form are analyzed, and determine have how many variablees, changes
The variable parameters such as the character of amount and the span of variable；
" algorithm general program module " is used for automatic founding mathematical models and calculating, is some in module
The algorithm general program write, but uncertain have taking of how many variablees, the character of variable and variable
The variable parameters such as value scope, the most uncertain concrete mathematical model, only algorithm flow, according to " becoming
Component analysis module " determined by variable parameter, the automatic founding mathematical models of system calculates；
System flow is: user enters Web Server by browser or application program connects APP
Server, selection algorithm, the data form algorithmically specified submission data, " data review module " is right
The data that user submits to check, " variable analysis module " according to algorithm and algorithm to data form
Requirement, the data submitting user to are analyzed, determine variable parameter, " algorithm general program module "
Calculate according to the automatic founding mathematical models of variable parameter that " variable analysis module " determines.
Described " algorithm general program module " includes " classified counting ", " cluster calculation ", " PCA
Calculate ", " association analysis calculating ", " sequence analysis calculating " and " text mining calculating " program.
For " classified counting " program: if user submits TXT data, system regulation data form to
By: the 1st behavioral data descriptive item is expert at；1st is classified as " identification id ", and last is classified as " decision-making
Variable " D, remaining be classified as m " conditional attribute variable " C1, C2 ..., Ci ..., Cm},
Between character string with separators such as space, comma, Tab separately；" variable analysis module " is come really with this
Determining the variable parameter such as variable name, span, " algorithm general program module " builds mathematical modulo with this
Type calculates.
For " classified counting " program: if data leave in data base, system regulation submits number to
According to form it is: include 1 " identification id ", 1 " decision variable " and m " conditional attribute variable "
C1, C2 ..., Ci ..., Cm} variable；Every one variable declaration of behavior, illustrates " variable in row
Attribute ", " variable name ", " data base's table name " and " field name "；" variable analysis module " is the most true
Determine variable name, composition SQL string, from data base, inquire about data, determine the variable parameters such as span；
" algorithm general program module " builds mathematical model with this and calculates.
For " cluster calculation " or " PCA calculating " program: if user submits TXT data to, be
System regulation data form by: the 1st behavioral data descriptive item is expert at；1st is classified as " identification id ", its
More than be classified as m " property variable " A1, A2 ..., Ai ..., Am}, with sky between character string
The separators such as lattice, comma, Tab are separately；" variable analysis module " determines the variablees such as variable name with this
Parameter, " algorithm general program module " builds mathematical model with this and calculates.
For " cluster calculation " or " PCA calculating " program: if data leave in data base,
System regulation submits to the data form to be: include 1 " identification id " and m individual " property variable " A1,
A2 ..., Ai ..., Am}；Every one variable declaration of behavior, illustrates " variable's attribute ", " becomes in row
Amount name ", " data base's table name " and " field name "；" variable analysis module " determine therefrom that variable name,
Composition SQL string, inquires about data from data base, determines the variable parameters such as variable name；" the general journey of algorithm
Sequence module " build mathematical model with this and calculate.
For " association analysis calculating " or " sequence analysis calculating " program: if user submits TXT to
Data, system regulation data form is: all data from the 1st row；1st is classified as " identification id ",
Remaining is classified as " things or commodity ", between character string with separators such as space, comma, Tab separately；
The columns of every record can differ；" variable analysis module " determines the variable ginsengs such as variable name with this
Number, " algorithm general program module " builds mathematical model with this and calculates.
For " association analysis calculating " or " sequence analysis calculating " program: if data leave number in
According in storehouse, system regulation submission data form is: include " identification id " and " things or commodity " two
Type variable；Every one variable declaration of behavior, including " variable's attribute ", " variable name ", " data
Storehouse table name " and " field name "；" variable analysis module " determines therefrom that variable name, composition SQL string,
From data base, inquire about data, determine the variable parameters such as variable name；" algorithm general program module " is with this
Build mathematical model to calculate.
For " text mining calculating " program: user selects a certain " text mining " algorithm, submit to
One group of text, selection text represent word quantity；" variable analysis module " is come really according to data form regulation
Determine amount of text and the variable parameter of algorithm needs；" algorithm general program module " builds mathematics with this
Model calculates.
The present invention compared with prior art, has the advantage that
1, grasp, without user, mathematical theory and the algorithm knowledge that classification, cluster, text mining etc. relate to,
Only it is to be understood that what classification, cluster, text mining can do, selection algorithm by specifying submission data,
Algorithm and the data of submission that system just can select according to user calculate, it is simple to non-data is excavated
Professional uses.
2, need not EXEC user defined variableEXEC, the quantity of explanatory variable and span, only need to carry by regulation
For data, system just can automatically determine variable quantity, title and span, selected by user
The algorithm selected, automatic founding mathematical models carries out excavating calculating.
As long as 3, anyone logs in cloud computing system web constructed by the present invention by Internet
Server, or it is connected to APP Server by application program, it is possible to carry out business intelligence cloud computing.
4, being easy to be embedded in application program, application program submits data to APP Server, from greatly
Data find knowledge, it is achieved data mining and the seamless connection of application program.
Detailed description of the invention
Internet or LAN sets up a Web Server or APP Server.
System is constituted and effect:
System is mainly made up of 3 program modules:
Module 1: data review module
Module is data check program, is used for checking data, the algorithm selected according to user and algorithm
Requirement to data form, the data that user is submitted to, if meet the data form that algorithm specifies and enter
Row checks.
Module 2: data analysis module
Module is DAP, determines variable for analytical data, the calculation selected according to user
The data that user is submitted to by method and algorithm by the requirement of data form are analyzed, and determine and have how many to become
The variable parameters such as the span of amount, the character of variable and variable, " module 3 " sets up mathematics accordingly
Model calculates.
Module 3: algorithm general program module
Module is algorithm general program, for automatic founding mathematical models and calculating, if module is
The dry algorithm general program write, but uncertain have how many variablees, the character of variable and variable
The variable parameters such as span, the most uncertain concrete mathematical model, only algorithm flow, " module 3 "
According to variable parameter determined by " module 2 ", automatic founding mathematical models calculates.
System also has other auxiliary programs, such as: result of calculation display module, application programming interfaces etc..
Working-flow:
Step one, user's selection algorithm, algorithmically submit data to the prescribed form of data；
The data of step 2, the algorithm that user is selected by " module 1 " and submission check, if not
Meet the requirements, return error message, otherwise, call " module 2 "；
Algorithm that step 3, " module 2 " select according to user and data format requirement, submit to user
Data be analyzed, determine the quantity of variable and the span of variable, call " module 3 "；
The variable parameter that step 4, " module 3 " basis " module 2 " determine sets up concrete mathematical modulo
Type, distribution memory element, corresponding algorithm general program calculates.
The specific implementation method of various algorithms:
One, sorting algorithm
Classify and belong to computer learning category, existing a lot of sorting algorithms, such as: Bayes's classification, ID3
Classification, Rough Sets Classification etc..The problem that classification is to be solved is: be provided with a sample set, including n bar
Know that the record of tag along sort, every record comprise 1 " identification id ", m " conditional attribute variable "
(C1, C2 ..., Ci ..., Cm) and 1 " decision variable " D, it is each that " conditional attribute becomes
Amount " Ci and " decision variable " D have several values.Every is recorded as an example, and m worked as in record
The timing of individual " conditional attribute variable " Ci value one, the value of " decision variable " D.
The purpose of classification is to excavate classifying rules from sample set: i.e., " conditional attribute variable " Ci
With functional relationship f (the C)=D of " decision variable " D, functional relationship is utilized to determine: as given m
During the value of individual Ci, the value of D or probability.
No matter use which kind of sorting algorithm, be required for being determined in advance the quantity of " conditional attribute variable " Ci
M, variable name and span, the variable name of " decision variable " D and span are the most permissible
Set up concrete mathematical model to calculate.
The present invention comes automatic situational variables quantity and span by the following method, automatically builds classification
Mathematical model calculates.
(1), data form regulation
User can submit TXT or two kinds of data of data base to:
1, TXT data form regulation
(1) require as TXT data；
(2) n bar record, one record of every behavior are included；
(3) every record by 1 " identification id ", m " conditional attribute variable Ci " C1, C2 ...,
Ci ..., Cm} and 1 " decision variable " D composition；
(4) the 1st behavioral data descriptive item of text is expert at；
(5) the 1st are classified as " identification id ", and last 1 is classified as " decision variable D " column, remaining
It is classified as " conditional attribute variable " Ci；
(6) separator such as character string space, comma, Tab is separately.
As user submits data to it is:
Limiting according to claim 3, data are resolved to by system: the 1st behavioral data descriptive item, the 2nd,
3 behavioral datas, the 1st is classified as " identification id " (" recording mechanism "), and last is classified as " decision variable " and (" purchases
Buy "), remaining 2,3,4,5 be classified as " conditional attribute variable " (" age ", " income ", " student is no ",
" prestige ").
The invention is not restricted to said method, it is also possible to other forms regulation TXT data form, system root
According to being embodied as data regulation, data are resolved.
2, the data form regulation in data base is left in
(1) include 1 " identification id ", m " conditional attribute variable Ci " C1, C2 ...,
Ci ..., Cm} and 1 " decision variable " D, three types variable；
(2) every one variable declaration of behavior, has 4 data description entries:
" variable's attribute ": " identification id ", " conditional attribute variable ", " decision variable "；
" variable name ": variable name during display；
" data base's table name ": data leave in which table of data base；
" field name ": the field name in database table.
(3) each data descriptive item angle brackets "<>" expand, and form is as follows:
<variable's attribute>,<variable name>,<data base's table name>,<field name>
As: "<identification id>,<recording mechanism>,<table 1>,<RecID>", limit according to claim 4
Fixed, data are resolved to by system:
" variable's attribute " is: " identification id ",
" variable name " is: " recording mechanism ",
" data base's table name " is: " table 1 ",
" field name " is: " RecID ".
As: "<conditional attribute variable>,<age>,<table 1>,<Age>", according to claim 4
Limiting, data are resolved to by system:
Variable's attribute is: " conditional attribute variable ",
Variable is entitled: " age ",
Data leave in " table 1 " of data base,
Field entitled " Age ".
As: "<decision variable>,<buying no>,<table 1>,<Buy>", limit according to claim 4
Fixed, data are resolved to by system:
Variable's attribute is: " decision variable ",
Variable is entitled: " buying no ",
Data leave in " table 1 " of data base,
Field entitled " Buy ".
The invention is not restricted to said method, it is also possible to other forms specify the data leaving in data base
Form, data are resolved by system according to being embodied as data regulation.
(2), system runs specific implementation method
User can log in Web Server or connect APP Server by application program.
1, log in Web Server to use
(1) user submits data to
User logs in Web Server by browser, and selection sort algorithm, according to algorithm to data form
Data are submitted in the requirement of regulation to, illustrate that the data submitted to are TXT or leave in data base, system
Call " module 1 ".
(2) data are checked
Algorithm that " module 1 " selects according to user and data format requirement, check whether data meet rule
Fixed, if against regulation, show error message, otherwise, call " module 2 ".
(3) analytical data
If A user submits TXT data to, " module 2 " is according to the algorithm selected by user and data
Specify to user submit to data be analyzed, with the 1st row of data determine Ci and D column,
Quantity m of " conditional attribute variable " Ci, the variable name of each Ci, the variable name of " decision variable " D,
Data are added up by the column according to variable, obtain the union of each variable-value, determine with this
The span of variable, calls " module 3 ".
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other TXT data forms,
It should be understood that can be by real data form provision discussion data.
If B data leave in data base, " module 3 " is according to the algorithm selected by user and data
Specifying that the data submitting user to are analyzed, set up data base and connect, composition SQL string, from data
Storehouse table inquires " identification id ", m " conditional attribute variable " Ci{C1, C2 ..., Ci ...,
Cm} and 1 " decision variable " D forms record set Set, the number of statistics " conditional attribute variable Ci "
Amount m, respectively " conditional attribute variable " Ci and " decision variable " D value in statistic record collection Set
Union, the span determining variable with this, call " module 3 ".
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other data bases deposits number
According to form, it should be understood that can be by actual prescribed form analytical data.
(4) founding mathematical models calculates
The variable that " module 3 " basis " module 2 " determines distributes memory element, founding mathematical models,
Corresponding general-purpose algorithm program calculates.
2, application program connects APP Server
(1) user submits data to
User connects APP Server by application program, submits algorithm mark to and meets the number that algorithm specifies
According to, illustrate that the data submitted to are TXT or leave in data base, call " module 1 "；
(2) data are checked
Algorithm that " module 1 " selects according to user and data format requirement, check the data that user submits to
Whether meet regulation, if against regulation, return error message, otherwise, call " module 2 "；
(3) analytical data
If A user submits TXT data to, " module 2 " is according to the algorithm selected by user and data
Specify to user submit to data be analyzed, with the 1st row of data determine Ci and D column,
Quantity m of " conditional attribute variable " Ci, the variable name of each Ci, the variable name of " decision variable " D,
Data are added up by the column according to variable, obtain the union of each variable-value, determine with this
The span of variable, calls " module 3 " and is modeled, calculates.
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other TXT data forms,
It should be understood that can be by real data prescribed form analytical data.
If B data leave in data base, " module 2 " is according to the algorithm selected by user and data
Specifying that the data submitting user to are analyzed, set up data base and connect, composition SQL string, from data
Storehouse table inquires " identification id ", m " conditional attribute variable " Ci{C1, C2 ..., Ci ...,
Cm} and 1 " decision variable " D forms record set Set, the number of statistics " conditional attribute variable Ci "
Amount m, respectively " conditional attribute variable " Ci and " decision variable " D value in statistic record collection Set
Union, the span determining variable with this, call " module 3 " and be modeled, calculate.
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other data bases deposits number
According to form, it should be understood that can be by actual prescribed form analytical data.
(4) founding mathematical models calculates
The variable that " module 3 " basis " module 2 " determines distributes memory element, founding mathematical models,
Corresponding general-purpose algorithm program calculates, and result of calculation is placed on APP Server, and user applies
Program can be used directly result of calculation.
Two, " clustering algorithm " and " PCA algorithm "
Cluster and belong to computer learning category, existing a lot of clustering algorithms, such as: k-means algorithm, mould
Stick with paste cluster, SOM neural network clustering etc..Cluster belongs to " unsupervised segmentation ", compares with classification,
N bar record in cluster sample set does not has " decision variable " D (tag along sort), and only m " poly-
Generic attribute variable " (A1, A2 ..., Ai ..., Am), owing to there is no " decision variable " D, because of
This, it is not known which classification every record belongs to, so claiming " unsupervised segmentation ".
The purpose of cluster is, according to the value of n bar record " cluster property variable " Ai in sample set,
Similar record is divided into identical classification, and the record similarity belonging to same category is maximum, belongs to
The record difference of different classification is maximum.
" PCA " algorithm is: assumes that a things is made up of many factors, is provided with n sample, often
Individual sample has m attribute, constitutes the compositional data matrix on n × m rank,
The purpose of PCA algorithm is:
(1) dimension is reduced
When the dimension m of matrix X is bigger, m-dimensional space is investigated problem cumbersome, need fall
Low dimensional, on the basis of not affecting things evaluation, selects less several leading indicator P (P < m)
Replace the most more variable index m.
(2) dependency between variable is eliminated
When describing things by multiple conditional-variables, dependency between variable, will be likely to be of, both, some
Can influence each other between variable, independent reflection features can not be waited.Owing to dimensionality reduction is just to use one
Handing over P the aggregative indicator that matrixing obtains, therefore, orthogonal matrix ensure that P aggregative indicator
Irrelevance, is independent of each other between variable, eliminates influencing each other of former m index；
(3) distinction to things of each index in analysis indexes system.Weigh a things quality
Determined by multiple indexs, but index had dividing of power to the distinction of things, is calculated by PCA,
Can analyze which index and have more preferable distinction, the distinction of which index is more weak.
Owing to " cluster " calculates identical to data demand with " PCA ", therefore, the present invention is returned
It it is a class.
The present invention carrys out automatic situational variables by the following method, structure " clusters " or " PCA " mathematics
Model calculates.
(1), data form regulation
User can submit TXT or two kinds of data of data base to:
1, TXT data form regulation
(1) require as TXT data；
(2) n bar record, one record of every behavior are included；
(3) every record by 1 " identification id ", m " property variable " A1, A2 ...,
Ai ..., the two kinds of variable of Am} forms；
(4) the 1st behavioral data descriptive item of text is expert at；
(5) the 1st are classified as " identification id ", and remaining is classified as " property variable " Ai；
(6) separator such as character string space, comma or Tab is separately.
As user submits data to it is:
Limiting according to claim 5, system is to data parsing: the 1st behavioral data descriptive item, the 2nd,
3 behavioral datas, the 1st is classified as " identification id " (" regional "), and remaining the 2nd, 3,4 is classified as " attribute
Variable " (" GDP ", " fixed assets ", " human capital ").
The invention is not restricted to said method, it is also possible to other forms regulation TXT data form, system root
According to being embodied as data regulation, data are resolved.
2, the data form regulation in data base is left in
(1) include 1 " identification id " and m " property variable " A1, A2 ..., Ai ...,
Am} two types variable；
(2) every one variable declaration of behavior, has 4 data description entries:
" variable's attribute ": " identification id ", " property variable "；
" variable name ": variable name during display；
" data base's table name ": data leave in which table of data base；
" field name ": the field name in database table.
(3) each data descriptive item angle brackets "<>" expand, and form is as follows:
<variable's attribute>,<variable name>,<data base's table name>,<field name>
As: "<identification id>,<area>,<table 1>,<Area>", limit according to claim 6,
Data are resolved to by system:
Variable's attribute is: " identification id ",
Variable is entitled: " regional ",
Data leave in " table 1 " of data base,
Field entitled " Area ".
As: "<property variable>,<output value>,<table 1>,<GDP>", limit according to claim 6,
Data are resolved to by system:
Variable's attribute is: " property variable ",
Variable is entitled: " output value ",
Data leave in " table 1 " of data base,
Field entitled " GDP ".
The invention is not restricted to said method, it is also possible to other forms specify the data leaving in data base
Form, data are resolved by system according to being embodied as data regulation.
(2), system runs specific implementation method
User can log in Web Server or connect APP Server by application program.
1, Web Server is logged in
(1) user submits data to
User logs in Web Server by browser, and selection sort algorithm, according to algorithm to data form
Requirement submit to data, illustrate submission data be TXT or leave in data base, system is called
" module 1 "；
(2) data are checked
Algorithm that " module 1 " selects according to user and data format requirement, check whether data meet rule
Fixed, if against regulation, show error message, otherwise, call " module 2 "；
(3) analytical data
If A user submits TXT data to, " module 2 " is according to the algorithm selected by user and data
Data are analyzed by regulation, determine " identification id ", " property variable " with the 1st row of TXT
The column of Ai, quantity m of " property variable " Ai, the variable name of Ai, call " module 3 ", enter
Row modeling, calculating；
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other TXT data forms,
It should be understood that can be by real data prescribed form analytical data.
If B data leave in data base, " module 2 " is according to the algorithm selected by user and data
Data are analyzed by regulation, set up data base and connect, composition SQL string, inquire about from database table
Go out to inquire " identification id " and " property variable " Ai from database table and form record set Set, system
Quantity m of meter " property variable " Ai, calls " module 3 " and is modeled, calculates.
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other data bases deposits number
According to form, it should be understood that can be by actual prescribed form analytical data.
(4) founding mathematical models calculates
The variable that " module 3 " basis " module 2 " determines distributes memory element, founding mathematical models,
Corresponding general-purpose algorithm program calculates, and shows result of calculation on a web browser.
2, application program connects APP Server
(1) user submits data to
User connects APP Server by application program, submits algorithm mark to and meets the number that algorithm specifies
According to, illustrate that the data submitted to are TXT or leave in data base, call " module 1 "；
(2) data are checked
Algorithm that " module 1 " selects according to user and data format requirement, check whether data meet rule
Fixed, if against regulation, return error message, otherwise, call " module 2 "；
(3) analytical data
If A user submits TXT data to, " module 2 " is according to the algorithm selected by user and data
Data are analyzed by regulation, determine " identification id ", " property variable " with the 1st row of TXT
The column of Ai, quantity m of " property variable " Ai, the variable name of Ai, call " module 3 ", enter
Row modeling, calculating；
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other TXT data forms,
It should be understood that can be by actual provision discussion data.
If B data leave in data base, " module 2 " is according to the algorithm selected by user and data
Data are analyzed by regulation, set up data base and connect, composition SQL string, inquire about from database table
Go out " identification id " and " property variable " Ai and form record set Set, statistics " property variable " Ai's
Quantity m, calls " module 3 ", is modeled, calculates.
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other data bases deposits number
According to form, it should be understood that can be by actual provision discussion data.
(4) founding mathematical models calculates
The variable that " module 3 " basis " module 2 " determines distributes memory element, founding mathematical models,
Corresponding general-purpose algorithm program calculates, and result of calculation is placed on APP Server, and user applies
Program can be used directly result of calculation.
Three, " association analysis calculating " and " sequence analysis calculating "
The purpose " associating " analytical calculation is, " things (or transaction) " record is by multiple " things
Part (or transaction) " constitute, by the statistical analysis to record, find " event (or transaction) "
" associate " rule.
The purpose of " sequence " analytical calculation is, by the statistical analysis to record, find " event (or
Transaction) " " sequence " rule of sequencing.
The present invention carrys out automatic situational variables by the following method, structure mathematical model calculates.
(1), data form regulation
User can submit TXT or two kinds of data of data base to:
1, TXT data form regulation
(1) require as TXT data；
(2) n bar record, one record of every behavior are included；
(3) all data from the 1st row；
(4) between the column and the column with separators such as space, comma or Tab separately；
(5) often row includes " identification id " and " event (or commodity) " two types variable, the 1st
Being classified as " identification id ", remaining is classified as " event (or commodity) "；
(6) columns of every record can differ.
As user submits data to it is:
T1 milk, bread
T2 milk, bread, pure water
T3 milk, pure water
Limit according to claim 7 and resolve: the 1st is classified as " identification id " (T1, T2, T3),
2nd row and later be classified as " event (or commodity) ".
The invention is not restricted to said method, it is also possible to other forms regulation TXT data form, system root
According to being embodied as data regulation, data are resolved.
2, the data form regulation in data base is left in
(1) " identification id " and " event (or commodity) " two types variable is included；
(2) every one variable declaration of behavior, has 4 data description entries:
" variable's attribute ": " identification id ", " event (or commodity) "；
" variable name ": variable name during display；
" data base's table name ": data leave in which table of data base；
" field name ": the field name in database table.
(3) each data descriptive item angle brackets "<>" expand, and form is as follows:
<variable's attribute>,<variable name>,<data base's table name>,<field name>
As: "<identification id>,<transaction record>,<table 1>,<T>", limit according to Claim 8
Data are resolved to:
Variable's attribute is: " identification id ",
Variable is entitled: " transaction record ",
Data leave in " table 1 " of data base,
Field entitled " T ".
As: "<event (or commodity)>,<purchase commodity>,<table 1>,<Goods>", according to right
Require that data are resolved to by 8 restrictions:
Variable's attribute is: " event (or commodity) ",
Variable is entitled: " purchase commodity ",
Data leave in " table 1 " of data base,
Field entitled " Goods ".
The invention is not restricted to said method, it is also possible to other forms specify the data leaving in data base
Form, data are resolved by system according to being embodied as data regulation.
(2), system runs specific implementation method
User can log in Web Server or connect APP Server by application program.
1, Web Server is logged in
(1) user submits data to
User logs in Web Server by browser, selects " association " or " sequence " algorithm, according to
The requirement of data form is submitted to data by algorithm, illustrates that the data submitted to are TXT or leave data in
In storehouse, call " module 2 "；
(2) data are checked
" module 1 ", according to the algorithm requirement to data, checks whether the data that user submits to meet regulation,
If against regulation, show error message, otherwise, call " module 2 "；
(3) analytical data
If A user submits TXT data to, the regulation of data form is divided by " module 2 " according to algorithm
Analysis data, determine " identification id ", the column of " event (or commodity) ", " event (or commodity) "
Quantity, call " module 3 ", be modeled, calculate；
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other TXT data forms,
It should be understood that can be by real data prescribed form analytical data.
If B data leave in data base, " module 2 " is according to the algorithm selected by user and data
Data are analyzed by regulation, set up data base and connect, composition SQL string, inquire about from database table
Go out " identification id " and " event (or commodity) " composition record set Set, only two of which field,
1st field is " identification id ", and the 2nd field is " event (or commodity) ", calls " module 3 ",
It is modeled, calculates.
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other data bases deposits number
According to form, it should be understood that can be by actual prescribed form analytical data.
(4) founding mathematical models calculates
The variable that " module 3 " basis " module 2 " determines distributes memory element, founding mathematical models,
General algorithm routine calculates.
2, application program connects APP Server
(1) user submits data to
User connects APP Server by application program, submits algorithm mark to and meets the number that algorithm specifies
According to, illustrate that the data submitted to are TXT or leave in data base, call " module 1 "；
(2) data are checked
" module 1 ", according to the algorithm regulation to data, checks whether data meet regulation, if be not inconsistent
The hop algorithm regulation to data, returns error message, otherwise " module 2 "；
(3) analytical data
If A user submits TXT data to, the regulation of data form is divided by " module 2 " according to algorithm
Analysis data, determine " identification id ", the column of " event (or commodity) ", " event (or commodity) "
Quantity, call " module 3 ", be modeled, calculate；
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other TXT data forms,
It should be understood that can be by real data prescribed form analytical data.
If B data leave in data base, " module 2 " is according to the algorithm selected by user and data
Data are analyzed by regulation, set up data base and connect, composition SQL string, inquire about from database table
Go out " identification id " and " event (or commodity) " composition record set Set, only two of which field,
1st field is " identification id ", and the 2nd field is " event (or commodity) ", calls " module 3 ",
It is modeled, calculates.
The invention is not restricted to above-mentioned data analysing method, if prescribed form is other data bases deposits number
According to form, it should be understood that can be by actual prescribed form analytical data.
(4) founding mathematical models calculates
The variable that " module 3 " basis " module 2 " determines distributes memory element, founding mathematical models,
General-purpose algorithm program calculates, and result of calculation is placed on APP Server, and user application can
Directly use result of calculation.
Four, text mining
So-called " text " is a substantial text strings sequence data, including the electricity such as webpage, Word
Subdocument." text mining " belongs to computer learning category, currently mainly has text mining to have: text
Classification, text cluster and content of text Similarity Measure etc..
The present invention carrys out automatic situational variables by the following method, structure " text mining " mathematical model enters
Row calculates.
(1), data form regulation
(1) one group of text data file；
(2) quantity of word in " representing phrase ", so-called " representing phrase " is that one group of weight is maximum
Word, represents, with this phrase, the content that document is stated.
(2), system runs specific implementation method
User can log in Web Server or connect APP Server by application program.
1, Web Server is logged in
(1) user logs in Web Server by browser, selects a certain " text mining " algorithm,
Submit one group of text to, select to determine the quantity of word in " representing phrase ", call " module 1 "；
(2) " module 1 " checks that user submits whether the data of data meet regulation to, if not meeting rule
Fixed, show error message, otherwise call " module 2 "；
(3) the data form provision discussion data that " module 2 " basis " text mining " calculates, determine
Amount of text and " representing phrase " in the quantity of word, call " module 3 ", be modeled, calculate；
(4) variable that " module 3 " basis " module 2 " determines distributes memory element, sets up mathematics
Model, algorithm general program calculates.
2, application program connects APP Server
(1) user connects APP Server by application program, submits algorithm mark to and meets algorithm rule
Fixed data, call " module 1 "；
(2) " module 1 " checks that user submits whether the data of data meet regulation to, if not meeting rule
Fixed, show error message, otherwise call " module 2 "；
(3) the data form provision discussion data that " module 2 " basis " text mining " calculates, determine
In the quantity of text and " representing phrase ", the quantity of word, calls " module 3 "；
(4) variable that " module 3 " basis " module 2 " determines distributes memory element, sets up mathematics
Model, algorithm general program calculates, and result of calculation is placed on APP Server, and user applies
Program can be used directly result of calculation.
Embodiment 1
If any classification problem: consumer purchases goods behavior to be carried out Bayes's classification calculating, excavate difference
The rule of probability of type customer purchasing behavior.Specific implementation method is:
1, selection algorithm
User's Website login, selects Bayesian Classification Arithmetic.
2, data are submitted to
As follows shown in " table 1 ", user submits data to by regulation, and system is called " module 1 " and checked number
According to whether meet regulation, if against regulation, point out error message, otherwise call " module 2 ".
Table 1: user submits data to by regulation
3, " module 2 " analytical data, determine variable parameter
Claim 3 limits: " the 1st behavioral data descriptive item is expert at, and the 1st is classified as " identification id ",
Last 1 is classified as " decision variable " D column, remaining be classified as " conditional attribute variable " C1, C2 ...,
Ci ..., Cm} ", accordingly, " module 2 " is analyzed as follows:
(1) decision variable analysis
By claim 3 limit understand, the first row last be classified as " decision variable " D, divide from data
Analyse to obtain variable entitled " purchase ", last string is added up D has two values { Y, N}；
(2) Conditional Variable Analysis
Shown in variable analysis result following " table 2 ", module 2 data analysis is as follows:
Being limited by claim 3 and understand, the 1st row is in addition to the 1st row and last string, and remaining is classified as " bar
Part property variable " Ci, the 1st row is analyzed variable name is respectively as follows:
" age " of the 2nd row；
" income " of 3rd row；
" student is no " of the 4th row；
" prestige " of the 5th row.
Arrange from the 2 of data, 3,4,5 and add up the span drawing Ci:
C1 (age)={ > 40 ,≤30,31～40}；
C2 (income)={ basic, normal, high }；
C3 (student is no)={ N, Y}；
C4 (prestige)={ good, poor }.
(3) sample set record analysis
Have 14 records, n=14；
Shown in analysis result following " table 2 ", then system calls " module 3 ".
Table 2: automatically identify variable
4, " module 3 " founding mathematical models automatically calculates
" module 2 " analytical data, determine variable parameter after, " module 3 " is set up according to above-mentioned analysis
Mathematical model, storage allocation, calls the general Bayesian probabilistic classifier write and calculates.
Embodiment 2
If any clustering problem: to 32 the provinces, cities and autonomous regions' levels of economic development in the whole nation, by GDP, fix
Assets and human capital 3 cluster.Specific implementation method is:
1, user's selection algorithm
User's Website login, selects fuzzy clustering algorithm；
2, TXT data are submitted to
As follows shown in " table 3 ", user submits data to by regulation.System is called " module 1 " and is checked number
According to whether meet regulation, if against regulation, point out error message, otherwise call " module 2 ".
Table 3: the user of example 2 submits data to by regulation
3, " module 2 " analytical data determines variable parameter
Claim 5: " the 1st behavioral data descriptive item is expert at, and the 1st is classified as " identification id " place
Row, remaining be classified as " property variable " A1, A2 ..., Ai ..., Am} column ", " module 2 "
It is analyzed as follows:
(1) variable analysis
Shown in variable analysis result following " table 4 ".Being limited by claim 5, " module 2 " is carried out
Analyzing, draw from the 1st row analysis, have 3 property variable Ai in addition to the 1st row, variable is entitled is not:
" GDP " of the 2nd row；
" fixed assets " of the 3rd row；
" human capital " of the 4th row.
(2) sample set record analysis
Have 32 records, n=32,3 ATTRIBUTE INDEX: GDP, fixed assets and human capital.
Shown in analysis result following " table 4 " (owing to screen limits, " table 4 " only shows 16 records,
32 records of display the most completely), then system calls " module 3 ".
Checking text
Table 4: example 2 automatically identify variable
4, " module 3 " founding mathematical models automatically calculates
" module 2 " analytical data, determine variable parameter after, " module 3 " is set up according to above-mentioned analysis
Mathematical model, storage allocation, calls the fuzzy clustering program write and calculates.
Examples of implementation 3
With " examples of implementation 2 " above, if any clustering problem: to 32 the provinces, cities and autonomous regions' economy in the whole nation
Level of development, is clustered by GDP, fixed assets and human capital 3.Method particularly includes:
1, user's selection algorithm
User application is connected to APP Sever, selects fuzzy clustering algorithm；
2, data are submitted to
Data leave in data base, and the data content that user submits to is:
<identification id>,<area>,<table 1>,<RecID>
<property variable>,<output value>,<table 2>,<GDP>
<property variable>,<fixed assets>,<table 3>,<GDZC>
<property variable>,<human resources>,<table 4>,<RLZY>
System call " module 1 " check data whether meet regulation, if against regulation, prompting
Error message, otherwise calls " module 2 ".
3, " module 2 " analytical data, determine variable parameter
" calculating for " cluster " or " PCA ", claim 6 limits: if sample data is deposited
It is placed in database table, one variable declaration of every behavior, including " variable's attribute ", " variable name ", " number
According to storehouse table name " and " field name ".
(1) variable analysis
According to claim 6, data are analyzed as follows by " module 2 ":
Data the 1st row is analyzed as follows by system:
<identification id>,<area>,<table 1>,<RecID>
Variable's attribute is: " identification id ",
Variable is entitled: " regional ",
Data leave in " table 1 " of data base,
Field entitled " RecID ".
Data the 2nd row is analyzed as follows:
"<property variable>,<output value>,<table 2>,<GDP>"
Variable's attribute is: " property variable ",
Variable is entitled: " output value ",
Data leave in " table 2 " of data base,
Field entitled " GDP ".
Data the 3rd row is analyzed as follows:
"<property variable>,<fixed assets>,<table 3>,<GDZC>"
Variable's attribute is: " property variable ",
Variable is entitled: " fixed assets ",
Data leave in " table 3 " of data base,
Field entitled " GDZC ".
Data the 4th row is analyzed as follows:
"<property variable>,<human resources>,<table 4>,<RLZY>"
Variable's attribute is: " property variable ",
Variable is entitled: " human resources ",
Data leave in " table 4 " of data base,
Field entitled " RLZY ".
The m=3 of statistics " property variable "；
Set up data base to connect, system composition SQL string, inquire about from the table 1,2,3,4 of data base
Go out related data, form record set Set, the record number n in statistics Set.
SQL is prior art, and available any form SQL, as long as identical result can be inquired.
4, " module 3 " founding mathematical models automatically calculates
" module 2 " analytical data, determine variable parameter after, " module 3 " is set up according to above-mentioned analysis
Mathematical model, storage allocation, calls the fuzzy clustering program write and calculates.
Claims (9)
1. a construction method for business intelligence cloud computing system, sets up a Web Server or APP Server on Internet or LAN, it is characterised in that:
System mainly includes " data review module ", " variable analysis module " and " algorithm general program module "；
" data review module " is used for checking data, the algorithm selected according to user and the algorithm requirement to data form, the data submitting user to, if meets the data form that algorithm specifies and checks；
" variable analysis module " determines variable for analytical data, the algorithm selected according to user and the algorithm requirement to data form, and the data submitting user to are analyzed, and determines the variable parameters such as the span that has how many variablees, the character of variable and variable；
" algorithm general program module " is used for automatic founding mathematical models and calculating, for some algorithm general programs write in module, but the variable parameters such as the uncertain span having how many variablees, the character of variable and variable, the most uncertain concrete mathematical model, only algorithm flow, according to variable parameter determined by " variable analysis module ", the automatic founding mathematical models of system calculates；
System flow is: user enters Web Server by browser or application program connects APP Server, selection algorithm, the data form algorithmically specified submit data to, the data that user is submitted to by " data review module " check, " variable analysis module " requirement to data form according to algorithm and algorithm, the data submitting user to are analyzed, determine variable parameter, and the automatic founding mathematical models of variable parameter that " algorithm general program module " basis " variable analysis module " determines calculates.
2. the construction method of a kind of business intelligence cloud computing system described in claim 1, it is characterised in that: described " algorithm general program module " includes " classified counting ", " cluster calculation ", " PCA calculating ", " association analysis calculating ", " sequence analysis calculating " and " text mining calculating " program.
3. the construction method of a kind of business intelligence cloud computing system described in claim 2, it is characterised in that: for " classified counting " program, if user submits TXT data to, system regulation data form by: the 1st behavioral data descriptive item is expert at；1st is classified as " identification id ", and last is classified as " decision variable " D, remaining be classified as m " conditional attribute variable " C1, C2 ..., Ci ..., between Cm} character string with separators such as space, comma, Tab separately；" variable analysis module " determines the variable parameter such as variable name, span with this, and " algorithm general program module " builds mathematical model with this and calculate.
4. the construction method of a kind of business intelligence cloud computing system described in claim 2, it is characterized in that: for " classified counting " program, if data leave in data base, system regulation submission data form is: include 1 " identification id ", 1 " decision variable " and m " conditional attribute variable " { C1, C2 ..., Ci, ..., Cm} variable；Every one variable declaration of behavior, illustrates " variable's attribute ", " variable name ", " data base's table name " and " field name " in row；" variable analysis module " determines therefrom that variable name, composition SQL string, inquires about data, determines the variable parameters such as span from data base；" algorithm general program module " builds mathematical model with this and calculates.
The construction method of a kind of business intelligence cloud computing system the most according to claim 2, it is characterized in that: for " cluster calculation " or " PCA calculating " program, if user submits TXT data to, system regulation data form by: the 1st behavioral data descriptive item is expert at；1st is classified as " identification id ", remaining be classified as m " property variable " A1, A2 ..., Ai ..., Am}, between character string with separators such as space, comma, Tab separately；" variable analysis module " determines the variable parameters such as variable name with this, and " algorithm general program module " builds mathematical model with this and calculate.
The construction method of a kind of business intelligence cloud computing system the most according to claim 2, it is characterized in that: for " cluster calculation " or " PCA calculating " program, if data leave in data base, system regulation submission data form is: include 1 " identification id " and m " property variable " { A1, A2 ..., Ai, ..., Am}；Every one variable declaration of behavior, illustrates " variable's attribute ", " variable name ", " data base's table name " and " field name " in row；" variable analysis module " determines therefrom that variable name, composition SQL string, inquires about data, determines the variable parameters such as variable name from data base；" algorithm general program module " builds mathematical model with this and calculates.
The construction method of a kind of business intelligence cloud computing system the most according to claim 2, it is characterized in that: for " association analysis calculating " or " sequence analysis calculating " program, if user submits to the TXT data, system regulation data form to be: all data from the 1st row；1st is classified as " identification id ", and remaining is classified as " things or commodity ", between character string with separators such as space, comma, Tab separately；The columns of every record can differ；" variable analysis module " determines the variable parameters such as variable name with this, and " algorithm general program module " builds mathematical model with this and calculate.
The construction method of a kind of business intelligence cloud computing system the most according to claim 2, it is characterized in that: for " association analysis calculating " or " sequence analysis calculating " program, if data leave in data base, system regulation submission data form is: include " identification id " and " things or commodity " two types variable；Every one variable declaration of behavior, including " variable's attribute ", " variable name ", " data base's table name " and " field name "；" variable analysis module " determines therefrom that variable name, composition SQL string, inquires about data, determines the variable parameters such as variable name from data base；" algorithm general program module " builds mathematical model with this and calculates.
The construction method of a kind of business intelligence cloud computing system the most according to claim 2: it is characterized in that: for " text mining calculating " program, user selects a certain " text mining " algorithm, submits one group of text to, selects text to represent word quantity；" variable analysis module " specifies to determine the variable parameter that amount of text and algorithm need according to data form；" algorithm general program module " builds mathematical model with this and calculates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310530032.8A CN103544299B (en) | 2013-10-30 | 2013-10-30 | A kind of construction method of business intelligence cloud computing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310530032.8A CN103544299B (en) | 2013-10-30 | 2013-10-30 | A kind of construction method of business intelligence cloud computing system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103544299A CN103544299A (en) | 2014-01-29 |
| CN103544299B true CN103544299B (en) | 2017-01-04 |
Family
ID=49967751
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310530032.8A Active CN103544299B (en) | 2013-10-30 | 2013-10-30 | A kind of construction method of business intelligence cloud computing system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103544299B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105989138B (en) * | 2015-02-27 | 2019-12-31 | 北大方正集团有限公司 | Data processing method, data processing system and server |
| CN105894019A (en) * | 2016-03-30 | 2016-08-24 | 北京京东尚科信息技术有限公司 | Database data classification method and apparatus |
| CN106482502B (en) * | 2016-10-10 | 2019-01-15 | 重庆科技学院 | The intelligence drying long-range control method and system recommended based on cloud platform big data |
| CN108170770A (en) * | 2017-12-26 | 2018-06-15 | 山东联科云计算股份有限公司 | A kind of analyzing and training platform based on big data |
| CN111694844B (en) * | 2020-05-28 | 2024-05-07 | 平安科技（深圳）有限公司 | Enterprise operation data analysis method and device based on configuration algorithm and electronic equipment |
| CN112199376B (en) * | 2020-11-05 | 2021-07-20 | 北京三维天地科技股份有限公司 | Standard knowledge base management method and system based on cluster analysis |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101169798A (en) * | 2007-12-06 | 2008-04-30 | 中国电信股份有限公司 | Data excavation system and method |
| CN103218691A (en) * | 2013-04-26 | 2013-07-24 | 吉林市赢科信息技术有限责任公司 | Embedded type business intelligent information management system and management method |
| CN103229198A (en) * | 2010-11-29 | 2013-07-31 | 国际商业机器公司 | Fast, dynamic, data-driven report deployment of data mining and predictive insight into business intelligence (BI) tools |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6636860B2 (en) * | 2001-04-26 | 2003-10-21 | International Business Machines Corporation | Method and system for data mining automation in domain-specific analytic applications |
-
2013
- 2013-10-30 CN CN201310530032.8A patent/CN103544299B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101169798A (en) * | 2007-12-06 | 2008-04-30 | 中国电信股份有限公司 | Data excavation system and method |
| CN103229198A (en) * | 2010-11-29 | 2013-07-31 | 国际商业机器公司 | Fast, dynamic, data-driven report deployment of data mining and predictive insight into business intelligence (BI) tools |
| CN103218691A (en) * | 2013-04-26 | 2013-07-24 | 吉林市赢科信息技术有限责任公司 | Embedded type business intelligent information management system and management method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103544299A (en) | 2014-01-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103544299B (en) | A kind of construction method of business intelligence cloud computing system | |
| JP7090936B2 (en) | ESG-based corporate evaluation execution device and its operation method | |
| Li et al. | Analyzing and predicting question quality in community question answering services | |
| CN109657947B (en) | Enterprise industry classification-oriented anomaly detection method | |
| CN105069122A (en) | Personalized recommendation method and recommendation apparatus based on user behaviors | |
| CN112835570A (en) | Machine learning-based visual mathematical modeling method and system | |
| CN115547466B (en) | Medical institution registration and review system and method based on big data | |
| CN112734569A (en) | Stock risk prediction method and system based on user portrait and knowledge graph | |
| Bhatia et al. | Machine Learning with R Cookbook: Analyze data and build predictive models | |
| CN116645129A (en) | Manufacturing resource recommendation method based on knowledge graph | |
| US9305261B2 (en) | Knowledge management engine for a knowledge management system | |
| US7899776B2 (en) | Explaining changes in measures thru data mining | |
| He et al. | Word embedding based document similarity for the inferring of penalty | |
| CN110059749B (en) | Method and device for screening important features and electronic equipment | |
| Jeyaraman et al. | Practical Machine Learning with R: Define, build, and evaluate machine learning models for real-world applications | |
| Yang et al. | Automatic machine learning-based OLAP measure detection for tabular data | |
| Viswanathan et al. | R: Recipes for analysis, visualization and machine learning | |
| CN103279549A (en) | Method and device for acquiring target data of target objects | |
| US20230126022A1 (en) | Automatically determining table locations and table cell types | |
| Venkateswara Rao et al. | The societal communication of the Q&A community on topic modeling | |
| CN116778210A (en) | Teaching image evaluation system and teaching image evaluation method | |
| CN113935819A (en) | Method for extracting checking abnormal features | |
| WO2021146175A1 (en) | Systems and method for dynamically updating materiality distributions and classifications | |
| Blanchard et al. | Data Science for Marketing Analytics: Achieve your marketing goals with the data analytics power of Python | |
| Sadeghi et al. | Developing a new assessment fuzzy model by focusing on improving the reliability of customers’ individual verbal judgment (An Internet Banking case study) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |