CN103838775B - Data analysing method and DAF - Google Patents
Data analysing method and DAF Download PDFInfo
- Publication number
- CN103838775B CN103838775B CN201210489648.0A CN201210489648A CN103838775B CN 103838775 B CN103838775 B CN 103838775B CN 201210489648 A CN201210489648 A CN 201210489648A CN 103838775 B CN103838775 B CN 103838775B
- Authority
- CN
- China
- Prior art keywords
- data
- recommending
- orderly
- user
- multiway tree
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
- G06N5/025—Extracting rules from data
Abstract
The invention discloses a kind of data analysing method, including：The recommending data set based on orderly multiway tree is built, one recommending data of each node on behalf of the orderly multiway tree connects the correlation that the weights of each node branch are represented between different recommending datas；Receive the first choice data from user；Determine position of the first choice data in the orderly multiway tree；And the position with where the first choice data is as father node, depth and/or breadth traversal are carried out in the recommending data set based on orderly multiway tree, so as to suitable one or more recommending datas of user output.The invention also discloses a kind of DAF.
Description
Technical field
The present invention relates to computer data analysis field, especially, it is related to generation recommending data to draw to user
The data analysing method and DAF led.
Background technology
Existing DAF is mostly based on Bayesian network model when analysis is predicted to build a shellfish
Leaf this network, to form digraph.But, the construction of Bayesian network be a task for complexity, it is necessary to knowledge engineer and
The participation of domain expert.Because its construction and analysis process are all sufficiently complex, it is easy to cause error.
Hidden Markov model may be considered a special case for Bayesian network model, and it is simpler in terms of structure
Single and convenience, it is not necessary to have the participation of knowledge engineer and domain expert.And its final result that builds is a chained list, from journey
It is more convenient and quick for sequence process angle.But, because hidden Markov model is a very rough simplification, so
To the less use hidden Markov model of forecast analysis accuracy requirement occasion higher.
The content of the invention
To solve the above problems, the present invention is from hidden Markov model carry out, and comprehensive Bayesian network model
Characteristic, propose brand-new a data analysing method and DAF based on orderly multiway tree.
According to an aspect of the invention, there is provided a kind of data analysing method, including：Build based on orderly multiway tree
Recommending data set, one recommending data of each node on behalf of the orderly multiway tree connects the power of each node branch
Value represents the correlation between different recommending datas；Receive the first choice data from user；Determine the first choice number
According to the position in the orderly multiway tree；And the position with where the first choice data is as father node, in the base
Depth and/or breadth traversal are carried out in the recommending data set of orderly multiway tree, to export suitable one to the user
Individual or multiple recommending datas.
Above-mentioned data analysing method may also include：Receive the second selection data from the user；Determine described second
Position of the selection data in the orderly multiway tree；And the position with where the first choice data is as father node, and
Position with where the described second selection data is carried out as child node in the recommending data set based on orderly multiway tree
Depth and/or breadth traversal, so as to suitable one or more recommending datas of user output.
In above-mentioned data analysing method, before one or more recommending datas are exported to the user, compare this one
Correlations individual or between multiple recommending datas and the first choice data.
In above-mentioned data analysing method, before one or more recommending datas are exported to the user, compare this one
Individual or multiple recommending datas select the correlation between data with the first choice data, described second.
In above-mentioned data analysing method, the recommending data set of the structure based on orderly multiway tree includes：Collect and use
Family historical selection data；User's history selection data are carried out regular；Data are selected to the user's history after regular
It is ranked up；And select data, recommendation number of the generation based on orderly multiway tree according to by the user's history after regular, sequence
According to set.
In above-mentioned data analysing method, the recommending data set of the structure based on orderly multiway tree also includes：From institute
State user's history selection data and select one or more data；With one or more of data as path, to the base for having generated
Traveled through in the recommending data set of orderly multiway tree；And it is each in the orderly multiway tree to adjust according to the result of traversal
The weights of branch.
It is important to point out that, above mentioned data analysing method is all applied to the analysis of Customer Shopping basket.
According to another aspect of the present invention, a kind of DAF is additionally provided, including：Construction device, for structure
The recommending data set based on orderly multiway tree is built, one recommending data of each node on behalf of the orderly multiway tree connects
Connect the correlation that the weights of each node branch are represented between different recommending datas；First receiving device, uses by oneself for receiving
The first choice data at family；First determining device, for determining position of the first choice data in the orderly multiway tree
Put；And first analytical equipment, for the position with where the first choice data as father node, described based on many in order
Carry out depth and/or breadth traversal in the recommending data set for pitching tree, so as to user output it is suitable one or more
Recommending data.
Above-mentioned DAF may also include：Second reception device, for receiving the second selection from the user
Data；Second determining device, for determining position of the second selection data in the orderly multiway tree；And second point
Analysis apparatus, for the position with where the first choice data as father node, and with the position where the described second selection data
Child node is set to, depth and/or breadth traversal is carried out in the recommending data set based on orderly multiway tree, so as to institute
State suitable one or more recommending datas of user's output.
In above-mentioned DAF, first analytical equipment also includes the first comparing unit, and described first compares
Unit is used to compare the correlation between one or more recommending datas and the first choice data.
In above-mentioned DAF, second analytical equipment also includes the second comparing unit, and described second compares
Unit is used to compare related between one or more recommending datas and the first choice data, the second selection data
Property.
In above-mentioned DAF, the construction device includes：Collector unit, for collecting user's history selection number
According to；Regular unit, it is regular for being carried out to user's history selection data；Sequencing unit, for the use after regular
Family historical selection data is ranked up；And generation unit, for selecting number according to by the user's history after regular, sequence
According to recommending data set of the generation based on orderly multiway tree.
In above-mentioned DAF, the construction device also includes：Select unit, for being selected from the user's history
Select data and select one or more data；Traversal Unit, for one or more of data as path, to the base for having generated
Traveled through in the recommending data set of orderly multiway tree；And adjustment unit, described in being adjusted according to the result for traveling through
The weights of orderly multiway tree Zhong Ge branches.
It is important to point out that, above mentioned DAF is all applied to the analysis of Customer Shopping basket.
The basic thought of the present inventor's foundation hidden Markov model, and Bayesian network model is merged, carry
Go out brand-new a data analysing method and DAF based on orderly multiway tree.The data analysing method and data point
Desorption device has done good balance between the two in computation complexity and the degree of accuracy, effectively improves inside computer system
Energy.
Brief description of the drawings
After specific embodiment of the invention has been read referring to the drawings, those skilled in the art will be more clearly
Solution various aspects of the invention.Skilled person would appreciate that：These accompanying drawings are used only for coordinating specific embodiment party
Formula illustrates technical scheme, and is not intended to be construed as limiting protection scope of the present invention.
Fig. 1 is the schematic diagram of data analysing method according to an embodiment of the invention；
Fig. 2 is the schematic diagram of DAF according to an embodiment of the invention；
Fig. 3 to Fig. 5 is the logical schematic of data analysing method according to an embodiment of the invention.
Specific embodiment
What is be described below is multiple some that may be in embodiment of the invention, it is desirable to provide to of the invention basic
Solution, it is no intended to confirm of the invention crucial or conclusive key element or limit scope of the claimed.It is readily appreciated that, according to this
The technical scheme of invention, in the case where connotation of the invention is not changed, those of ordinary skill in the art can be proposed can be mutual
Other implementations replaced.Therefore, detailed description below and accompanying drawing are only the examples to technical scheme
Property explanation, and be not to be construed as whole of the invention or be considered as to define or limit technical solution of the present invention.
With reference to Fig. 1, it shows a kind of data analysing method, including：Step 110, builds pushing away based on orderly multiway tree
Data acquisition system is recommended, one recommending data of each node on behalf of the orderly multiway tree connects the weights of each node branch
Represent the correlation between different recommending datas；Step 120, receives the first choice data from user；Step 130, it is determined that
Position of the first choice data in the orderly multiway tree；And step 140, with where the first choice data
Position is father node, and depth and/or breadth traversal are carried out in the recommending data set based on orderly multiway tree, so as to
Described suitable one or more recommending datas of user's output.
In a specific embodiment, building the recommending data set based on orderly multiway tree may include following step
Suddenly：Collect user's history selection data；User's history selection data are carried out regular；Data are selected to the user's history after regular
It is ranked up；And select data, recommendation number of the generation based on orderly multiway tree according to by the user's history after regular, sequence
According to set.
After recommending data set of the generation based on orderly multiway tree, the recommending data set for generating can also be carried out
Amendment.In a specific embodiment, build the recommending data set based on orderly multiway tree and may also include：Gone through from user
History selection data select one or more data；With one or more data as path, to having generated based on orderly multi-fork
The recommending data set of tree is traveled through；And the weights of orderly multiway tree Zhong Ge branches are adjusted according to the result of traversal.
In a preferred embodiment, before one or more recommending datas are exported to user, compare this or
Correlation between multiple recommending datas and first choice data（Connect the weights of branch）.Compared by this, it may be determined which
A little recommending datas are more related to first choice data, such that it is able to selectively export suitable recommending data to user.
In practice, DAF or other devices to suitable one or more recommending datas of user's output it
Afterwards, user may make further selection.So, data analysing method above-mentioned may also include：Receive from institute
State the second selection data of user；Determine position of the second selection data in the orderly multiway tree；And with described
Position where first choice data is father node, and position with where the described second selection data is as child node, described
Depth and/or breadth traversal are carried out in recommending data set based on orderly multiway tree, so as to suitable to user output
One or more recommending datas.
In a preferred embodiment, before one or more recommending datas are exported to user, compare this or
Correlation between multiple recommending datas and first choice data, the second selection data（Connect the weights of branch）.By this
Compare, it may be determined which recommending data and first choice data, the second selection data are more related, such that it is able to selectively to
User exports suitable recommending data.
With reference to Fig. 2, it shows a kind of DAF 200.The DAF 200 include construction device 210,
First receiving device 220, the first determining device 230 and the first analytical equipment 240.Construction device 210 is based on for structure to be had
The recommending data set of sequence multiway tree, one recommending data of each node on behalf of the orderly multiway tree connects each node
The weights of branch represent correlation between different recommending datas.First receiving device 220 is used to receive from user the
One selection data.First determining device 230 is used to determine position of the first choice data in the orderly multiway tree.The
Position that one analytical equipment 240 is used for where the first choice data as father node, described based on orderly multiway tree
Depth and/or breadth traversal are carried out in recommending data set, so as to suitable one or more the recommendation numbers of user output
According to.
In a specific embodiment, construction device 210 may include with lower unit：Collector unit, uses for collecting
Family historical selection data；Regular unit, it is regular for being carried out to user's history selection data；Sequencing unit, for after regular
User's history selection data be ranked up；And generation unit, for according to by the user's history selection after regular, sequence
Data, recommending data set of the generation based on orderly multiway tree.
After generation unit recommending data set of the generation based on orderly multiway tree, can also be to the recommending data of generation
Set is modified.In a specific embodiment, construction device 210 may also include：Select unit, for being gone through from user
History selection data select one or more data；Traversal Unit, for one or more data as path, to what is generated
Recommending data set based on orderly multiway tree is traveled through；And adjustment unit, adjusted for the result according to traversal
The weights of sequence multiway tree Zhong Ge branches.
In a preferred embodiment, the first analytical equipment 240 also includes the first comparing unit（Not shown in Fig. 2）.
First comparing unit compared one or more recommending datas and before one or more recommending datas are exported to user
Correlation between one selection data（Connect the weights of branch）.Which compared by this, it may be determined that recommending data and first
Selection data are more related, such that it is able to selectively export suitable recommending data to user.
In practice, in DAF 200 or other devices to suitable one or more the recommendation numbers of user's output
After, user may make further selection.So, DAF above-mentioned may also include：Second connects
Receiving apparatus, for receiving the second selection data from the user；Second determining device, for determining the second selection number
According to the position in the orderly multiway tree；And second analytical equipment, for the position where the first choice data
Be father node, and with described second selection data where position as child node, in the recommendation number based on orderly multiway tree
According to depth and/or breadth traversal is carried out in set, so as to suitable one or more recommending datas of user output.
In a preferred embodiment, the second analytical equipment may also include the second comparing unit.Second comparing unit
Can before one or more recommending datas are exported to user, compare one or more recommending datas and first choice data,
Correlation between second selection data（Connect the weights of branch）.Which compared by this, it may be determined that recommending data and
One selection data, the second selection data are more related, such that it is able to selectively export suitable recommending data to user.
Data analysing method and DAF based on orderly multiway tree more than, can be in computation complexity and standard
Exactness makes balance well between the two, so as to effectively improve inside computer system performance.
In a specific embodiment, data analysing method and data based on orderly multiway tree above-mentioned point
Desorption device can be applied to the analysis of Customer Shopping basket.
Before Customer Shopping basket analysis process is specifically introduced, it is necessary to first some technical terms are explained.
Term " historical trading data collection system " refers to a persistence platform for historical trading data.The platform is external
There is provided the query interface specific manifestation of historical trading can be：The collection of database, data warehouse, data file or data file
Close.
Term " type of merchandize analysis system " refers to a type of merchandize maintenance system, and it is mainly responsible in whole system
Involved all commodity carry out Classification Management according to certain rule.
Term " advertisement recommending data " is the recommending data stored with orderly multiway tree, and its storage location does not do specific limit
System, can be internal memory, database, single file or multiple files.Each node of the orderly multiway tree is a business
The description of product, all sections included between the first level of child nodes of the orderly multiway tree to the node any one terminal node
The subset of the purchase commodity of at least one user of set of point.After the commodity that user have purchased father node, this is bought orderly
The probability of commodity must not drop below the probability of its right brotgher of node representated by the left child node of multiway tree.It is of course also possible to according to
Situation is come orderly multiway tree as setting：After the commodity that user have purchased father node, the right son of the orderly multiway tree is bought
The probability of commodity representated by node must not drop below the probability of its left brotgher of node.The degree of the orderly multi-fork root vertex can not surpass
Cross the species of the commodity that type of merchandize analysis system is safeguarded.
Term " the regular system of transaction data " is come to historical trading by the regular parameter configuration files of transaction data
Data carry out it is regular, to form a system for logic single shopping basket list.
Term " the regular parameter configuration files of transaction data " is used for the constraint of the item property for describing to meet single shopping basket
Condition, a configuration file that can be for reference is described as follows shown in table：
| Whether exclusion is returned goods | |
| Exchange hour scope | |
| Whether ship-to is identical | |
| Whether order IP is identical | |
| … | |
| Exclude repeat buying |
Term " single shopping basket commodity sort generator " is to sort coefficient configuration file to shopping basket according to type of merchandize
In commodity be ranked up, so as to form a maker for unique shopping basket commodity sequence.
Term " type of merchandize sequence coefficient configuration file " is used for describing the weights list of shopping basket commodity ordering attribute, one
Individual configuration file that can be for reference is described as follows shown in table：
| Trade name | |
| Item property | |
| Brand | |
| Commodity price | |
| … | |
| Time buying |
Term " ad data generation system ", as the term suggests, that is, the system for generating ad data.The execution step of the system
It is divided into following three step：The first step, takes out first commodity, the first of advertisement recommending data from sorted shopping basket
Searched in level child node, if do not found, insert the commodity data to the first order child node of advertisement recommending data
Most right (left side) side；If finding the commodity comprising the node, the weights of adjustment advertisement recommending data root node to the node,
And first order child node is ranked up.Second step, takes out the n-th (n>1) part commodity, from first order in advertisement recommending data
It is father node A that node is begun look for first commodity, right with its subsequent article is its child node (n-1)th commodity node N
The child node of the node is searched, if finding n-th commodity, the weights of concept transfer N to the node, and to node
N is ranked up；Otherwise insert n-th commodity data to most right (left side) side of node N.3rd step, repeats second step, until commodity
All commodity in basket take out and finish.
Term " commodity sequencing weight coefficient configuration file " is used for describing path of the commodity in ad data generation system
Weights, a configuration file that can be for reference is described as follows shown in table：
| Customer name | |
| Affiliated depth | |
| Trade name | |
| Item property | |
| … | |
| Time buying |
Term " shopping basket commercial articles searching device " is entered in advertisement recommending data according to the shopping basket commodity path sorted
Row maximizes matching, and the searcher being compared to matching result.Commercial articles searching knot is generated if search effect is not up to standard
Fruit coefficient of deviation table.One of commercial articles searching result coefficient of deviation table can be described as follows shown in table for reference：
| Customer name | |
| Affiliated depth | |
| Trade name | |
| Item property | |
| … | |
| Time buying | |
| Recommended location information | |
| Expect recommended location information |
Term " commodity sequence coefficients generator " is the maker for generating commodity sequence coefficient, and the maker is according to purchase
The Search Results coefficient of deviation table that thing basket commercial articles searching device is provided is modified to type of merchandize sequence coefficient configuration file.
Term " shopping basket commercial product recommending device " is a commercial product recommending device, and the commercial product recommending device is according to single shopping basket commodity
The selected items list of client that sort generator is generated generates path and is searched in advertisement recommending data according to it, finds
The node N of selected last part commodity, is had concurrently with node N as root node carries out extreme saturation or breadth traversal or both,
One Recommendations list is generated according to traversing result.
Term " Recommendations list " is the inventory that can finally be provided to user, and the inventory is available for user to enter
Row commodity selection.One of Recommendations list can be described as follows for reference：
| Optimal purchase inventory | The right side (left side) descendant nodes to node N carry out an extreme saturation |
| Most probable buys inventory | To the child node of node N since right (left side) a breadth traversal |
In upper table, optimal purchase inventory needs to carry out an extreme saturation by the right side (left side) descendant nodes to node N
To obtain, and most probable purchase inventory then needs to the child node of node N that a breadth traversal is obtained since right (left side).
Term " history single shopping basket commodity generate system " is that historical trading data is captured, so as to extract it
In original some the true market basket datas for meeting shopping basket definition system.
Customer Shopping basket analysis process will be below described using above-mentioned term.
The structure of forecasting system
One complete forecasting system for carrying out Customer Shopping basket analysis can be built by following four step.
First, the basic framework preparation of ad data is built according to type of merchandize analysis system, by historical trading
Data gathering system carries out output preparation to historical trading data.
Second, it is driven by transaction history data collection system, the regular system of transaction data is regular using transaction data
The rule that parameter configuration files are provided carries out regular to the historical trading data that transaction history data collection system is provided, and generates
One virtual shopping basket.
3rd, it is driven by the regular system of transaction data, single shopping basket commodity sort generator is according to type of merchandize
The configuration information that sequence coefficient configuration file is provided is carried out to the commodity in the virtual shopping basket of the regular system generation of transaction data
Sequence work.
4th, driven by single shopping basket commodity sort generator, ad data generates system according to commodity sequencing weight
The configuration data that coefficient configuration file is provided carries out commodity insertion work to advertisement recommending data.
The structure of four steps more than, a complete prediction system tentatively builds the transfer of the data between completion, each system
By can be as shown in Figure 3.
The use of forecasting system
The use of forecasting system is a process for moving in circles, and the starting of system depends on a choosing for commodity of user
It is fixed, the output of Recommendations list is ended in, but the whole shopping process of user is not terminated with the end of system.Each system
The transfer of the data between system is by as shown in Figure 4.Specifically, the use of forecasting system includes following steps：
First, after user selects first commodity, a virtual shopping basket maker is generated by the commodity.Work as later
After user often selects a commodity, the commodity that will all choose are put into the virtual shopping basket, and adding procedure is all each time
The generation of following second, third and four steps will be caused.
Second, it is driven by virtual shopping basket maker, single shopping basket commodity sort generator passes through type of merchandize
Sequence coefficient configuration file is processed the virtual shopping basket.
3rd, it is driven by single shopping basket commodity sort generator, the generation result of the maker is transferred to shopping
Basket commercial product recommending device is processed in advertisement recommending data and is generated Recommendations list.
4th, displaying Recommendations list buys next commodity as guidance to user to user.
The amendment of forecasting system
The amendment of forecasting system belongs to a process for iteration study, it is necessary to repeat until without Search Results
Untill coefficient of deviation table is generated.After Search Results coefficient of deviation table is not regenerated, sort coefficient configuration file to type of merchandize
Contrasted with the type of merchandize sequence coefficient configuration file used when carrying out system constructing, if content is different, it is right to need
System is rebuild, and the transfer of the data between each system is by as shown in Figure 5.Specifically, the amendment bag of forecasting system
Include following steps：
First, history single shopping basket commodity generate system by the original number to history shopping basket in historical trading data
According to extraction, obtain one group of authentic and valid historical trading basket data object.
Second, it is driven by history single shopping basket commodity generation system, single shopping basket commodity sort generator leads to
Cross type of merchandize sequence coefficient configuration file and generate a commodity traverse path.
3rd, it is driven by single shopping basket commodity sort generator, shopping basket commercial articles searching device is done shopping according to single
The commodity traverse path that basket commodity sort generator is provided is scanned for involved commodity.And Search Results are sentenced
It is fixed, if meet search be expected, release subsequent step.Otherwise generating Search Results coefficient of deviation table and continuation is carried out downwards.
4th, it is driven by shopping basket commercial articles searching device, commodity sort coefficients generator according to shopping basket commercial articles searching
Information in the Search Results coefficient of deviation table of device generation is modified to type of merchandize sequence coefficient configuration file
The algorithm the convergence speed for carrying out Customer Shopping basket analysis according to the method described above is only needed to historic customer quickly, typically
Shopping basket scan with this.Also, the structure of Customer Shopping basket analysis system is relatively easy, whole process to knowledge engineer and
The participation dependence of domain expert is smaller.
Certainly, it will be understood by those skilled in the art that data analysing method as herein described and DAF can also be amplified
To in other decision systems, such as GPS satellite navigation.Specifically, GPS satellite navigation equipment can in advance build one based on having
The recommending data model of sequence multiway tree, then GPS satellite navigation equipment receive the first choice from user（For example, selection mesh
Ground be A）, then GPS satellite navigation equipment its position in orderly multiway tree is determined according to the selected destination A of user
Put, last GPS satellite navigation equipment provides a user with optimal course by carrying out extreme saturation to orderly multiway tree.
Proposed by the invention data analysing method and DAF based on orderly multiway tree being capable of effective Horizons
Weighing apparatus both computation complexity and the degree of accuracy, so as to improve inside computer system performance.
Above, specific embodiment of the invention is described with reference to the accompanying drawings.But, those skilled in the art
It is understood that without departing from the spirit and scope of the present invention, can also make each to specific embodiment of the invention
Plant change and replace.These changes and replacement all fall in claims of the present invention limited range.
Claims (9)
1. a kind of data analysing method, including：
Build the recommending data set based on orderly multiway tree, one recommendation number of each node on behalf of the orderly multiway tree
According to the weights of each node branch of connection represent the correlation between different recommending datas；
Receive the first choice data from user；
Determine position of the first choice data in the orderly multiway tree；And
Position with where the first choice data as father node, in the recommending data set based on orderly multiway tree
Depth and/or breadth traversal are carried out, to export suitable one or more recommending datas to the user,
Wherein, the recommending data set of the structure based on orderly multiway tree includes：
Collect user's history selection data；
User's history selection data are carried out regular；
User's history selection data after regular are ranked up；And
Data, recommending data set of the generation based on orderly multiway tree are selected according to by the user's history after regular, sequence.
2. data analysing method as claimed in claim 1, also includes：
Receive the second selection data from the user；
Determine position of the second selection data in the orderly multiway tree；And
Position with where the first choice data as father node, and with described second selection data where position for son section
Point, depth and/or breadth traversal are carried out in the recommending data set based on orderly multiway tree, so as to defeated to the user
Go out suitable one or more recommending datas.
3. data analysing method as claimed in claim 1, wherein, to the user export one or more recommending datas it
Before, compare the correlation between one or more recommending datas and the first choice data.
4. data analysing method as claimed in claim 2, wherein, to the user export one or more recommending datas it
Before, compare the correlation between one or more recommending datas and the first choice data, the second selection data.
5. data analysing method as claimed in claim 1, wherein, it is described to build the recommending data set based on orderly multiway tree
Also include：
One or more data are selected from user's history selection data；
With one or more of data as path, the recommending data set based on orderly multiway tree to having generated is carried out time
Go through；And
The weights of the orderly multiway tree Zhong Ge branches are adjusted according to the result for traveling through.
6. a kind of DAF, including：
Construction device, for building the recommending data set based on orderly multiway tree, each node of the orderly multiway tree
A recommending data is represented, the correlation that the weights of each node branch are represented between different recommending datas is connected；
First receiving device, for receiving the first choice data from user；
First determining device, for determining position of the first choice data in the orderly multiway tree；And
First analytical equipment, for the position with where the first choice data as father node, described based on orderly multi-fork
Carry out depth and/or breadth traversal in the recommending data set of tree, so as to user output it is suitable one or more push away
Data are recommended, the construction device includes：
Collector unit, for collecting user's history selection data；
Regular unit, it is regular for being carried out to user's history selection data；
Sequencing unit, for being ranked up to the user's history selection data after regular；And
Generation unit, for selecting data, generation pushing away based on orderly multiway tree according to by the user's history after regular, sequence
Recommend data acquisition system.
7. DAF as claimed in claim 6, also includes：
Second reception device, for receiving the second selection data from the user；
Second determining device, for determining position of the second selection data in the orderly multiway tree；And
Second analytical equipment, for the position with where the first choice data as father node, and with the described second selection number
It is child node according to the position at place, depth and/or range time is carried out in the recommending data set based on orderly multiway tree
Go through, so as to suitable one or more recommending datas of user output.
8. DAF as claimed in claim 6, wherein, first analytical equipment also includes the first comparing unit,
First comparing unit is used to compare the correlation between one or more recommending datas and the first choice data.
9. DAF as claimed in claim 7, wherein, second analytical equipment also includes the second comparing unit,
Second comparing unit is used to compare one or more recommending datas with the first choice data, the second selection number
Correlation between.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210489648.0A CN103838775B (en) | 2012-11-27 | 2012-11-27 | Data analysing method and DAF |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210489648.0A CN103838775B (en) | 2012-11-27 | 2012-11-27 | Data analysing method and DAF |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103838775A CN103838775A (en) | 2014-06-04 |
| CN103838775B true CN103838775B (en) | 2017-06-06 |
Family
ID=50802284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210489648.0A Active CN103838775B (en) | 2012-11-27 | 2012-11-27 | Data analysing method and DAF |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103838775B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893697A (en) * | 2016-04-22 | 2016-08-24 | 北京交通大学 | System reliability assessment method based on Bayesian network reasoning |
| CN109325027A (en) * | 2018-08-21 | 2019-02-12 | 朱常林 | One kind is based on the analysis of cloud data, Situation Awareness algorithm |
| CN110825750A (en) * | 2019-11-07 | 2020-02-21 | 杭州电弧数据科技有限公司 | Big data storage service method |
| CN112285131A (en) * | 2020-11-02 | 2021-01-29 | 泰州帕沃能源科技有限公司 | Method for judging quality of X-ray image for detecting power switch |
| CN113781087A (en) * | 2021-01-29 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Recall method and device of recommended object, storage medium and electronic equipment |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101986299A (en) * | 2010-10-28 | 2011-03-16 | 浙江大学 | Multi-task personalized web service method based on hypergraph |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090094285A1 (en) * | 2007-10-03 | 2009-04-09 | Mackle Edward G | Recommendation apparatus |
-
2012
- 2012-11-27 CN CN201210489648.0A patent/CN103838775B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101986299A (en) * | 2010-10-28 | 2011-03-16 | 浙江大学 | Multi-task personalized web service method based on hypergraph |
Non-Patent Citations (1)
| Title |
|---|
| 基于加权关联规则和浏览行为的个性化推荐;陈华月;《中国优秀硕士学位论文全文数据库信息科技辑》;20060115(第01期);第1-51页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103838775A (en) | 2014-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103838775B (en) | Data analysing method and DAF | |
| Adulyasak et al. | Optimization-based adaptive large neighborhood search for the production routing problem | |
| CN110458627A (en) | A kind of commodity sequence personalized recommendation method of user oriented preference of dynamic | |
| CN106250568B (en) | Service range determines method, logistics service provider recommended method and related device | |
| CN105574025B (en) | For calculating sequence point and establishing the method, apparatus and commercial product recommending system of model | |
| KR20200088131A (en) | Item recommendation method and apparatus using graph database | |
| CN108121737A (en) | A kind of generation method, the device and system of business object attribute-bit | |
| CN103548019A (en) | Method and system for providing statistical from a data warehouse | |
| US7680697B2 (en) | Searching for a seller of a product | |
| CN102419779A (en) | Method and device for personalized searching of commodities sequenced based on attributes | |
| CN102609869A (en) | Commodity purchasing system and method | |
| CN105138690A (en) | Method and device for determining keywords | |
| CN103903148A (en) | C2C system, B2C system, and application methods thereof | |
| CN106846088A (en) | A kind of Method of Commodity Recommendation of the product electric business website that disappears soon | |
| Hajiaghaei-Keshteli et al. | Addressing the freight consolidation and containerization problem by recent and hybridized meta-heuristic algorithms | |
| Aljubayrin et al. | Skyline trips of multiple POIs categories | |
| KR102538398B1 (en) | Method, Apparatus and System for Recommendation in Groups Using Bigdata | |
| CN106502881A (en) | A kind of method of testing of commodity ordering rule and device | |
| GÜVEN et al. | Machine Learning Algorithms with Intermittent Demand Forecasting: An Application in Retail Apparel with Plenty of Predictors | |
| KR20220001616A (en) | Method, Apparatus and System for Constructing Bigdata Based on Generating United Identifier of Customer | |
| CN112559807A (en) | Graph pattern matching method based on multi-source point parallel exploration | |
| KR20200088132A (en) | Item recommendataion method and apparatus between group member and group item | |
| CN104462480B (en) | Comment big data method for digging based on typicalness | |
| Eisner et al. | Sequenced route queries: getting things done on the way back home | |
| Pennacchioli et al. | Overlap versus partition: marketing classification and customer profiling in complex networks of products |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |