CN106600356A - Multi-platform electronic commerce information aggregation method and system - Google Patents
Multi-platform electronic commerce information aggregation method and system Download PDFInfo
- Publication number
- CN106600356A CN106600356A CN201610957162.3A CN201610957162A CN106600356A CN 106600356 A CN106600356 A CN 106600356A CN 201610957162 A CN201610957162 A CN 201610957162A CN 106600356 A CN106600356 A CN 106600356A
- Authority
- CN
- China
- Prior art keywords
- platform
- electric business
- information
- commodity
- business platform
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0641—Shopping interfaces
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a multi-platform electronic commerce information aggregation method and system. The method includes the following steps that: the page category information about sample commodities in electronic commerce platforms is acquired; similarity analysis is performed, the minimum information elements of the sample commodities are analyzed based on an analysis result, the minimum information elements are merged, so that the minimum information element set of each sample commodity in the electronic commerce platforms is obtained, and the minimum information element sets of the sample commodities constitute an integrated sample set; with the minimum information element set and category information obtained from one electronic commerce platform adopted as a training sample pair, each electronic commerce platform is trained with a neural network of which the weight coefficient is 0 or 1; and the minimum information element sets of commodities to be aggregated are inputted according to the neural network aggregation model of the different electronic commerce platforms, and aggregated category information satisfying the commodity page requirements of corresponding electronic commerce platforms is obtained through model calculation. With the multi-platform electronic commerce information aggregation method and system adopted, defects such as large maintenance workload, low efficiency, error-proneness and incapability of timely discovering change which are caused by the disunity of platform interfaces can be eliminated.
Description
Technical field
The present invention relates to electronic commerce information technology, more particularly to one kind can be used for multi-platform cross-border ecommerce and live
Dynamic commodity are issued and the Information Syndication safeguarded.
Background technology
Now, with world economic integration and the development of globalization, economic development and society of the ecommerce to various countries
The impact of transition gradually strengthens.There are huge market basis and economic vitality in China in occupation of population most in the world.It is early
In 2010, China just became global manufacturing output highest country.The ecommerce in the range of our times is rapidly sent out
Under the background of exhibition, Small and Medium Enterprises in China combines e-commerce technology, especially cross-border ecommerce, borrows emerging trade modes and comes
Promote Chinese Enterprises to enter international market, become that increasingly there is realistic feasibility.Small and Medium Enterprises in China utilizes cross-border electronics business
Business enters international market, for the international market share for expanding China's product, expands foreign trade marketing network, breeds new outlet competing
Advantage is striven, changes the mode of doing business of traditional international trade, promote the industrial transformation and industrial upgrading of China's foreign trade, improved outside China
Competitiveness of the trade in international market has important and far-reaching meaning.
So-called cross-border ecommerce, refers to the transaction agent for adhering to country variant or area separately, complete by e-commerce platform
Into payment and settlement of concluding the business, carry out, and by cross-border logistics distribution commodity, complete a kind of international business's trade activity concluded the business.Its
Feature has:Global, instantaneity and convenience.The cross-border electric business transaction size of China in 2013 is 3.1 trillion yuans, and rate of increase is
31.3%, account for the 11.9% of total import and export volume, although accounting is relatively low, but with the globalization of Electronic Commerce in China, across
Status meeting more and more higher of the border ecommerce in China's Foreign Trade, it is contemplated that cross-border electric business was in total import and export volume in 2017
In ratio be up to 20% or so.
The Chinese government actively promotes cross-border e-commerce development, at present Shanghai, Chongqing, Hangzhou, Ningbo, Zhengzhou, depth
7 cities such as Zhen Qianhai and Qingdao or area achieve cross-border ecommerce experimental city in succession.Meanwhile, a series of cross-border electric business
The intensive appearance of policy also makes policy environment, law, regulation, standards system and the support safeguard level of the cross-border ecommerce of China
Etc. having obtained significantly being lifted, the impetus of cross-border e-commerce development is powerful.
Domestic medium-sized and small enterprises pass through into the high cost that one of the significant obstacle of international market is acquisition of information, ecommerce
Optimization to supply chains process, reduces the cost of acquisition of information, and improves the integration running of enterprise supply chain management
Level, and then management cost, acquisition of information cost be greatly reduced.Enterprise set foot in cross-border ecommerce firstly the need of to company from
The electronic information level of body is built and is safeguarded, is engaged in the medium-sized and small enterprises of cross-border ecommerce, only 250,000 or so
The enterprise of 7000 or so creates the electronic information platform of oneself, and other numerous medium-sized and small enterprises are still needed by the 3rd
Fang Pingtai carries out foreign trade.However, overseas market is huge, the cross-border e-platform of third party is numerous, the issue of its merchandise news
Specification and requirement are different, and the workload and degree of specialization of maintenance of information is higher, at present, the maintenance of most of different platforms
Professional is needed to have been manually done, for the little Wei enterprises of resource-constrained, the maintenance of multi-platform e-commerce system becomes
White elephant.The structure of cross-border managing eBusiness platform needs to integrate the information resources of a large amount of Dispersed heterogeneous, these moneys
Source is likely distributed in different third-party platform systems, and this is that one numerous and diverse and difficult task.How not fellow disciple is effectively reused
The information resources at family, build the information management maintenance system of commodity automatically, and the commodity of different platform are carried out issuing automatically and
Maintenance work, so as to increase substantially cross-border ecommerce commodity information management efficiency, reduces commodity cross-border in multiple third parties
The issue of e-commerce platform and maintenance cost oneself become a key technical problem urgently to be resolved hurrily.
At present, cross-border e-commerce platform is numerous, and the platform such as speed of main flow sells logical, Amazon, eBay, Wish, The Orchid Pavilion
Collection gesture, Dunhuang, Walmart, Newegg etc., the technical standard of each platform employing, background data base are different, cross-border for being engaged in
Enterprise for, need respectively in different platform according to the platform requirement publishing commodity information and description, and to oneself
Commodity enter Mobile state management, including carrying out timely synchronized update to the information of same commodity different platform；A certain platform is become
Descriptive labelling project more remaps etc.；The demand that to enter edlin to the attribute of new product different to adapt to each platform
And push to the platform and complete commodity and get in stocks.In the past, the issue of commodity, maintenance and renewal needs artificial treatment, and workload is big, effect
Rate is low, easily error, therefore, some platforms have issued data-interface, and cross-border enterprise can carry out batch by data-interface
The issue of product and renewal, however, the interface of different platform is with data form and differs, for while carrying out in multiple platforms
The cross-border enterprise of operation, it is still desirable to carry out substantial amounts of artificial operation, the computer that realize the process is automatically processed, specifically needs
Solve problems with：
1) destructuring of data.In most of data message of cross-border e-commerce platform now, it is flooded with big
The non-structured data of amount, the ratio shared by structural data is still very low, and this is a disaster for the access of data base
Topic, these unstructured datas are only converted into the structural data of unified data form.
2) mapping between commodity minimal information element with combine, the data category difference between each platform, each platform
For the description classification of commodity is differed, some attributes (such as trade name) are the predicables of each platform, are had
Then that respective platform is exclusive, although and other be each platform have, description require it is different with information fusion degree,
Therefore some differences are had, is more also the complex properties that simple attributes are combined.And cross-border enterprise needs to carry out commodity
Unified maintenance, it is therefore desirable to set up a big commodity data classification comprising all platform properties, commodity data carried out
Unified tissue and planning, need that original attribute synthesis set is screened and merged.
3) Mass production platform descriptive labelling respectively, each platform are carried out to the set of commodity classification according to commodity place platform
Descriptive labelling classification and form it is different, need to set up different platform information and close with the dynamic mapping of description information set
System, so that according to attribute synthesis set and target platform, dynamic generates the item property list of target platform and content, so as to right
The merchandise news of different platform carries out unifying to update and safeguarding.
In cross-border e-commerce initiative, a large amount of medium-sized and small enterprises for carrying out foreign trade by third-party platform are there are, should
Class medium-sized and small enterprises lack qualified technical personnel, and human resourcess are deficient, and the problems referred to above become the medium and small cross-border enterprise of impact and expand product rapidly
One of selling market is badly in need of the technical barrier for solving.
The content of the invention
The technical problem to be solved is to provide a kind of multi-platform electronic commerce information polymerization and system, gram
Take multiple cross-border e-commerce platform commodity in prior art issue, information updating when because tieing up caused by each platform interface disunity
Shield workload is big, inefficiency, easily error, can not in time find that plateform system changes and cannot carry out information updating etc. and lack
Fall into.
To solve the above problems, the present invention proposes a kind of multi-platform electronic commerce information polymerization, comprises the following steps：
Obtain page category information of each electric business platform with regard to sample commodity；
Whole category informations of the same sample commodity on the different electric business platform pages to obtaining carry out similarity analysis,
The minimal information element of sample commodity is parsed based on similarity analysis result, is merged minimal information element and is obtained the sample commodity
Minimal information element set, the minimal information element set of multiple sample commodity constitutes comprehensive sample set；
For each electric business platform, with the minimal information element set of sample commodity and the classification letter obtained from an electric business platform
Breath carries out the neural metwork training that weight coefficient value is 0 or 1 as training sample pair, obtains the nerve net of each electric business platform
Network polymerization model；
According to the neutral net polymerization model of different electric business platforms, the minimal information element set of commodity to be polymerized is input into
Close, Jing models are calculated the category information of correspondence electric business platform commodity page request.
According to one embodiment of present invention, also include：
When merchandise news is changed, the neutral net polymerization is re-entered into according to the minimal information element set for updating
Model, obtains the category information that each electric business platform updates, and is uploaded to corresponding electric business platform.
According to one embodiment of present invention, also include：
When the category information of the electric business platform sample commodity page for being monitored changes, similarity analysis and solution are carried out
Analysis merging obtains new minimal information element set, builds the new sample pair of target electric business platform, the new nerve of re -training
Network polymerization model, original minimal information element set of commodity to be polymerized is input to the neutral net polymerization mould after updating
Category information after being changed in type, and it is uploaded to the target electric business platform.
According to one embodiment of present invention, by capturing each electric business based on target data model orientation using focused crawler
The dependent merchandise web page resources of platform, to obtain category information of each electric business platform with regard to commodity.
According to one embodiment of present invention, each neutral net polymerization model is minimal information element set to an electric business
The mapping model of the category information of platform, including one layer of input layer, multilamellar or monolayer hidden layer, one layer of output layer, input number of nodes
Determined according to the element number of minimum unit information aggregate, hidden layers numbers and nodes determine according to the granularity of information fusion, defeated
The classification quantity of the corresponding electric business platform of the several evidences of egress determines.
According to one embodiment of present invention, the ginseng using neural network feedback correcting algorithm to neutral net polymerization model
Number is trained and optimizing.
According to one embodiment of present invention, described being trained includes with optimizing：
Using binary GA algorithms, the weight parameter value of neutral net polymerization model is 0 or 1, hidden node and output
Node is weighted summation to input information, obtains the synthesis of different information elements；
The output valve of comparative neural network polymerization model and the output valve of sample pair, enter according to comparative result to weight parameter
Row change, to obtain being adapted to the neutral net polymerization model of each electric business platform.
According to one embodiment of present invention, neural metwork training includes off-line training pattern, comprises the following steps：
A1：The chromosome set of binary GA algorithms is generated at random, selects an initial chromosome as the first of neutral net
Beginning weight coefficient value, and fitness value is calculated, initial chromosome is assigned to into history optimal value；
A2：Chromosome set is replicated, made a variation and is intersected, the chromosome population after being replicated, made a variation and being intersected is closed
Collection, calculates each fitness value using chromosome therein as the weight coefficient value of neutral net respectively；
A3：The chromosome of minimum fitness in each fitness value is selected, if the minimum fitness is optimum less than history
Value, then be assigned to history optimal value by corresponding chromosome, otherwise keeps history optimal value constant, the new dye produced to variation
Colour solid set is screened based on fitness function using roulette mode, obtains follow-on chromosome set, waits iteration；
A4：When judging that the chromosome of history optimal value is applied in neutral net polymerization model, the model of all samples pair
Whether output is all consistent with sample output, if then terminating iterative learning, exports the optimum chromosome, otherwise selects of future generation
Chromosome set return to step A2.
According to one embodiment of present invention, neural metwork training is additionally included in line operational mode, comprises the following steps：
B1：According to the weight coefficient and structure of the corresponding neutral net polymerization model of target electric business platform selecting, phase is generated
The neutral net polymerization model answered；
B2：Input minimal information element set, calls neutral net polymerization model to calculate the commodity page of target electric business platform
The category information in face.
The present invention also provides a kind of multi-platform electronic commerce information paradigmatic system, including：
Platform information acquisition module, to obtain page category information of each electric business platform with regard to sample commodity；
Information similarity analysis integrate module, to the same sample commodity to obtaining on the different cross-border electric business platform pages
Whole category informations carry out similarity analysis, the minimal information element of sample commodity is parsed based on similarity analysis result,
Merge the minimal information element set that minimal information element obtains the sample commodity, the minimal information element set of multiple sample commodity
Constitute comprehensive sample set；
Polymerization model training module, to for each electric business platform, with the minimal information element set of sample commodity with from
The category information that one electric business platform is obtained carries out the neural metwork training that weight coefficient value is 0 or 1 as training sample pair,
Obtain the neutral net polymerization model of each electric business platform；
Aggregation information computing module, to the neutral net polymerization model according to different electric business platforms, is input into and waits to gather
The minimal information element set of commodity is closed, Jing models are calculated the category information of correspondence electric business platform commodity page request.
According to one embodiment of present invention, also include：
Polymerization model dynamic update module, to when merchandise news is changed, according to the minimal information element set for updating
The neutral net polymerization model is re-entered into, the category information that each electric business platform updates is obtained, and is uploaded to corresponding electricity
Business's platform；And/or, when the category information of the electric business platform sample commodity page for being monitored changes, to carry out similar
Degree analysis and parsing merging obtain new minimal information element set, build the new sample pair of target electric business platform, instruct again
The new neutral net polymerization model of white silk, by original minimal information element set of commodity to be polymerized the nerve after updating is input to
Category information after being changed in network polymerization model, and it is uploaded to the target electric business platform.
After above-mentioned technical proposal, the present invention has the advantages that compared to existing technology：
Neutral net polymerization model is set up by 0-1 connection weight coefficients, merchandise news can be effectively carried out substantially first
The polymerization of element, quickly obtains meeting the category information of each target electric business platform page, can be used for towards multi-platform cross-border electronics business
The information fusion of business；
Using continuous parameter model and sigmoid functions, the present invention is in order to more preferably reach information for traditional neural network model
The purpose of polymerization, specially proposes a kind of neural network model of redesign, and its parameter adopts 0-1 integers, transfer function to adopt
With addition function, using 2 system GA algorithms, these have no in tradition or at present technology and address for training；
Because the introducing of integer 0-1 neutral net polymerization models result in model parameter discontinuous problem, and binary system GA
Chromosome and neutral net polymerization model weight mapping method and training method, efficiently solve neutral net polymerization model
Study and training problem on sample machine；
The dynamic detection and update mechanism of system, is become to merchandise news by dynamic with cross-border electric business platform classes of pages purpose
Change situation is detected, once detecting change, page classification content is carried out at once and is recalculated or neutral net polymerization model
Re -training, update asynchronous problem so as to solve time lag that Traditional Man generation patterns bring and content of pages.
Description of the drawings
Fig. 1 is the schematic flow sheet of the multi-platform electronic commerce information polymerization of one embodiment of the invention；
Fig. 2 is the structured flowchart of the neutral net polymerization model of one embodiment of the invention；
Fig. 3 is the operation principle schematic diagram of the multi-platform electronic commerce information polymerization of one embodiment of the invention；
Fig. 4 is the schematic flow sheet of the neural metwork training of the off-line training pattern of one embodiment of the invention；
The schematic flow sheet that Fig. 5 works for the neutral net of the on-line operation pattern of one embodiment of the invention；
Fig. 6 updates workflow schematic diagram for the online dynamic of neutral net polymerization model of one embodiment of the invention；
Fig. 7 is the neutral net polymerization model weight coefficient pass corresponding with binary system GA chromosomes of one embodiment of the invention
System.
Specific embodiment
It is understandable to enable the above objects, features and advantages of the present invention to become apparent from, below in conjunction with the accompanying drawings to the present invention
Specific embodiment be described in detail.
Elaborate many details in order to fully understand the present invention in the following description.But the present invention can be with
Much it is different from alternate manner described here to implement, those skilled in the art can be in the situation without prejudice to intension of the present invention
Under do similar popularization, therefore the present invention is not embodied as being limited by following public.
Referring to Fig. 1, the multi-platform electronic commerce information polymerization of the embodiment of the present invention, comprise the following steps：
S1：Obtain page category information of each electric business platform with regard to sample commodity；
S2：Whole category informations to obtaining carry out similarity analysis, and based on similarity analysis result sample business is parsed
The minimal information element of product, merges the minimal information element set that minimal information element obtains the sample commodity, multiple sample business
The minimal information element set of product constitutes comprehensive sample set；
S3：For each electric business platform, with the minimal information element set of sample commodity and the class obtained from an electric business platform
Mesh information carries out the neural metwork training that weight coefficient value is 0 or 1 as training sample pair, obtains the god of each electric business platform
Jing network polymerization models；
S4：According to the neutral net polymerization model of different electric business platforms, the minimal information unit of commodity to be polymerized is input into
Element set, Jing models are calculated the category information of correspondence electric business platform commodity page request.
Each electric business platform is classified to the item property of itself commodity page and description data can have more or less difference,
For example, the volume of the commodity of electric business platform definition, can then be defined as length, width and height in another electric business platform.In step S1
In, them can be obtained by capturing the commodity page of each electric business platform respectively for the different attribute classification of commodity and is described
Data, naturally it is also possible to obtained using other modes.In other words, first have to obtain classification letter of each electric business platform with regard to commodity
Breath.Category information be electric business platform for the attribute of commodity carries out sorted information, can include corresponding classification classification and
Classification content.
Optionally, in step sl, by capturing each electric business platform based on target data model orientation using focused crawler
Dependent merchandise web page resources, to obtain category information of each electric business platform with regard to commodity.Specifically, by controller, solution
Parser, resources bank three is partially completed.The groundwork of controller is responsible for being shared out the work to each reptile thread in multithreading
Task.Controller is the central controller of web crawlers, it be mainly responsible for according to system be transmitted through come URL link, distribution one
Thread, then starts the process that thread dispatching crawls webpage.The groundwork of resolver is to download webpage, carries out the place of the page
Reason, mainly falls the contents processings such as some JS script tags, CSS code contents, space character, html tag, the base of reptile
This work is completed by resolver, and job content has：The function of webpage is downloaded, the text of webpage is processed, such as filter work(
Can, extract the function of special html tag, analytical data function etc..Resources bank is for depositing the web page resources for downloading to one
As all adopt large-scale database purchase, such as oracle database, and index is set up to it, for storing webpage in download
Data record container, and provide generate index target source.Medium-and-large-sized database product has：Oracle、Sql
Server etc..
In step s 2, the whole category informations to obtaining in step sl carry out similarity analysis, based on similarity point
Analysis result parses each minimal information element, can be specifically that similar element is retained into one, dissimilar element whole guarantor
Stay, the element to being combined is disassembled, such as length, width and height element is decomposed into three elements of length, so as to obtain business
The whole minimal information element of product, merges these minimal information elements and obtains minimal information element set, as merchandise news
Comprehensive sample sets, for study, training and operation.
In step s3, for each electric business platform, each electric business platform trains respective neutral net polymerization model, often
The secondary minimal information element set obtained using in step S2, as sample pair, is carried out with the category information obtained from an electric business platform
Weight coefficient value is 0 or 1 neural metwork training, obtains the neutral net polymerization model of an electric business platform, to each electricity
Business's platform is trained so as to obtain each respective neural network model of electric business platform.
Each neutral net polymerization model is mapping mould of the minimal information element set to the category information of an electric business platform
Type, including one layer of input layer, multilamellar or monolayer hidden layer, one layer of output layer, input number of nodes is according to minimum unit information aggregate
Element number determines that hidden layers numbers and the number of hidden nodes determine according to the granularity of information fusion, for realizing between different information
Synthesis, the classification quantity (can be understood as the data bulk in category information) of the corresponding electric business platform of the several evidences of output node is really
It is fixed.Input number of nodes, the output node number of model can enter Mobile state adjustment, the i.e. structure and parameter of model according to practical situation
All it is dynamic change, to adapt to the maintenance upgrade requirement of commodity self information and the platform page.
In one embodiment, the parameter of neutral net polymerization model is instructed using neural network feedback correcting algorithm
Practice and optimizing.Training algorithm preferably using the GA algorithms of 2 systems, solves the discontinuous parameter neural network model that adopted
A feedback learning difficult problem.
Specifically, being trained can include with optimizing：Using binary GA algorithms, the power of neutral net polymerization model
Weight parameter value is 0 or 1, and hidden node and output node are weighted summation to input information, obtain the comprehensive of different information elements
Close；The output valve of comparative neural network polymerization model and the output valve of sample pair, are carried out more according to comparative result to weight parameter
Change, to obtain being adapted to the neutral net polymerization model of each electric business platform.
In step s 4, after training is completed, model running just can be carried out, obtains the category information of each electric business platform, root
According to the neutral net polymerization model of different electric business platforms, the minimal information element set of commodity to be polymerized is input into, Jing models are calculated
Obtain the category information of correspondence electric business platform commodity page request.Eliminate the artificial operation of prior art to set up and maintenance class
The form of mesh information, the speed of service is compared and is greatly improved for artificial operation, and accuracy rate is greatly improved.Business to be polymerized
The minimal information element set of product can be first according to the minimal information of sample commodity by the item property provided commodity producer
The form of element set carries out division and sums up acquisition, can be continuing with later after acquisition, sums up without the need for carrying out division again.
Because the category information of the commodity page of the information and electric business platform of commodity can change, accordingly, it would be desirable to dynamic
Go to state to carry out the renewal of commodity page category information and the renewal of polymerization model, and by the electric business platform commodity page after renewal
The classification and content uploading of face requirement is to electric business platform.
Preferably, multi-platform electronic commerce information polymerization can also include：When merchandise news is changed, that is, most
Element change in little information element set, according to the minimal information element set for updating neutral net polymerization mould is re-entered into
Type, obtains the category information that each electric business platform updates, and is uploaded to corresponding electric business platform.
Preferably, multi-platform electronic commerce information polymerization can also include：As the electric business platform sample business for being monitored
(the paradigmatic relation i.e. between the minimal information element set and page category information of commodity when the category information of the product page changes
When changing), carry out similarity analysis and parsing merging obtains new minimal information element set, build target electric business platform
New sample pair, the new neutral net polymerization model of re -training, by original minimal information element set of commodity to be polymerized
Conjunction is input to the category information after being changed in the neutral net polymerization model after updating, and it is flat to be uploaded to the target electric business
Platform.
The method of the embodiment of the present invention is expanded on further below.
As shown in Fig. 2 the input of the neutral net polymerization model is the minimal information element set of merchandise news, it is output as
The category information of cross-border electric business platform commodity page request, the input node number of neutral net is merchandise news minimal information unit
Number j of element, output node number for category information bar number i, depending on hidden node number and the number of plies are by the degree of information fusion,
In the present embodiment, it is k to choose hidden node number, and the hidden node number of plies selects one layer, arrives hidden from input node p (p ∈ [1,2 ... j])
The weight coefficient of node q (q ∈ [1,2 ... k]) is ωpq, the weight coefficient of (m ∈ [1,2 ... n]) from hidden node q to output node m
For νqm, wherein ωpqAnd νqmValue be 0 or 1, i.e. ωpq∈ [0,1], νqm∈ [0,1], cross-border electronic commerce information nerve
The weight coefficient of network polymerization model value among 0 and 1, therefore can be described as integer 0-1 neutral nets.The hidden section of neutral net
Point selects linear adder function, i.e. hidden node output y with the transmission function of outputq=Σ xp*ωpq,yqIt is defeated for q-th hidden node
Go out, xpIt is input into for p-th node, output node is similar with hidden node, and transmission function is also adopted by linear adder function, using integer
0-1 neutral nets are able to express and are combined by different merchandise news least members from the purpose of linear adder function,
The category information of cross-border electric business platform commodity page request is described.During the neural metwork training, calculated based on binary system GA
Method, according to the sample pair that merchandise news least member is constituted with the categories of information of cross-border electric business platform commodity page request, to god
Connection weight parameter ω of Jing networkspqAnd νqmIt is trained to obtain that different ecommerce page infos and commodity can be reflected
The model of combination and mapping relations between information least member.
Neural metwork training may include off-line training pattern.The input of off-line training pattern is through similarity analysis and whole
, to set, sample is to including commodity for the sample of the cross-border electric business platform commodity page info least member information aggregate composition after conjunction
Category information Y of minimal information element set U and the page, such as " knapsack " and " man " are two least members, then the platform of commodity
Title classification corresponding content " man's knapsack " is the combinatorial mapping result of two least members；In " size " classification, knapsack
The long 480mm of three minimal information elements, high 350mm, depth 150mm collectively form specification category information in the platform page
" 350mm*480mm*150mm ", different platform has different mapping relations, has been manually done minimum letter by professional at present
Cease the matching of page classification content and combine, and the present invention can automatically generate combinations thereof relation by computer program, with
Convenient enterprise carries out the renewal of dependent merchandise content and polymerization model.
The set of the above-mentioned classification content that same cross-border electric business platform is obtained and the sample of least member set to constituting
Neutral net polymerization model in Fig. 2 is input to, wherein merchandise news least member collection is combined into input value, commodity page category information
For output valve, by comparing to determine, if the model output valve of all samples is matched completely with commodity page category information, error
It is zero, illustrates that the model is capable of the descriptive labelling mapping of the true match electric business platform and syntagmatic, if there is project not meet,
Then illustrate that the connection weight of the model is not also trained to complete, it is still necessary to which GA is optimized, then go successively to iteration optimization training
Process, till every model output valve all meets with commodity page info classification content on training sample set, obtains final product
To on the cross-border electric business platform of target from merchandise news least member collection to the correct mapping of commodity page info classification content and group
Conjunction relation.
Specifically, referring to Fig. 3 and Fig. 4, the neural metwork training of off-line training pattern comprises the following steps A1-A4.
A1：The chromosome set of binary GA algorithms is generated at random, selects an initial chromosome as the first of neutral net
Beginning weight coefficient value, and fitness value is calculated, initial chromosome is assigned to into history optimal value.
More specifically, if N is ωpqAnd νqmNumber summation, i.e. N=p*q+q*m, then in GA algorithms chromosome be one group
The length of random generation is the binary code (such as 10 ... 01 ... 101 ... 10 ... 1) of N, if chromosome set is S=[S1,S2,…,
Sr-1,Sr], r is chromosome number, and r=20 in this example distributes to correspondingly integer 0-1 neutral nets according to Fig. 7 modes
Weights, then select the corresponding neutral net of first chromosome, calculate its fitness value, assign it to history optimal value
Sbest。
A2：Chromosome set is replicated, made a variation and is intersected, the chromosome population after being replicated, made a variation and being intersected is closed
Collection, calculates each fitness value using chromosome therein as the weight coefficient value of neutral net respectively.
More specifically, carry out the system GA of standard 2 to all 20 chromosomes to select, intersect and mutation operation, and with reference to Fig. 6
Parameter is mapped to into integer 0-1 neutral nets.
I, chance is chosen by what select probability p (si) was determined, choose 1 chromosome at random from S every time and answered
System, does altogether n times, then will replicate the N chromosome sets that obtain into colony S1；
Ii, the Chromosome number c intersected by the participation for determining, determine at random c chromosome from S1, and pairing carries out intersection behaviour
Make, and replace original chromosome with the chromosome for producing, constitute colony S2；
Iii, by the variation number of times m for determining, determine m chromosome at random from S2, carry out mutation operation respectively, and with product
Raw new chromosome replaces original chromosome, constitutes colony S3；
By the integer 0-1 neutral nets of the brand-new generation obtained by the intersection being made up of S1, S2 and S3 subset mapping, in sample
Fitness value is calculated on collection.
A3：The chromosome of minimum fitness in each fitness value is selected, if the minimum fitness is optimum less than history
Value, then be assigned to history optimal value by corresponding chromosome, otherwise keeps history optimal value constant, the new dye produced to variation
Colour solid set is screened based on fitness function using roulette mode, obtains follow-on chromosome set, waits iteration.
More specifically, sort from small to large according to fitness function, select the chromosome of minimum fitness, if its adaptation
Degree is than history optimal value SbestAlso little, then it is better than history optimal value Sbest, assign it to history optimal value Sbest, otherwise, keep
History optimal value SbestIt is constant, subsequently roulette mode is adopted based on fitness function to the new chromosome congression that variation is produced
Screened, obtained the chromosome set of chromosome of future generation.
A4：When judging that the chromosome of history optimal value is applied in neutral net polymerization model, the model of all samples pair
Whether output is all consistent with sample output, if then terminating iterative learning, exports the optimum chromosome, otherwise selects of future generation
Chromosome set return to step A2.
In one embodiment, neural metwork training is additionally included in line operational mode, referring to Fig. 3 and Fig. 5, on-line operation mould
Neural metwork training under formula is comprised the following steps：B1：According to the corresponding neutral net polymerization model of target electric business platform selecting
Weight coefficient and structure, generate corresponding neutral net polymerization model；B2：Input minimal information element set, calls nerve
Network polymerization model calculates the category information of the commodity page of target electric business platform.
On-line operation pattern i.e. on the basis of the neutral net polymerization model obtained by off-line training pattern, by commodity most
Little information element set is input into polymerization model with the category information of target platform, by being calculated the cross-border electric business platform business of target
Product page category information.By the pattern, system can at any time be dynamically generated the cross-border electric business platform commodity page classification of target
Information, compared with the past dependence manually carries out the traditional mode of finish message integration, can automatically generate commodity page by computer
Surface information classification content, greatly improves work efficiency, reduces error rate.
In actual applications, with differences such as supplier and raw materials, sometimes merchandise news needs to be changed, additionally,
The page-describing classification of cross-border electric business platform can change as the platform of cross-border electric business service provider updates, manually not
Tracking can in real time be gone, and these change and carry out the renewal of model and information, therefore, in the base for obtaining neutral net polymerization model
On plinth, merchandise news is monitored by dynamic detection and is changed with cross-border electric business platform sample commodity classes of pages purpose, to page classification
Content and classification carry out dynamic optimization and adjustment, as shown in fig. 6, its idiographic flow is as follows：
(1) the monitoring cycle, it is flat with cross-border electric business that dynamic access is stored in local system article minimal information set
The category information of the platform sample commodity page；
(2) judge being stored in local commodity minimal information collection situation of change, can compare by way of reality
It is existing, if do not changed, step (3) is proceeded to, in the event of changing, then carry out following process：
I, neutral net polymerization model is re-entered in the commodity least member set of renewal calculated, obtain what is updated
Page category information；
Ii, the page classification information output of renewal is updated to target electric business platform；
Iii, judge whether all platforms update and finish, in this way, then into step (3), if not being, return the circulation of the i-th step
Perform to the commodity to be updated in the page of all electric business platforms and finish；
(3) the category information situation of change of the electric business platform sample commodity page to being monitored judges, if do not sent out
Changing, goes back to step (1) into next monitoring cycle, in the event of changing, then carries out following process：
I, the new learning sample of commodity page info structure for obtaining change platform；
Ii, re -training is carried out to target platform neutral net polymerization model；
Iii, by original minimal information element set of commodity to be polymerized be input to update after neutral net polymerization model
Calculated, obtained the commodity page category information after paradigmatic relation change, and be uploaded to the cross-border electric business platform of target, gone back to step
Suddenly (1) is into next monitoring cycle.
The present invention also provides a kind of multi-platform electronic commerce information paradigmatic system, including：
Platform information acquisition module, to obtain page category information of each electric business platform with regard to sample commodity；
Information similarity analysis integrate module, to the same sample commodity to obtaining on the different cross-border electric business platform pages
Whole category informations carry out similarity analysis, the minimal information element of sample commodity is parsed based on similarity analysis result,
Merge the minimal information element set that minimal information element obtains the sample commodity, the minimal information element set of multiple sample commodity
Constitute comprehensive sample set；
Polymerization model training module, to for each electric business platform, with the minimal information element set of sample commodity with from
The category information that one electric business platform is obtained carries out the neural metwork training that weight coefficient value is 0 or 1 as training sample pair,
Obtain the neutral net polymerization model of each electric business platform；
Aggregation information computing module, to the neutral net polymerization model according to different electric business platforms, is input into and waits to gather
The minimal information element set of commodity is closed, Jing models are calculated the category information of correspondence electric business platform commodity page request.
According to one embodiment of present invention, also include：
Polymerization model dynamic update module, to when merchandise news is changed, according to the minimal information element set for updating
The neutral net polymerization model is re-entered into, the category information that each electric business platform updates is obtained, and is uploaded to corresponding electricity
Business's platform；And/or, to (the i.e. commodity when the category information of the electric business platform sample commodity page for being monitored changes
When paradigmatic relation between minimal information element set and page category information changes), carry out similarity analysis and parsing is closed
And new minimal information element set is obtained, build the new sample pair of target electric business platform, the new neutral net of re -training
Polymerization model, original minimal information element set of commodity to be polymerized is input in the neutral net polymerization model after updating
Category information after being changed, and it is uploaded to the target electric business platform.
Aforementioned multi-platform electronics is may refer to regard to the particular content of multi-platform electronic commerce information paradigmatic system of the invention
The description of commercial matters information polymerization part, specifically repeats no more.
Although the present invention is disclosed as above with preferred embodiment, it is not for limiting claim, any this area
Technical staff without departing from the spirit and scope of the present invention, can make possible variation and modification, therefore the present invention
The scope that protection domain should be defined by the claims in the present invention is defined.
Claims (11)
1. a kind of multi-platform electronic commerce information polymerization, it is characterised in that comprise the following steps：
Obtain page category information of each electric business platform with regard to sample commodity；
Whole category informations of the same sample commodity on the different electric business platform pages to obtaining carry out similarity analysis, are based on
Similarity analysis result parses the minimal information element of sample commodity, merges minimal information element and obtains the sample commodity most
Little information element set, the minimal information element set of multiple sample commodity constitutes comprehensive sample set；
For each electric business platform, made with the category information obtained from an electric business platform with the minimal information element set of sample commodity
For training sample pair, the neural metwork training that weight coefficient value is 0 or 1 is carried out, the neutral net for obtaining each electric business platform is gathered
Matched moulds type；
According to the neutral net polymerization model of different electric business platforms, the minimal information element set of commodity to be polymerized is input into,
Jing models are calculated the category information of correspondence electric business platform commodity page request.
2. multi-platform electronic commerce information polymerization as claimed in claim 1, it is characterised in that also include：
When merchandise news is changed, the neutral net polymerization mould is re-entered into according to the minimal information element set for updating
Type, obtains the category information that each electric business platform updates, and is uploaded to corresponding electric business platform.
3. multi-platform electronic commerce information polymerization as claimed in claim 1, it is characterised in that also include：
When the category information of the electric business platform sample commodity page for being monitored changes, carry out similarity analysis and parsing is closed
And new minimal information element set is obtained, build the new sample pair of target electric business platform, the new neutral net of re -training
Polymerization model, original minimal information element set of commodity to be polymerized is input in the neutral net polymerization model after updating
Category information after being changed, and it is uploaded to the target electric business platform.
4. the multi-platform electronic commerce information polymerization as described in any one in claim 1-3, it is characterised in that pass through
The dependent merchandise web page resources of each electric business platform of crawl are oriented based on target data model using focused crawler, to obtain each electric business
Category information of the platform with regard to commodity.
5. multi-platform electronic commerce information polymerization as claimed in claim 1, it is characterised in that each neutral net polymerization
Model is mapping model of the minimal information element set to the category information of an electric business platform, including one layer of input layer, multilamellar or
Monolayer hidden layer, one layer of output layer, input number of nodes is according to the determination of the element number of minimum unit information aggregate, hidden layers numbers and section
Points determine that the classification quantity of the corresponding electric business platform of the several evidences of output node determines according to the granularity of information fusion.
6. the multi-platform electronic commerce information polymerization as described in claim 1 or 5, it is characterised in that adopt neutral net
Feedback compensation algorithm is trained and optimizing to the parameter of neutral net polymerization model.
7. multi-platform electronic commerce information polymerization as claimed in claim 6, it is characterised in that it is described be trained and
Optimizing includes：
Using binary GA algorithms, the weight parameter value of neutral net polymerization model is 0 or 1, hidden node and output node
Summation is weighted to input information, the synthesis of different information elements is obtained；
The output valve of comparative neural network polymerization model and the output valve of sample pair, are carried out more according to comparative result to weight parameter
Change, to obtain being adapted to the neutral net polymerization model of each electric business platform.
8. the multi-platform electronic commerce information polymerization as described in any one in claim 1-3,5, it is characterised in that god
Jing network trainings include off-line training pattern, comprise the following steps：
A1：The chromosome set of binary GA algorithms is generated at random, selects an initial chromosome as the initial power of neutral net
Weight coefficient value, and fitness value is calculated, initial chromosome is assigned to into history optimal value；
A2：Chromosome set is replicated, made a variation and is intersected, the chromosome population intersection after being replicated, made a variation and being intersected, point
Chromosome therein each fitness value is not calculated into as the weight coefficient value of neutral net；
A3：The chromosome of minimum fitness in each fitness value is selected, if the minimum fitness is less than history optimal value,
Corresponding chromosome is assigned to into history optimal value, otherwise keeps history optimal value constant, the new chromosome produced to variation
Set is screened based on fitness function using roulette mode, obtains follow-on chromosome set, waits iteration；
A4：When judging that the chromosome of history optimal value is applied in neutral net polymerization model, the model output of all samples pair
Whether all it is consistent with sample output, if then terminating iterative learning, exports the optimum chromosome, otherwise selects follow-on dye
Colour solid group return to step A2.
9. multi-platform electronic commerce information polymerization as claimed in claim 8, it is characterised in that neural metwork training is also wrapped
On-line operation pattern is included, is comprised the following steps：
B1：According to the weight coefficient and structure of the corresponding neutral net polymerization model of target electric business platform selecting, generate corresponding
Neutral net polymerization model；
B2：Input minimal information element set, calls neutral net polymerization model to calculate the commodity page of target electric business platform
Category information.
10. a kind of multi-platform electronic commerce information paradigmatic system, it is characterised in that include：
Platform information acquisition module, to obtain page category information of each electric business platform with regard to sample commodity；
Information similarity analysis integrate module, complete on the different cross-border electric business platform pages to the same sample commodity to obtaining
Portion's category information carries out similarity analysis, and based on similarity analysis result the minimal information element of sample commodity is parsed, and merges
Minimal information element obtains the minimal information element set of the sample commodity, and the minimal information element set of multiple sample commodity is constituted
Comprehensive sample set；
Polymerization model training module, it is electric with from one with the minimal information element set of sample commodity to for each electric business platform
The category information that business's platform is obtained carries out the neural metwork training that weight coefficient value is 0 or 1 as training sample pair, obtains
The neutral net polymerization model of each electric business platform；
Aggregation information computing module, to the neutral net polymerization model according to different electric business platforms, is input into business to be polymerized
The minimal information element set of product, Jing models are calculated the category information of correspondence electric business platform commodity page request.
11. multi-platform electronic commerce information paradigmatic systems as claimed in claim 10, it is characterised in that also include：
Polymerization model dynamic update module, to when merchandise news is changed, according to the minimal information element set for updating again
The neutral net polymerization model is input to, the category information that each electric business platform updates is obtained, and it is flat to be uploaded to corresponding electric business
Platform；And/or, when the category information of the electric business platform sample commodity page for being monitored changes, to carry out similarity point
Analysis and parsing merging obtain new minimal information element set, build the new sample pair of target electric business platform, and re -training is new
Neutral net polymerization model, by original minimal information element set of commodity to be polymerized be input to update after neutral net
Category information after being changed in polymerization model, and it is uploaded to the target electric business platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610957162.3A CN106600356B (en) | 2016-10-27 | 2016-10-27 | Multi-platform e-commerce information aggregation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610957162.3A CN106600356B (en) | 2016-10-27 | 2016-10-27 | Multi-platform e-commerce information aggregation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106600356A true CN106600356A (en) | 2017-04-26 |
| CN106600356B CN106600356B (en) | 2019-12-24 |
Family
ID=58590357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610957162.3A Active CN106600356B (en) | 2016-10-27 | 2016-10-27 | Multi-platform e-commerce information aggregation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106600356B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107464170A (en) * | 2017-08-15 | 2017-12-12 | 合肥蓝胖子科技有限公司 | Mobile terminal electric business service system based on internet |
| CN107483582A (en) * | 2017-08-16 | 2017-12-15 | 广州环球梦电子商务股份有限公司 | It is a kind of to borrow the method for pushing and system for selling platform commodity |
| CN108257031A (en) * | 2017-11-09 | 2018-07-06 | 中国平安人寿保险股份有限公司 | Medical insurance product dissemination method, device and storage medium |
| CN108335164A (en) * | 2017-01-20 | 2018-07-27 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and electronic equipment for realizing shopping at network |
| CN108665358A (en) * | 2018-08-03 | 2018-10-16 | 北京引领海逛科技有限公司 | The method and apparatus of the multiple platforms of product information Rapid matching |
| CN109389450A (en) * | 2017-08-08 | 2019-02-26 | 阿里巴巴集团控股有限公司 | A kind of method and system of relating attribute information and generation data acquisition system |
| CN109947782A (en) * | 2017-11-03 | 2019-06-28 | ***通信有限公司研究院 | A kind of update method of big data real-time application system, apparatus and system |
| CN112085568A (en) * | 2020-09-09 | 2020-12-15 | 百度在线网络技术（北京）有限公司 | Commodity and rich media aggregation display method and device, electronic device and medium |
| CN112231311A (en) * | 2020-10-23 | 2021-01-15 | 食亨(上海)科技服务有限公司 | Method, system and computer readable medium for integrating cross-platform commodity data |
| CN112685320A (en) * | 2021-01-09 | 2021-04-20 | 北京信息科技大学 | Software defect repairing method and device based on multiple candidate programs |
| CN109224453B (en) * | 2018-06-28 | 2021-07-09 | 平安科技（深圳）有限公司 | Game supervision method and system, computer equipment and computer-readable storage medium |
| WO2022269336A1 (en) * | 2021-06-23 | 2022-12-29 | Ddb Mudra Max Private Limited | System and method for intelligent cross platform marketing resource optimization |
| CN117763422A (en) * | 2023-12-25 | 2024-03-26 | 北京燃数科技有限公司 | SaaS (software as a service) investment system for solving investment strategy |
| CN118247020A (en) * | 2024-05-27 | 2024-06-25 | 深圳美云集网络科技有限责任公司 | Multi-platform commodity release method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101118610A (en) * | 2007-09-10 | 2008-02-06 | 东北大学 | Sparseness data process modeling approach |
| US20080306830A1 (en) * | 2007-06-07 | 2008-12-11 | Cliquality, Llc | System for rating quality of online visitors |
| CN103914478A (en) * | 2013-01-06 | 2014-07-09 | 阿里巴巴集团控股有限公司 | Webpage training method and system and webpage prediction method and system |
| CN104504086A (en) * | 2014-12-25 | 2015-04-08 | 北京国双科技有限公司 | Clustering method and device for webpage |
| CN104765728A (en) * | 2014-01-08 | 2015-07-08 | 富士通株式会社 | Method and device for training neural network and method for determining sparse feature vector |
| CN105005917A (en) * | 2015-07-07 | 2015-10-28 | 上海晶赞科技发展有限公司 | Universal method for correlating single items of different e-commerce websites |
-
2016
- 2016-10-27 CN CN201610957162.3A patent/CN106600356B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080306830A1 (en) * | 2007-06-07 | 2008-12-11 | Cliquality, Llc | System for rating quality of online visitors |
| CN101118610A (en) * | 2007-09-10 | 2008-02-06 | 东北大学 | Sparseness data process modeling approach |
| CN103914478A (en) * | 2013-01-06 | 2014-07-09 | 阿里巴巴集团控股有限公司 | Webpage training method and system and webpage prediction method and system |
| CN104765728A (en) * | 2014-01-08 | 2015-07-08 | 富士通株式会社 | Method and device for training neural network and method for determining sparse feature vector |
| CN104504086A (en) * | 2014-12-25 | 2015-04-08 | 北京国双科技有限公司 | Clustering method and device for webpage |
| CN105005917A (en) * | 2015-07-07 | 2015-10-28 | 上海晶赞科技发展有限公司 | Universal method for correlating single items of different e-commerce websites |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108335164A (en) * | 2017-01-20 | 2018-07-27 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and electronic equipment for realizing shopping at network |
| CN109389450B (en) * | 2017-08-08 | 2022-04-19 | 阿里巴巴集团控股有限公司 | Method and system for associating attribute information and generating data set |
| CN109389450A (en) * | 2017-08-08 | 2019-02-26 | 阿里巴巴集团控股有限公司 | A kind of method and system of relating attribute information and generation data acquisition system |
| CN107464170A (en) * | 2017-08-15 | 2017-12-12 | 合肥蓝胖子科技有限公司 | Mobile terminal electric business service system based on internet |
| CN107483582A (en) * | 2017-08-16 | 2017-12-15 | 广州环球梦电子商务股份有限公司 | It is a kind of to borrow the method for pushing and system for selling platform commodity |
| CN109947782A (en) * | 2017-11-03 | 2019-06-28 | ***通信有限公司研究院 | A kind of update method of big data real-time application system, apparatus and system |
| CN108257031A (en) * | 2017-11-09 | 2018-07-06 | 中国平安人寿保险股份有限公司 | Medical insurance product dissemination method, device and storage medium |
| CN109224453B (en) * | 2018-06-28 | 2021-07-09 | 平安科技（深圳）有限公司 | Game supervision method and system, computer equipment and computer-readable storage medium |
| CN108665358A (en) * | 2018-08-03 | 2018-10-16 | 北京引领海逛科技有限公司 | The method and apparatus of the multiple platforms of product information Rapid matching |
| CN112085568A (en) * | 2020-09-09 | 2020-12-15 | 百度在线网络技术（北京）有限公司 | Commodity and rich media aggregation display method and device, electronic device and medium |
| CN112085568B (en) * | 2020-09-09 | 2023-10-27 | 百度在线网络技术（北京）有限公司 | Commodity and rich media aggregation display method and equipment, electronic equipment and medium |
| CN112231311A (en) * | 2020-10-23 | 2021-01-15 | 食亨(上海)科技服务有限公司 | Method, system and computer readable medium for integrating cross-platform commodity data |
| CN112685320B (en) * | 2021-01-09 | 2023-04-25 | 北京信息科技大学 | Software defect repairing method and device based on multiple candidate programs |
| CN112685320A (en) * | 2021-01-09 | 2021-04-20 | 北京信息科技大学 | Software defect repairing method and device based on multiple candidate programs |
| WO2022269336A1 (en) * | 2021-06-23 | 2022-12-29 | Ddb Mudra Max Private Limited | System and method for intelligent cross platform marketing resource optimization |
| CN117763422A (en) * | 2023-12-25 | 2024-03-26 | 北京燃数科技有限公司 | SaaS (software as a service) investment system for solving investment strategy |
| CN118247020A (en) * | 2024-05-27 | 2024-06-25 | 深圳美云集网络科技有限责任公司 | Multi-platform commodity release method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106600356B (en) | 2019-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106600356A (en) | Multi-platform electronic commerce information aggregation method and system | |
| Özceylan et al. | Fuzzy multi-objective linear programming approach for optimising a closed-loop supply chain network | |
| CN106326248B (en) | The storage method and device of database data | |
| Lambert | Do We Need a'real'Taxonomy of E-business Models? | |
| Kuo et al. | The integration of association rule mining and artificial immune network for supplier selection and order quantity allocation | |
| CN107292775A (en) | A kind of service system for intellectual property rights and its implementation | |
| Drechsler | Model-based integration of ecology and socio-economics for the management of biodiversity and ecosystem services: State of the art, diversity and current trends | |
| Magliocca et al. | Using Pattern‐oriented Modeling (POM) to Cope with Uncertainty in Multi‐scale Agent‐based Models of Land Change | |
| Zhang et al. | A study of rural logistics center location based on intuitionistic fuzzy TOPSIS | |
| Zhang | [Retracted] Optimization of the Marketing Management System Based on Cloud Computing and Big Data | |
| CN110310012A (en) | Data analysing method, device, equipment and computer readable storage medium | |
| Nodoust et al. | An evidential reasoning approach for production modeling with deteriorating and ameliorating items | |
| Xiao et al. | Modeling and intelligent optimization of social collective behavior with online public opinion synchronization | |
| Wu | [Retracted] Optimization of E‐Commerce Supply Chain Management Process Based on Internet of Things Technology | |
| CN114861050A (en) | Feature fusion recommendation method and system based on neural network | |
| Qu et al. | Parallel genetic algorithm model based on AHP and neural networks for enterprise comprehensive business | |
| Zhang et al. | A hybrid attention and time series network for enterprise sales forecasting under digital management and edge computing | |
| Yan et al. | A data-mining approach for product conceptualization in a web-based architecture | |
| Xiong et al. | WebATP: a Web-based flexible available-to-promise computation system | |
| Zhang | B2C E-commerce logistics network optimization model | |
| Wang et al. | A Sustainable Supply Chain Design for Personalized Customization in Industry 5.0 Era | |
| Mohanty et al. | Application of fuzzy multi attribute decision making analysis to rank web services | |
| Liu et al. | Research of second-hand real estate price forecasting based on data mining | |
| KR20220108970A (en) | Method of risk analysis and technology transfer opportunity recommendation by technology-product-material based on machine learning or AI | |
| Liu | [Retracted] Fractal Art Pattern Generation Based on Genetic Algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201010 Address after: Room 301-303, 305-308, building 2, No. 567, Jiangling Road, Xixing street, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: HANGZHOU PINGPONG INTELLIGENT TECHNICAL Co.,Ltd. Address before: Hangzhou City, Zhejiang province 310000 Binjiang District West Street No. 567 Jiangling Road, building 2, floor 5 Patentee before: HANGZHOU WANGDAO TECHNOLOGY Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Room 301-303, 305-308, building 2, No. 567, Jiangling Road, Xixing street, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou ping pong Intelligent Technology Co., Ltd Address before: Room 301-303, 305-308, building 2, No. 567, Jiangling Road, Xixing street, Binjiang District, Hangzhou City, Zhejiang Province Patentee before: HANGZHOU PINGPONG INTELLIGENT TECHNICAL Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 310000 rooms 301-303, 305-308, building 2, No.567 Jiangling Road, Xixing street, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou ping pong Intelligent Technology Co.,Ltd. Address before: 310000 rooms 301-303, 305-308, building 2, No.567 Jiangling Road, Xixing street, Binjiang District, Hangzhou City, Zhejiang Province Patentee before: Hangzhou ping pong Intelligent Technology Co., Ltd |