CN106127493A - A kind of method and device analyzing customer transaction behavior - Google Patents
A kind of method and device analyzing customer transaction behavior Download PDFInfo
- Publication number
- CN106127493A CN106127493A CN201610460883.3A CN201610460883A CN106127493A CN 106127493 A CN106127493 A CN 106127493A CN 201610460883 A CN201610460883 A CN 201610460883A CN 106127493 A CN106127493 A CN 106127493A
- Authority
- CN
- China
- Prior art keywords
- commodity
- tree
- transaction
- data
- 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.)
- Pending
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
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of method and device analyzing customer transaction behavior, by user's purchasing behavior data compression for buying tree data, buy tree data and be more conducive to storage and the process of the big data of user, improve the effect of cluster after data reasonably being compressed, be greatly improved actual application value；Simultaneously, in the clustering method of tree is bought in transaction, the method using spectral clustering, the problem dexterously problem of a NP difficulty being converted into Laplacian Matrix eigenvalue (vectorial), discrete clustering problem being relaxed as continuous print characteristic vector, minimum series of features vector correspond to the serial division methods that figure is optimum, and remaining is only by the problem of laxization discretization again, will characteristic vector is subdivided opens, just can obtain corresponding classification.The cluster result obtained by above procedure, has not only been evaded the dependence to user's static attribute of traditional cluster, and user data has carried out reasonable dimensionality reduction and has obtained good user grouping result.
Description
Technical field
The present embodiments relate to the technical field of customer transaction, particularly relate to a kind of method analyzing customer transaction behavior
And device.
Background technology
At present, Customer Shopping analysis is the most increasingly paid attention to by retailer.It is association rule that tree analysis is bought in transaction
Then in an important application of retail business, it, by finding the contact between customer purchase commodity, analyzes the purchase row of client
For and assist retailer formulate marketing strategy.The tree of buying generally said analyzes the friendship referred to by showing in purchase tree
Easily information analyzes the purchasing behavior of client, client during buying commodity would generally the multiple commodity of single purchase, thus
Make to have between these commodity the strongest relatedness.Therefore, the purchasing behavior that one can consider that client is a kind of entirety
The most whether behavior, buy commodity and influence whether the purchase of other commodity, thus have influence on each profit buying tree.Institute
With, the target buying tree analysis finds out important and valuable purchase tree exactly, analyzes client frequent from retail records
The combination of the commodity simultaneously bought, thus the beneficially sales promotion of commodity, rocker, logistics etc..
Current tree analytical technology of buying is mainly based upon the purchase tree analysis method of clustering algorithm.Wherein, application is the most
Being exactly HAC (hierarchical agglomerative clustering) algorithm widely, it mainly includes classification and surveys
Away from two processes.The hierarchical clustering of cohesion is a kind of bottom-up strategy, first using each object as one bunch, then closes
And these clusters are increasing bunch, until all of object is all in one bunch, or certain finish condition is satisfied,
Most hierarchy clustering methods belong to this class, they simply bunch between similarity definition on different.By each right
As being classified as a class, there are N class, every class only comprises an object.Distance between class and class is exactly the object that they are comprised
Between distance.Finding immediate two classes and be merged into a class, the most total class number has lacked one.Recalculate new class
And the distance between had been friends in the past class.Repeat step above, be to the last merged into till a class that (this type of contains N number of right
As).
The static attribute information such as traditional clustering method geographical position based on client, demographic characteristics, cluster result is not
Necessarily there is an identical purchasing behavior, and achievement data has the advantages that private ownership is difficult to obtain, cause obtaining good
Clustering Effect.Along with the arrival of information age, being analyzed processing to substantial amounts of data is a job the hugest, and this just closes
The problem being tied to a computational efficiency.When the most many clustering methods process small-scale data and low-dimensional data, Performance comparision is good,
But when data scale increases, and dimension raises, performance will drastically decline, such as conventional process data time property on a small scale
Very well, but can be as data volume and increase, efficiency is just gradually reduced, and real-life data major part broadly falls into scale
The data set that bigger, dimensional comparison is high.
Summary of the invention
The purpose of the embodiment of the present invention is to propose a kind of method and device analyzing customer transaction behavior, it is intended to solve such as
The problem what improves Clustering Effect.
For reaching this purpose, the embodiment of the present invention by the following technical solutions:
First aspect, a kind of method analyzing customer transaction behavior, described method includes:
Reading transaction data from retailer data base, described transaction data includes Transaction Identification Number, exchange hour, trade name
Claim, parent title belonging to sales volume and commodity；
According to described Transaction Identification Number, described transaction data being carried out packet aggregation, described transaction data is the complete of transaction record
Collection, the transaction record of one commodity of every behavior of described transaction data；
Described transaction data is divided in groups according to client, merges set up the purchase tree of each user to often organizing data；From
In described transaction record, extraction is without repeating exhaustively commodity, according to described commodity with hierarchical information set up commodity tree；
Each user buys the similarity between tree and uses similar matrix to represent, uses spectral clustering to enter Laplacian Matrix
Row cluster, circulates cluster process, chooses best cluster result, described best cluster result from least one cluster result
For user grouping result.
Preferably, described according to described Transaction Identification Number, described transaction data is carried out packet aggregation, including:
The form of every transaction record is T=[TID, < i1,i2...,in>] form；
Wherein, T is for once to conclude the business, and TID is Transaction Identification Number, inFor certain commodity, < i1,i2...,in> it is that this transaction is purchased
The commodity set bought.
Preferably, described from described transaction record extraction without repeating exhaustively commodity, according to described commodity with
Hierarchical information sets up commodity tree, including:
Being numbered parent title belonging to described trade name and described commodity, from transaction record set, extraction is without weight
Multiple exhaustively commodity set；
The sub-categories relation carried according to described commodity, top-down sets up a commodity tree, and in tree, each node contains
Having a key-value pair, key is trade name or item name, is worth for reference numeral.
Preferably, described to described transaction data according to client divide in groups, to often organize data merge set up each user
Purchase tree, including:
Described transaction data is divided in groups by different client, sets up one and buy tree often organizing commodity successively, described
The each node bought in tree contains number value.
Preferably, described each user buys the similarity between tree and uses similar matrix to represent, uses spectral clustering to drawing
This matrix of pula clusters, including:
(i j) calculates user and buys the similarity between tree i and j, added up by each column element to use similarity matrix S
Put composition N*N similarity matrix S on the diagonal；
Similarity matrix S is transformed into Laplacian Matrix L, obtains front k eigenvalue and characteristic of correspondence vector,
Arrange described eigenvalue and the matrix of described characteristic vector one N*k of composition；
By the dimensionality reduction mode dimensionality reduction of laplacian eigenmaps, the characteristic vector obtained is carried out K-means cluster.
Second aspect, a kind of device analyzing customer transaction behavior, described device includes:
Read module, for reading transaction data from retailer data base, described transaction data includes Transaction Identification Number, friendship
Easily parent title belonging to time, trade name, sales volume and commodity；
Grouping module, for described transaction data being carried out packet aggregation according to described Transaction Identification Number, described transaction data is
The complete or collected works of transaction record, the transaction record of one commodity of every behavior of described transaction data；
Set up module, for described transaction data is divided in groups according to client, set up each use to often organizing data merging
The purchase tree at family；From described transaction record, extraction is without repeating exhaustively commodity, according to described commodity with hierarchical information
Set up commodity tree；
Cluster module, buys the similarity between tree for each user and uses similar matrix to represent, use spectral clustering pair
Laplacian Matrix clusters；
Choose module, be used for circulating cluster process, from least one cluster result, choose best cluster result, described
Best cluster result is user grouping result.
Preferably, described grouping module, it is used for:
The form of every transaction record is T=[TID, < i1,i2...,in>] form；
Wherein, T is for once to conclude the business, and TID is Transaction Identification Number, inFor certain commodity, < i1,i2...,in> it is that this transaction is purchased
The commodity set bought.
Preferably, described set up module, be used for:
Being numbered parent title belonging to described trade name and described commodity, from transaction record set, extraction is without weight
Multiple exhaustively commodity set；
The sub-categories relation carried according to described commodity, top-down sets up a commodity tree, and in tree, each node contains
Having a key-value pair, key is trade name or item name, is worth for reference numeral.
Preferably, described set up module, be additionally operable to:
Described transaction data is divided in groups by different client, sets up one and buy tree often organizing commodity successively, described
The each node bought in tree contains number value.
Preferably, described cluster module, it is used for:
(i j) calculates user and buys the similarity between tree i and j, added up by each column element to use similarity matrix S
Put composition N*N similarity matrix S on the diagonal；
Similarity matrix S is transformed into Laplacian Matrix L, obtains front k eigenvalue and characteristic of correspondence vector,
Arrange described eigenvalue and the matrix of described characteristic vector one N*k of composition；
By the dimensionality reduction mode dimensionality reduction of laplacian eigenmaps, the characteristic vector obtained is carried out K-means cluster.
The embodiment of the present invention provides a kind of method and device analyzing customer transaction behavior, by user's purchasing behavior data pressure
It is condensed to buy tree data, buys tree data and be more conducive to storage and the process of the big data of user, after data reasonably being compressed
Improve the effect of cluster, be greatly improved actual application value；Meanwhile, in the clustering method of tree is bought in transaction, use spectral clustering
Method, the problem dexterously problem of a NP difficulty being converted into Laplacian Matrix eigenvalue (vectorial), by discrete
Clustering problem relaxes as continuous print characteristic vector, and minimum series of features vector correspond to the serial division methods that figure is optimum, surplus
Under be only by the problem of laxization discretization again, will characteristic vector is subdivided opens, just can obtain corresponding classification.Pass through
The cluster result that above procedure obtains, has not only evaded the dependence to user's static attribute of traditional cluster, and to user data
Carry out reasonable dimensionality reduction and obtain good user grouping result.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet that the embodiment of the present invention provides a kind of method analyzing customer transaction behavior；
Fig. 2 is the structural representation of a kind of commodity tree that the embodiment of the present invention provides；
Fig. 3 is that a kind of transaction that the embodiment of the present invention provides buys the structural representation of tree；
Fig. 4 is the high-level schematic functional block diagram of a kind of device analyzing customer transaction behavior that the embodiment of the present invention provides.
Detailed description of the invention
With embodiment, the embodiment of the present invention is described in further detail below in conjunction with the accompanying drawings.It is understood that this
Specific embodiment described by place is used only for explaining the embodiment of the present invention, rather than the restriction to the embodiment of the present invention.Additionally also
It should be noted that for the ease of describing, accompanying drawing illustrate only the part relevant to the embodiment of the present invention rather than entire infrastructure.
It is the flow process signal of a kind of method analyzing customer transaction behavior that the embodiment of the present invention provides with reference to Fig. 1, Fig. 1
Figure.
As it is shown in figure 1, the method for described analysis customer transaction behavior includes:
Step 101, reads transaction data from retailer data base, when described transaction data includes Transaction Identification Number, transaction
Between, parent title belonging to trade name, sales volume and commodity；
Step 102, carries out packet aggregation according to described Transaction Identification Number to described transaction data, and described transaction data is transaction note
The complete or collected works of record, the transaction record of one commodity of every behavior of described transaction data；
Preferably, described according to described Transaction Identification Number, described transaction data is carried out packet aggregation, including:
The form of every transaction record is T=[TID, < i1,i2...,in>] form；
Wherein, T is for once to conclude the business, and TID is Transaction Identification Number, inFor certain commodity, < i1,i2...,in> it is that this transaction is purchased
The commodity set bought.
Step 103, divides in groups according to client described transaction data, merges and sets up each user to often organizing data and purchase
Buy tree；From described transaction record, extraction is without repeating exhaustively commodity, according to described commodity with hierarchical information set up business
Product tree.
Preferably, described from described transaction record extraction without repeating exhaustively commodity, according to described commodity with
Hierarchical information sets up commodity tree, including:
Being numbered parent title belonging to described trade name and described commodity, from transaction record set, extraction is without weight
Multiple exhaustively commodity set；
The sub-categories relation carried according to described commodity, top-down sets up a commodity tree, and in tree, each node contains
Having a key-value pair, key is trade name or item name, is worth for reference numeral.
Concrete, the commodity sold in market can construct commodity tree according to merchandise classification, and such as, Fig. 2 is that certain surpasses
The commodity tree construction exemplary plot in city.Wherein, " Television&Video " of commodity tree comprises two subclass item property " 4K
Ultra HD TVs " and " Smart TVs ".
Preferably, described to described transaction data according to client divide in groups, to often organize data merge set up each user
Purchase tree, including:
Described transaction data is divided in groups by different client, sets up one and buy tree often organizing commodity successively, described
The each node bought in tree contains number value.
Concrete, obtain transaction after using Spectral Clustering and buy tree, as Fig. 3 buys tree exemplary plot for transaction.First structure
Build a commodity tree, for each client, build individual commodity tree respectively, referred to as buy tree.To all of transaction data
Divide in groups according to client, merge set up the purchase tree of this user to often organizing data.User buys the similarity between tree and uses
Similarity matrix calculates, and uses spectral clustering to produce cluster result.So carrying out assignment herein by multiple parameters, see
Examine cluster result, choose reasonable cluster result, finally give ideal effect.
Step 104, each user buys the similarity between tree and uses similar matrix to represent, uses spectral clustering to La Pula
This matrix clusters, and circulates cluster process, chooses best cluster result from least one cluster result, described best
Cluster result is user grouping result.
Preferably, described each user buys the similarity between tree and uses similar matrix to represent, uses spectral clustering to drawing
This matrix of pula clusters, including:
(i j) calculates user and buys the similarity between tree i and j, added up by each column element to use similarity matrix S
Put composition N*N similarity matrix S on the diagonal；
Similarity matrix S is transformed into Laplacian Matrix L, obtains front k eigenvalue and characteristic of correspondence vector,
Arrange described eigenvalue and the matrix of described characteristic vector one N*k of composition；
By the dimensionality reduction mode dimensionality reduction of laplacian eigenmaps, the characteristic vector obtained is carried out K-means cluster.
The embodiment of the present invention provides a kind of method analyzing customer transaction behavior, by user's purchasing behavior data compression for purchasing
Buy tree data, buy tree data and be more conducive to storage and the process of the big data of user, improve poly-after data are reasonably compressed
The effect of class, is greatly improved actual application value；Meanwhile, in the clustering method of tree is bought in transaction, the method using spectral clustering,
The problem dexterously problem of a NP difficulty being converted into Laplacian Matrix eigenvalue (vectorial), by discrete clustering problem
Relaxing as continuous print characteristic vector, minimum series of features vector correspond to the serial division methods that figure is optimum, remaining is only
By the problem of laxization discretization again, will characteristic vector is subdivided opens, just can obtain corresponding classification.Pass through above procedure
The cluster result obtained, has not only evaded the dependence to user's static attribute of traditional cluster, and has been closed user data
Reason dimensionality reduction obtains good user grouping result.
Show with reference to the functional module that Fig. 4, Fig. 4 are a kind of devices analyzing customer transaction behavior that the embodiment of the present invention provides
It is intended to.
As shown in Figure 4, the device of described analysis customer transaction behavior includes:
Read module 401, for reading transaction data from retailer data base, described transaction data includes transaction
Number, parent title belonging to exchange hour, trade name, sales volume and commodity；
Grouping module 402, for carrying out packet aggregation, described transaction data according to described Transaction Identification Number to described transaction data
It is the complete or collected works of transaction record, the transaction record of one commodity of every behavior of described transaction data；
Set up module 403, for described transaction data is divided in groups according to client, set up each to often organizing data merging
The purchase tree of user；From described transaction record, extraction is without repeating exhaustively commodity, according to described commodity with level letter
Breath sets up commodity tree.
Cluster module 404, buys the similarity between tree for each user and uses similar matrix to represent, use spectral clustering
Laplacian Matrix is clustered；
Choose module 405, be used for circulating cluster process, from least one cluster result, choose best cluster result,
Described best cluster result is user grouping result.
Preferably, described grouping module 402, it is used for:
The form of every transaction record is T=[TID, < i1,i2...,in>] form；
Wherein, T is for once to conclude the business, and TID is Transaction Identification Number, inFor certain commodity, < i1,i2...,in> it is that this transaction is purchased
The commodity set bought.
Preferably, described set up module 403, be used for:
Being numbered parent title belonging to described trade name and described commodity, from transaction record set, extraction is without weight
Multiple exhaustively commodity set；
The sub-categories relation carried according to described commodity, top-down sets up a commodity tree, and in tree, each node contains
Having a key-value pair, key is trade name or item name, is worth for reference numeral.
Preferably, described set up module 403, be additionally operable to:
Described transaction data is divided in groups by different client, sets up one and buy tree often organizing commodity successively, described
The each node bought in tree contains number value.
Preferably, described cluster module 404, it is used for:
(i j) calculates user and buys the similarity between tree i and j, added up by each column element to use similarity matrix S
Put composition N*N similarity matrix S on the diagonal；
Similarity matrix S is transformed into Laplacian Matrix L, obtains front k eigenvalue and characteristic of correspondence vector,
Arrange described eigenvalue and the matrix of described characteristic vector one N*k of composition；
By the dimensionality reduction mode dimensionality reduction of laplacian eigenmaps, the characteristic vector obtained is carried out K-means cluster.
The embodiment of the present invention provides a kind of device analyzing customer transaction behavior, by user's purchasing behavior data compression for purchasing
Buy tree data, buy tree data and be more conducive to storage and the process of the big data of user, improve poly-after data are reasonably compressed
The effect of class, is greatly improved actual application value；Meanwhile, in the clustering method of tree is bought in transaction, the method using spectral clustering,
The problem dexterously problem of a NP difficulty being converted into Laplacian Matrix eigenvalue (vectorial), by discrete clustering problem
Relaxing as continuous print characteristic vector, minimum series of features vector correspond to the serial division methods that figure is optimum, remaining is only
By the problem of laxization discretization again, will characteristic vector is subdivided opens, just can obtain corresponding classification.Pass through above procedure
The cluster result obtained, has not only evaded the dependence to user's static attribute of traditional cluster, and has been closed user data
Reason dimensionality reduction obtains good user grouping result.
The know-why of the embodiment of the present invention is described above in association with specific embodiment.These describe and are intended merely to explain this
The principle of inventive embodiments, and the restriction to embodiment of the present invention protection domain can not be construed to by any way.Based on herein
Explanation, those skilled in the art need not to pay performing creative labour, and can to associate other of the embodiment of the present invention concrete
Embodiment, within these modes fall within the protection domain of the embodiment of the present invention.
Claims (10)
1. the method analyzing customer transaction behavior, it is characterised in that described method includes:
From retailer data base read transaction data, described transaction data include Transaction Identification Number, exchange hour, trade name,
Parent title belonging to sales volume and commodity；
According to described Transaction Identification Number, described transaction data being carried out packet aggregation, described transaction data is the complete or collected works of transaction record, institute
State the transaction record of one commodity of every behavior of transaction data；
Described transaction data is divided in groups according to client, merges set up the purchase tree of each user to often organizing data；From described
In transaction record, extraction is without repeating exhaustively commodity, according to described commodity with hierarchical information set up commodity tree；
Each user buys the similarity between tree and uses similar matrix to represent, uses spectral clustering to gather Laplacian Matrix
Class, circulates cluster process, chooses best cluster result from least one cluster result, and described best cluster result is for using
Family group result.
Method the most according to claim 1, it is characterised in that described according to described Transaction Identification Number, described transaction data is carried out
Packet aggregation, including:
The form of every transaction record is T=[TID, < i1,i2...,in>] form；
Wherein, T is for once to conclude the business, and TID is Transaction Identification Number, inFor certain commodity, < i1,i2...,in> it is the business bought in this transaction
Product set.
Method the most according to claim 1, it is characterised in that described extraction nothing repetition exhaustive from described transaction record
Commodity, according to described commodity with hierarchical information set up commodity tree, including:
Being numbered parent title belonging to described trade name and described commodity, from transaction record set, extraction is without repeating nothing
The commodity set omitted；
The sub-categories relation carried according to described commodity, top-down sets up a commodity tree, and in tree, each node contains one
Individual key-value pair, key is trade name or item name, is worth for reference numeral.
Method the most according to claim 1, it is characterised in that described to described transaction data according to client divide in groups,
Merge set up the purchase tree of each user to often organizing data, including:
Described transaction data is divided in groups by different client, sets up one and buy tree, described purchase often organizing commodity successively
Each node in tree contains number value.
Method the most according to claim 1, it is characterised in that described each user buys the similarity between tree and uses phase
Represent like matrix, use spectral clustering that Laplacian Matrix is clustered, including:
(i j) calculates user and buys the similarity between tree i and j, added up by each column element and be placed on to use similarity matrix S
N*N similarity matrix S is formed on diagonal；
Similarity matrix S is transformed into Laplacian Matrix L, obtains front k eigenvalue and characteristic of correspondence vector, arrangement
Described eigenvalue and the matrix of described characteristic vector one N*k of composition；
By the dimensionality reduction mode dimensionality reduction of laplacian eigenmaps, the characteristic vector obtained is carried out K-means cluster.
6. the device analyzing customer transaction behavior, it is characterised in that described device includes:
Read module, for reading transaction data, when described transaction data includes Transaction Identification Number, transaction from retailer data base
Between, parent title belonging to trade name, sales volume and commodity；
Grouping module, for described transaction data being carried out packet aggregation according to described Transaction Identification Number, described transaction data is transaction
The complete or collected works of record, the transaction record of one commodity of every behavior of described transaction data；
Set up module, for described transaction data is divided in groups according to client, set up each user's to often organizing data merging
Buy tree；From described transaction record, extraction is without repeating exhaustively commodity, according to described commodity with hierarchical information set up
Commodity tree；
Cluster module, buys the similarity between tree for each user and uses similar matrix to represent, use spectral clustering general to drawing
Lars matrix clusters；
Choose module, be used for circulating cluster process, from least one cluster result, choose best cluster result, described best
Cluster result be user grouping result.
Device the most according to claim 6, it is characterised in that described grouping module, is used for:
The form of every transaction record is T=[TID, < i1,i2...,in>] form；
Wherein, T is for once to conclude the business, and TID is Transaction Identification Number, inFor certain commodity, < i1,i2...,in> it is the business bought in this transaction
Product set.
Device the most according to claim 6, it is characterised in that described set up module, is used for:
Being numbered parent title belonging to described trade name and described commodity, from transaction record set, extraction is without repeating nothing
The commodity set omitted；
The sub-categories relation carried according to described commodity, top-down sets up a commodity tree, and in tree, each node contains one
Individual key-value pair, key is trade name or item name, is worth for reference numeral.
Device the most according to claim 6, it is characterised in that described set up module, is additionally operable to:
Described transaction data is divided in groups by different client, sets up one and buy tree, described purchase often organizing commodity successively
Each node in tree contains number value.
Device the most according to claim 6, it is characterised in that described cluster module, is used for:
(i j) calculates user and buys the similarity between tree i and j, added up by each column element and be placed on to use similarity matrix S
N*N similarity matrix S is formed on diagonal；
Similarity matrix S is transformed into Laplacian Matrix L, obtains front k eigenvalue and characteristic of correspondence vector, arrangement
Described eigenvalue and the matrix of described characteristic vector one N*k of composition；
By the dimensionality reduction mode dimensionality reduction of laplacian eigenmaps, the characteristic vector obtained is carried out K-means cluster.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610460883.3A CN106127493A (en) | 2016-06-23 | 2016-06-23 | A kind of method and device analyzing customer transaction behavior |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610460883.3A CN106127493A (en) | 2016-06-23 | 2016-06-23 | A kind of method and device analyzing customer transaction behavior |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106127493A true CN106127493A (en) | 2016-11-16 |
Family
ID=57268998
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610460883.3A Pending CN106127493A (en) | 2016-06-23 | 2016-06-23 | A kind of method and device analyzing customer transaction behavior |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106127493A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106779074A (en) * | 2017-01-22 | 2017-05-31 | 腾云天宇科技（北京）有限公司 | Market brand combination prediction method and prediction server |
| WO2018059015A1 (en) * | 2016-09-29 | 2018-04-05 | 深圳大学 | Transaction data-based customer classification method, and system thereof |
| CN108268898A (en) * | 2018-01-19 | 2018-07-10 | 大象慧云信息技术有限公司 | A kind of electronic invoice user clustering method based on K-Means |
| CN109034853A (en) * | 2017-06-09 | 2018-12-18 | 北京京东尚科信息技术有限公司 | Similar users method, apparatus, medium and electronic equipment are found based on seed user |
| CN114418652A (en) * | 2022-01-27 | 2022-04-29 | 中国农业银行股份有限公司 | Customer group determination method and related equipment |
| CN114444577A (en) * | 2021-12-31 | 2022-05-06 | 广州盖盟达工业品有限公司 | Automatic product classification method and device, computer equipment and storage medium |
| CN116342168A (en) * | 2023-05-23 | 2023-06-27 | 山东灵动电子商务有限公司 | Information big data intelligent acquisition management system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101996215A (en) * | 2009-08-27 | 2011-03-30 | 阿里巴巴集团控股有限公司 | Information matching method and system applied to e-commerce website |
| CN103412948B (en) * | 2013-08-27 | 2017-10-24 | 北京交通大学 | The Method of Commodity Recommendation and system of collaborative filtering based on cluster |
-
2016
- 2016-06-23 CN CN201610460883.3A patent/CN106127493A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101996215A (en) * | 2009-08-27 | 2011-03-30 | 阿里巴巴集团控股有限公司 | Information matching method and system applied to e-commerce website |
| CN103412948B (en) * | 2013-08-27 | 2017-10-24 | 北京交通大学 | The Method of Commodity Recommendation and system of collaborative filtering based on cluster |
Non-Patent Citations (2)
| Title |
|---|
| 徐艳: "基于卷积核的营口港客户细分研究", 《中国优秀硕士学位论文全文数据库-经济与管理科学辑》 * |
| 褚维伟等: "一种带约束条件的购物篮分析方法", 《HTTP://KNS.CNKI.NET/KXREADER/DETAIL?TIMESTAMP=636809273834218750&DBCOD》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018059015A1 (en) * | 2016-09-29 | 2018-04-05 | 深圳大学 | Transaction data-based customer classification method, and system thereof |
| CN106779074A (en) * | 2017-01-22 | 2017-05-31 | 腾云天宇科技（北京）有限公司 | Market brand combination prediction method and prediction server |
| CN109034853A (en) * | 2017-06-09 | 2018-12-18 | 北京京东尚科信息技术有限公司 | Similar users method, apparatus, medium and electronic equipment are found based on seed user |
| CN108268898A (en) * | 2018-01-19 | 2018-07-10 | 大象慧云信息技术有限公司 | A kind of electronic invoice user clustering method based on K-Means |
| CN114444577A (en) * | 2021-12-31 | 2022-05-06 | 广州盖盟达工业品有限公司 | Automatic product classification method and device, computer equipment and storage medium |
| CN114418652A (en) * | 2022-01-27 | 2022-04-29 | 中国农业银行股份有限公司 | Customer group determination method and related equipment |
| CN116342168A (en) * | 2023-05-23 | 2023-06-27 | 山东灵动电子商务有限公司 | Information big data intelligent acquisition management system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106127493A (en) | A kind of method and device analyzing customer transaction behavior | |
| CN106529968B (en) | Customer classification method and system based on transaction data | |
| TWI512653B (en) | Information providing method and apparatus, method and apparatus for determining the degree of comprehensive relevance | |
| Annie et al. | Market basket analysis for a supermarket based on frequent itemset mining | |
| JP5965911B2 (en) | Data processing based on online trading platform | |
| Ahmed et al. | E-banking customer satisfaction and loyalty: Evidence from serial mediation through modified ES-QUAL model and second-order PLS-SEM | |
| Verma et al. | An intelligent approach to Big Data analytics for sustainable retail environment using Apriori-MapReduce framework | |
| CN101454771A (en) | System and method of segmenting and tagging entities based on profile matching using a multi-media survey | |
| CN103136683A (en) | Method and device for calculating product reference price and method and system for searching products | |
| CN102609854A (en) | Client partitioning method and device based on unified similarity calculation | |
| Mesforoush et al. | Customer profitability segmentation for SMEs case study: network equipment company | |
| Maulina et al. | Data mining approach for customer segmentation in B2B settings using centroid-based clustering | |
| CN109977299A (en) | A kind of proposed algorithm of convergence project temperature and expert's coefficient | |
| Zhang et al. | Measuring customer similarity and identifying cross-selling products by community detection | |
| Lewaaelhamd | Customer segmentation using machine learning model: an application of RFM analysis | |
| Mostafa | Knowledge discovery of hidden consumer purchase behaviour: a market basket analysis | |
| Yang et al. | Discovery of online shopping patterns across websites | |
| Rezaeian et al. | Measuring Customers Satisfaction of ECommerce Sites Using Clustering Techniques: Case Study of Nyazco Website. | |
| Triandini et al. | Hierarchical Clustering for Functionalities E-Commerce Adoption | |
| CN110020918B (en) | Recommendation information generation method and system | |
| Gholamian et al. | Improving electronic customers' profile in recommender systems using data mining techniques | |
| Cho et al. | Clustering method using weighted preference based on RFM score for personalized recommendation system in u-commerce | |
| Guo et al. | EC‐Structure: Establishing Consumption Structure through Mining E‐Commerce Data to Discover Consumption Upgrade | |
| Faridizadeh et al. | Market basket analysis using community detection approach: A real case | |
| Granov | Customer loyalty, return and churn prediction through machine learning methods: for a Swedish fashion and e-commerce company |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Peng Si Inventor after: Chen Xiaojun Inventor after: Huang Zhexue Inventor after: Ming Yong Inventor before: Chen Xiaojun Inventor before: Peng Si Inventor before: Huang Zhexue Inventor before: Ming Yong |
|
| COR | Change of bibliographic data | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161116 |