WO2017162074A1 - 商品映射方法、装置及设备 - Google Patents
商品映射方法、装置及设备 Download PDFInfo
- Publication number
- WO2017162074A1 WO2017162074A1 PCT/CN2017/076708 CN2017076708W WO2017162074A1 WO 2017162074 A1 WO2017162074 A1 WO 2017162074A1 CN 2017076708 W CN2017076708 W CN 2017076708W WO 2017162074 A1 WO2017162074 A1 WO 2017162074A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- information
- product
- mapping
- item
- commodity
- 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.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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]
- G06Q30/0621—Electronic shopping [e-shopping] by configuring or customising goods or services
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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—Electronic shopping [e-shopping] utilising user interfaces specially adapted for shopping
- G06Q30/0643—Electronic shopping [e-shopping] utilising user interfaces specially adapted for shopping graphically representing goods, e.g. 3D product representation
Definitions
- the present invention relates to the field of electronic commerce, and in particular, to a commodity mapping method, apparatus and device.
- E-commerce refers to the use of browser/server applications in business network activities, in the open Internet environment, where buyers and sellers do not face all kinds of business activities, realizing online shopping for consumers and online transactions between merchants. And a new type of business operation model for online electronic payments and various business activities, trading activities, financial activities and related integrated service activities.
- the realization of e-commerce depends on the e-commerce platform, referred to as the e-commerce platform.
- the e-commerce platform is a platform for companies or individuals to provide online trading negotiations.
- the enterprise e-commerce platform is a virtual network space for business activities on the Internet and a management environment for ensuring the smooth operation of business. It is an important place for coordinating and integrating information flow, goods flow, orderly flow, correlation and efficient flow of funds. Enterprises and merchants can make full use of the shared resources such as network infrastructure, payment platform, security platform and management platform provided by e-commerce platform to carry out their own business activities effectively and at low cost.
- the e-commerce platform is developing towards internationalization, offline and platform.
- an urgent need is to convert the goods of other e-commerce platforms into commodities of their own e-commerce platform, and then quickly release them to their e-commerce platform.
- the product description formats of the e-commerce platforms are completely different, it is necessary to map the products across the e-commerce platform so that the product formats of other e-commerce platforms are compatible with their own e-commerce platforms.
- the merchandise mapping of the cross-e-commerce platform is mainly performed by manual means.
- the manual mapping of goods across E-commerce platforms is not only inefficient, but also requires a large amount of labor costs, resulting in an increase in the operating costs of the e-commerce platform.
- the object of the present invention is to provide a commodity mapping method, device and device, which realize automatic mapping of products across e-commerce platforms and improve mapping efficiency.
- the present invention provides a commodity mapping method, including:
- the first product information is mapped to the second product information according to the pre-trained product mapping model, and the second product information is the product information of the second e-commerce platform product.
- the above method may further have the following features, and further includes:
- the commodity mapping model is trained by using the acquired second commodity information and search record information.
- the foregoing method may further have the following feature: the using the acquired second commodity information and the search record information to train the commodity mapping model, including:
- the search record information includes a search phrase and a user clicked product corresponding to the search phrase, and the user category of the clicked product is used as a product category associated with the search phrase, and the search phrase is associated with the product category.
- Item recorded as the first related item;
- the search phrase is decomposed to obtain at least one participle, and the search phrase is replaced with each participle, and the item of the word segmentation and the item category is obtained, which is recorded as the second item.
- mapping the first product information to the second product information according to the pre-trained product mapping model including:
- the foregoing method may further have the following feature: the predicting the category information in the second product information from the product title information in the first product information according to the second related item, including:
- the item category information included in the found second related item is used as the category information in the second item information.
- mapping the first product information to the second product information according to the pre-trained product mapping model including:
- the product category information in the first product information is directly mapped to the category information in the second product information by literal matching.
- the above method may further have the following feature, the utilizing the acquired second commodity information and the search record Information training results in the commodity mapping model, including:
- the annotation data is trained to obtain an attribute pair extraction model, and the attribute extraction model is used to extract the attribute pair information from the unstructured text.
- mapping the first product information to the second product information according to the pre-trained product mapping model including:
- mapping the first product information to the second product information according to the pre-trained product mapping model including:
- the attribute pair information in the first item information is directly mapped to the attribute pair information in the second item information by literal matching.
- mapping the first product information to the second product information according to the pre-trained product mapping model including:
- mapping the first product information to the second product information according to the pre-trained product mapping model including:
- the product mapping method of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, improve the product mapping efficiency of the cross-e-commerce platform, and reduce the operating cost of the e-commerce platform.
- the present invention also provides a commodity mapping apparatus, comprising:
- a first acquiring module configured to acquire commodity information of the first e-commerce platform commodity, and record the first commodity information
- mapping module configured to map, according to the pre-trained product mapping model, the first product information acquired by the first acquiring module into second product information, where the second product information is product information of the second e-commerce platform product .
- the above device may further have the following features, and further includes:
- a second acquiring module configured to acquire a second quantity of the second item information and the user on the second e-commerce platform Search record information
- the training module is configured to train the commodity mapping model by using the second commodity information and the search record information acquired by the second acquiring module.
- the training module includes:
- the search record information includes a search phrase and a user click item corresponding to the search phrase
- the creating unit is configured to create a search by using the item category of the item clicked by the user as the item category associated with the search phrase.
- the association between the phrase and the product category is recorded as the first association item;
- Decomposing the association unit configured to decompose the search phrase, obtain at least one word segment, replace the search phrase with each participle, and obtain a term item and a product category association item, and record the second association item.
- mapping module includes:
- the first category mapping unit is configured to predict category information in the second commodity information from the commodity title information in the first commodity information according to the second associated item obtained by the decomposition association unit.
- the first category mapping unit includes:
- a word segment subunit configured to decompose the product category information and/or the product title information in the first product information into at least one word segment, and record it as a matching word segment;
- a prediction subunit configured to use the commodity category information included in the found second association item as the category information in the second commodity information.
- mapping module includes:
- the second category mapping unit is configured to directly map the product category information in the first product information to the category information in the second product information by literal matching.
- the training module includes:
- An extracting unit configured to extract product title information and item detail information from the second item information
- a matching unit configured to use the product title information and the product detail information extracted by using the structured attribute data of the second e-commerce platform to obtain the attribute-to-label data of the second e-commerce platform in the unstructured text.
- a training unit configured to use the attribute of the second e-commerce platform to train the annotation data to obtain an attribute pair extraction model, and the attribute extraction model is used to extract the attribute pair information from the unstructured text.
- mapping module includes:
- the first attribute pair mapping unit is configured to extract attribute pair information of the second item information from the item title information and the item detail information of the first item information according to the attribute pair extraction model.
- mapping module includes:
- the second attribute pair mapping unit is configured to directly map the attribute pair information in the first item information to the attribute pair information in the second item information by literal matching.
- mapping module includes:
- the title mapping unit is configured to convert the product title information in the first product information into the title information in the second product information according to the first text conversion model in the product mapping model.
- mapping module includes:
- the detail mapping unit is configured to convert the commodity detail information in the first commodity information into the detail information in the second commodity information according to the second text conversion model in the commodity mapping model.
- the commodity mapping device of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, improve the product mapping efficiency of the cross-e-commerce platform, and reduce the operating cost of the e-commerce platform.
- the present invention also provides a commodity mapping method, including:
- the commodity information of the matching commodity is mapped to the commodity information of the second e-commerce platform as the commodity of the first e-commerce platform.
- the product mapping method of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, improve the product mapping efficiency of the cross-e-commerce platform, and reduce the operating cost of the e-commerce platform.
- the present invention also provides a commodity mapping apparatus, comprising:
- a barcode obtaining module configured to obtain barcode information of the first e-commerce platform product
- a barcode matching module configured to search for a commodity having the barcode information from a second e-commerce platform, and record it as a matching commodity
- the barcode mapping module is configured to map the commodity information of the matching commodity as the commodity of the first e-commerce platform to the commodity information of the second e-commerce platform.
- the commodity mapping device of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, improve the product mapping efficiency of the cross-e-commerce platform, and reduce the operating cost of the e-commerce platform.
- the present invention also provides a commodity mapping apparatus comprising the commodity mapping apparatus according to any of the preceding claims.
- the commodity mapping device of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, improve the product mapping efficiency of the cross-e-commerce platform, and reduce the operating cost of the e-commerce platform.
- FIG. 1 is a flowchart of a method for mapping goods in a first embodiment of the present invention
- FIG. 2 is a flowchart of a method for mapping goods in a second embodiment of the present invention.
- FIG. 3 is a structural block diagram of a commodity mapping apparatus according to Embodiment 3 of the present invention.
- FIG. 4 is a flowchart of a method for mapping goods in a fourth embodiment of the present invention.
- FIG. 5 is a structural block diagram of a commodity mapping apparatus according to Embodiment 5 of the present invention.
- FIG. 6 is a structural block diagram of a commodity mapping device according to Embodiment 6 of the present invention.
- FIG. 7 is a structural block diagram of a commodity mapping device in Embodiment 7 of the present invention.
- FIG. 8 is a schematic diagram of commodity mapping in an embodiment of the present invention.
- first, second, third, etc. may be used to describe XXX in embodiments of the invention, these XXX should not be limited to these terms. These terms are only used to distinguish XXX from each other.
- first XXX may also be referred to as a second XXX without departing from the scope of the embodiments of the present invention.
- second XXX may also be referred to as a first XXX.
- FIG. 1 is a flowchart of a method for mapping goods in a first embodiment of the present invention.
- the commodity mapping method may include the following steps:
- Step S101 Obtain product information of the first e-commerce platform product, and record it as the first product information;
- the item information may include item title information, item category information, item attribute pair information, item detail information, and the like.
- specific product information of the merchandise of the booth can be obtained.
- the product description format of each e-commerce platform is completely different, it generally includes the most basic information such as product title information, product category information, and product detail information, and the commodity information of some e-commerce platforms also includes product attribute pair information.
- the first e-commerce platform is a platform for commodity mapping, and this article is called the source e-commerce platform.
- the second e-commerce platform below is the target platform to which the goods are to be mapped, and is referred to herein as the target e-commerce platform.
- Step S102 Map the first product information to the second product information according to the pre-trained product mapping model, wherein the second product information is the product information of the second e-commerce platform product.
- the merchandise mapping model is trained based on the merchandise data of the second e-commerce platform. According to the commodity mapping model, The product information of the source e-commerce platform is mapped to the target e-commerce platform.
- the mapping of the commodities may include mappings of categories, attribute pairs, titles, details, and the like.
- One of the example mapping manners of the product category may be: directly mapping the product category information in the first product information to the category information in the second product information by literal matching.
- literal matching refers to the matching by the editing distance of the string literal. For example, if the item category information in the first item information is “winter dress” and the category information in the second item information is “dress”, the “winter dress” in the first item information is directly mapped by literal matching. It is the "dress” in the second product information.
- the second example of the mapping of product categories will be described in the second embodiment below.
- One of the example mapping manners of the commodity attribute pair may be: directly mapping the attribute pair information in the first item information to the attribute pair information in the second item information by literal matching.
- the attribute pair information in the first item information is “pink”
- the attribute pair information in the second item information is “pink”
- the “pink” in the first item information is directly mapped to the second by literal matching.
- "Pink” in the product information is directly mapped to the second by literal matching.
- the mapping manner may be: converting the product title information in the first product information into the title information in the second product information according to the first text conversion model in the product mapping model.
- the first text conversion model may be a machine translation model, such as translating the product title information of the traditional characters into the product title information of the simplified characters.
- the mapping manner may be: converting the product detailed information in the first product information into the detailed information in the second product information according to the second text conversion model in the product mapping model.
- the second text conversion model may be a machine translation model, such as translating the English product title information into the simplified Chinese product title information.
- the second text conversion model may be the same as the first text conversion model described above.
- the commodity mapping method of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, and does not require manual intervention, thereby greatly improving the commodity mapping efficiency of the cross-e-commerce platform. Moreover, the commodity mapping method of the embodiment of the present invention saves labor costs and reduces the operating cost of the e-commerce platform because it does not need to invest a large amount of manpower.
- the commodity mapping method may include the following steps:
- Step S201 acquiring second item information of a set number or more and a search record letter of the user on the second e-commerce platform interest;
- the second product information is product information of the second e-commerce platform product.
- Step S202 training the commodity mapping model by using the acquired second commodity information and the search record information
- the acquired commodity category mapping model, the mapping model of the commodity attribute pair, and the like can be trained by using the acquired second commodity information and the search record information.
- mapping model of the commodity category can be trained.
- the training of the commodity mapping model by using the acquired second commodity information and the search record information may include:
- the search record information may include a search phrase and a user clicked product corresponding to the search phrase, and the product category of the user clicked on the product as a product category associated with the search phrase, and creates an association term between the search phrase and the product category, which is recorded as the first Related item
- the second related item is the mapping model of the commodity category obtained by the training, and the second related item may be referred to as the category prediction model.
- the search phrase input by the user on the e-commerce platform is generally a description of the characteristics of the product that the user is looking for, and the click data after the search phrase is generally the product that the user is looking for. Therefore, the item that the user clicks on should be the item associated with the search phrase entered by the user.
- the accuracy of the category prediction can be improved according to the user's search phrase and the user's click on the product training category prediction model corresponding to the search phrase.
- the user inputs a search phrase "winter dress" on the second e-commerce platform, and after the search, the user clicks on the product a, the product b, and the product c, and the product categories of the product a, the product b, and the product c are respectively the category 1 , Category 2, Category 3.
- search phrase "winter dress” is broken down into two participles “winter” and "dress”, and the search term “winter dress” is replaced with the participle "winter” to get 3 second related items, namely:
- mapping model of the commodity attribute pair can be trained.
- the training of the commodity mapping model by using the acquired second commodity information and the search record information may include:
- the attribute data extraction model is trained on the annotation data by using the attribute of the second e-commerce platform, and the attribute extraction model is used to extract the attribute pair information from the unstructured text.
- the structured attribute data refers to the description structure of the goods by the e-commerce platform.
- the "baby details" of the product has the following contents: item number, style, combination form, skirt length, sleeve length, collar type, sleeve type, waist type, and the like.
- These contents are the structured attribute data of Taobao.com's e-commerce platform.
- the attribute pair extraction model is a mapping model of the pair of commodity attributes obtained by the training.
- the product title information in the second product information is: Haiqing Blue 2016 Spring New Women's Vintage Print Commuter
- Haiqing Blue 2016 Spring New Women's Vintage Print Commuter A-line skirt round neck sleeve loose dress, marked with the attribute pairs in this title, such as color: Haiqing blue, listing year :2016, style: retro, pattern: print, collar type: round neck, sleeve length: middle sleeve.
- HMM Hidden Markov Model
- CRF Consumer Random Field
- Step S203 acquiring product information of the first e-commerce platform product, and recording it as the first product information
- Step S204 Map the first product information to the second product information according to the pre-trained product mapping model.
- step S202 the second example mapping method of the commodity category and the second exemplary mapping manner of the commodity attribute pair are explained.
- the second example of the product category mapping manner may be: predicting the category information in the second product information from the product title information in the first product information according to the second related item obtained by the foregoing training. .
- the first item is predicted from the item title information in the first item information.
- the category information in the second product information may include: decomposing the product category information and/or the product title information in the first product information into at least one participle, as a matching participle; and searching for the second related item including the matching participle
- the item category information included in the found second related item is used as the category information in the second item information.
- the product category of the source product is "long-sleeved dress", which breaks down the “long-sleeved dress” into two participles “long sleeve” and “dress”, and "long sleeve” and “dress” are matching participles.
- there are three second related items (dress---category 1; dress---category 2; dress---category 3) including "dress”, and therefore, the product category of the target product is predicted.
- category 1 category 2
- dress---category 3 including "dress”
- the product category of the target product is predicted.
- category 1 category 2
- dress---category 3 that is, these three categories are the commodity categories of the target goods.
- the second example of the product attribute pair may be: extracting the attribute of the second item information from the item title information and the item detail information of the first item information according to the attribute pair extraction model obtained by the foregoing training. For information.
- the commodity mapping method of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, does not require manual intervention, greatly improves the commodity mapping efficiency of the cross-e-commerce platform, and, since it is no longer necessary to invest a large amount of manpower, It also saves labor costs and reduces the operating costs of the e-commerce platform.
- the model trained by the product data of the target e-commerce platform predicts the related product information of the source product, and the accuracy of the mapping can be improved.
- the present invention also proposes the following embodiment of the commodity mapping apparatus.
- FIG. 3 is a structural block diagram of a commodity mapping apparatus according to Embodiment 3 of the present invention.
- the commodity mapping apparatus 300 may include a first obtaining module 310 and a mapping module 320.
- the first obtaining module 310 is configured to obtain commodity information of the first e-commerce platform product, and record it as the first product information.
- the mapping module 320 is configured to map the first product information acquired by the first obtaining module 310 into the second product information according to the pre-trained product mapping model, wherein the second product information is the product information of the second e-commerce platform product.
- the commodity mapping apparatus 300 may further include a second acquiring module and a training module.
- the second obtaining module is configured to acquire the second item information of the set quantity or more and the search record information of the user on the second e-commerce platform;
- the training module is configured to use the second item information and the search acquired by the second acquiring module. Record information training to obtain a commodity mapping model.
- the training module may include a creation unit and a decomposition association unit.
- the search record information may include a search phrase and a user click item corresponding to the search phrase, and the creating unit is configured to use the product category of the user to click the item as the item category associated with the search phrase, and create a relationship between the search phrase and the item category. Recorded as first Associated item.
- the decomposition association unit is configured to decompose the search phrase to obtain at least one word segment, replace the search phrase with each participle, and obtain the item of association between the word segment and the product category, and record it as the second association item.
- the mapping module 320 may include a first category mapping unit.
- the first category mapping unit is configured to predict the category information in the second commodity information from the commodity title information in the first commodity information according to the second associated item obtained by the decomposition association unit.
- the first category mapping unit may include a word segment subunit, a lookup subunit, and a prediction subunit.
- the word segment sub-unit is configured to decompose the product category information and/or the product title information in the first product information into at least one word segment, and record it as a matching word segment.
- the lookup subunit is used to find a second associated item that contains a matching participle.
- the prediction subunit is configured to use the commodity category information included in the found second association item as the category information in the second commodity information.
- the mapping module 320 may include a second category mapping unit.
- the second category mapping unit is configured to directly map the product category information in the first product information to the category information in the second product information by literal matching.
- the training module may include an extracting unit, a matching unit, and a training unit.
- the extracting unit is configured to extract the item title information and the item detail information from the second item information.
- the matching unit is configured to use the product title information and the product detail information extracted by the structural attribute data of the second e-commerce platform to obtain the attribute pair label data of the second e-commerce platform in the unstructured text.
- the training unit is configured to use the attribute of the second e-commerce platform to train the annotation data to obtain an attribute pair extraction model, and the attribute extraction model is used to extract the attribute pair information from the unstructured text.
- the mapping module 320 may include a first attribute pair mapping unit.
- the first attribute pair mapping unit is configured to extract the attribute pair information of the second item information from the item title information and the item detail information of the first item information according to the attribute pair extraction model.
- the mapping module 320 may include a second attribute pair mapping unit.
- the second attribute pair mapping unit is configured to directly map the attribute pair information in the first item information to the attribute pair information in the second item information by literal matching.
- the mapping module 320 may include a title mapping unit.
- the title mapping unit is configured to convert the product title information in the first product information into the title information in the second product information according to the first text conversion model in the product mapping model.
- the mapping module 320 may include a detail mapping unit.
- the detail mapping unit is used according to The second text conversion model in the product mapping model converts the product detail information in the first product information into the detailed information in the second product information.
- the product mapping device in this embodiment can perform the product mapping method described in the foregoing embodiment of the present invention.
- the commodity mapping device of the embodiment of the invention can realize automatic mapping of goods between different e-commerce platforms, does not require manual intervention, greatly improves the product mapping efficiency of the cross-e-commerce platform, and, since it is no longer necessary to invest a large amount of manpower, It also saves labor costs and reduces the operating costs of the e-commerce platform. Further, the product mapping apparatus according to the embodiment of the present invention can predict the accuracy of the mapping by predicting the related product information of the source product using the model trained by the product data of the target e-commerce platform.
- the commodity mapping method may include the following steps:
- Step S401 acquiring barcode information of the first e-commerce platform product
- the same product has the same bar code information, so if the bar code information of the goods of the two e-commerce platforms is the same, then the two goods are the same goods. In this way, the same barcode information can be used to directly map the commodity A on one e-commerce platform to the commodity B on the other e-commerce platform having the same barcode information as the commodity A.
- Step S402 searching for a commodity having the barcode information from the second e-commerce platform, and recording the product as a matching commodity;
- step S403 the product information of the matching commodity is mapped to the commodity information of the second e-commerce platform as the commodity of the first e-commerce platform.
- the commodity mapping method of the embodiment of the invention can utilize the bar code information of the commodity to realize automatic mapping of goods between different e-commerce platforms, without manual intervention, greatly improving the commodity mapping efficiency of the cross-e-commerce platform, and A large amount of manpower is required, labor costs are saved, and the operating cost of the e-commerce platform is reduced.
- FIG. 5 is a structural block diagram of a commodity mapping apparatus according to Embodiment 5 of the present invention.
- the commodity mapping apparatus 500 may include a barcode acquisition module 510, a barcode matching module 520, and a barcode mapping module 530.
- the barcode obtaining module 510 is configured to obtain barcode information of the first e-commerce platform product.
- the barcode matching module 520 is configured to search for the commodity having the barcode information from the second e-commerce platform, and record it as a matching commodity.
- the barcode mapping module 530 is configured to map the commodity information of the matching commodity as the commodity of the first e-commerce platform to the commodity information of the second e-commerce platform.
- the merchandise mapping device in the embodiment of the present invention can perform the merchandise mapping method described in the foregoing embodiment of the present invention. A description of the application.
- the commodity mapping device of the embodiment of the invention can utilize the bar code information of the commodity to realize automatic mapping of goods between different e-commerce platforms, without manual intervention, greatly improving the commodity mapping efficiency of the cross-e-commerce platform, and A large amount of manpower is required, labor costs are saved, and the operating cost of the e-commerce platform is reduced.
- FIG. 6 is a structural block diagram of a commodity mapping device according to Embodiment 6 of the present invention.
- the commodity mapping device 600 may include a commodity mapping device 300.
- the commodity mapping device 300 may be any one of the commodity mapping devices of the embodiment shown in FIG. 1 or FIG. 2 of the present invention.
- the commodity mapping device 600 can be a server, a computer, or the like.
- the product mapping device 300 can be configured to acquire the product information of the first e-commerce platform product, and record the product information as the first product information; and map the acquired first product information to the second product information according to the pre-trained product mapping model.
- the second product information is product information of the second e-commerce platform product.
- the commodity mapping device of the embodiment of the invention includes the commodity mapping device, which can realize automatic mapping of goods between different e-commerce platforms, does not require manual intervention, greatly improves the commodity mapping efficiency of the cross-e-commerce platform, and, since it is no longer needed Investing a large amount of manpower also saves labor costs and reduces the operating costs of the e-commerce platform.
- FIG. 7 is a structural block diagram of a commodity mapping device according to Embodiment 7 of the present invention.
- the commodity mapping device 700 may include a commodity mapping device 500.
- the commodity mapping device 500 may be any one of the commodity mapping devices of the embodiment shown in FIG. 5 of the present invention.
- the commodity mapping device 700 can be a server, a computer, or the like.
- the product mapping device 500 may be configured to obtain barcode information of the first e-commerce platform product; and search for the product having the barcode information from the second e-commerce platform, and record the product as the matching product; The merchandise platform's merchandise is mapped to the merchandise information of the second e-commerce platform.
- the commodity mapping device of the embodiment of the invention includes a commodity mapping device, which can utilize the bar code information of the commodity to realize automatic mapping of goods between different e-commerce platforms, without manual intervention, and greatly improve the commodity mapping efficiency of the cross-e-commerce platform. Moreover, since it is no longer necessary to invest a large amount of manpower, it also saves labor costs and reduces the operating cost of the e-commerce platform.
- FIG. 8 is a schematic diagram of commodity mapping in an embodiment of the present invention.
- the product information (ie, the first product information) of the first e-commerce platform product is mapped to the product information (ie, the second product information) of the second e-commerce platform product after passing through the product mapping model.
- the product information ie, the first product information
- the product information ie, the second product information
- any one of the foregoing embodiments of the present invention may be employed.
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
一种商品映射方法、装置及设备。其中,商品映射方法包括:获取第一电商平台商品的商品信息,记为第一商品信息(S101);根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,所述第二商品信息为第二电商平台商品的商品信息(S102)。能够实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率。并且,所述商品映射方法,由于不再需要投入大量的人力,因此还节省了人工费用,降低了电商平台的运营成本。
Description
本申请要求2016年03月25日递交的申请号为201610180436.2、发明名称为“商品映射方法、装置及设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本发明涉及电子商务领域,尤其涉及一种商品映射方法、装置及设备。
在信息网络已广泛覆盖的当今时代,以信息网络为基础的电子商务得到了迅速发展。电子商务是指在商业贸易活动中,在因特网开放的网络环境下,基于浏览器/服务器应用方式,买卖双方不谋面地进行各种商贸活动,实现消费者的网上购物、商户之间的网上交易和在线电子支付以及各种商务活动、交易活动、金融活动和相关的综合服务活动的一种新型的商业运营模式。
电子商务的实现依赖于电子商务平台,简称电商平台。电商平台是一个为企业或个人提供网上交易洽谈的平台。企业电商平台是建立在因特网上进行商务活动的虚拟网络空间和保障商务顺利运营的管理环境,是协调、整合信息流、货物流、资金流有序、关联、高效流动的重要场所。企业、商家可以充分利用电商平台提供的网络基础设施、支付平台、安全平台、管理平台等共享资源有效地、低成本地开展自己的商业活动。
当前,电商平台正在向国际化、线下化、平台化发展。在此过程中,对于任何一个电商平台来说,一个迫切的需求是把其他电商平台的商品转换为自己电商平台的商品,然后快速发布到自己的电商平台上。由于各电商平台的商品描述格式完全不同,因此需要跨电商平台的商品映射,以便让其他电商平台的商品格式兼容于自己的电商平台。
相关技术中,跨电商平台的商品映射主要通过人工方式来进行。人工进行跨电商平台的商品映射不仅效率低下,而且由于需要投入大量的人工费用,导致电商平台的运营成本升高。
发明内容
本发明的目的在于提供一种商品映射方法、装置及设备,实现跨电商平台商品的自动映射,提高映射效率。
为实现上述目的,本发明提出了一种商品映射方法,包括:
获取第一电商平台商品的商品信息,记为第一商品信息;
根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,所述第二商品信息为第二电商平台商品的商品信息。
进一步地,上述方法还可具有以下特点,还包括:
获取设定数量以上的第二商品信息和用户在所述第二电商平台的搜索记录信息;
利用获取的第二商品信息和搜索记录信息训练得到所述商品映射模型。
进一步地,上述方法还可具有以下特点,所述利用获取的第二商品信息和搜索记录信息训练得到所述商品映射模型,包括:
所述搜索记录信息包括搜索短语和所述搜索短语对应的用户点击商品,将所述用户点击商品的商品类目作为与所述搜索短语关联的商品类目,创建搜索短语与商品类目的关联项,记为第一关联项;
对所述搜索短语进行分解,得到至少一个分词,用每一个分词替换所述搜索短语,得到分词与商品类目关联项,记为第二关联项。
进一步地,上述方法还可具有以下特点,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:
根据所述第二关联项,从所述第一商品信息中的商品标题信息预测出所述第二商品信息中的类目信息。
进一步地,上述方法还可具有以下特点,所述根据所述第二关联项,从所述第一商品信息中的商品标题信息预测出所述第二商品信息中的类目信息,包括:
将所述第一商品信息中的商品类目信息和/或商品标题信息分解为至少一个分词,记为匹配分词;
查找包含所述匹配分词的第二关联项;
将查找到的第二关联项中包含的商品类目信息作为所述第二商品信息中的类目信息。
进一步地,上述方法还可具有以下特点,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:
将所述第一商品信息中的商品类目信息通过字面匹配直接映射为所述第二商品信息中的类目信息。
进一步地,上述方法还可具有以下特点,所述利用获取的第二商品信息和搜索记录
信息训练得到所述商品映射模型,包括:
从所述第二商品信息中提取商品标题信息和商品详情信息;
用所述第二电商平台的结构化属性数据全匹配提取出的所述商品标题信息和商品详情信息,得到无结构文本中所述第二电商平台的属性对标注数据;
利用所述第二电商平台的属性对标注数据训练得到属性对抽取模型,所述属性对抽取模型用于从无结构文本中抽取出属性对信息。
进一步地,上述方法还可具有以下特点,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:
根据所述属性对抽取模型,从所述第一商品信息的商品标题信息和商品详情信息中抽取所述第二商品信息的属性对信息。
进一步地,上述方法还可具有以下特点,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:
将所述第一商品信息中的属性对信息通过字面匹配直接映射为所述第二商品信息中的属性对信息。
进一步地,上述方法还可具有以下特点,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:
根据所述商品映射模型中的第一文本转换模型,将所述第一商品信息中的商品标题信息转换为所述第二商品信息中的标题信息。
进一步地,上述方法还可具有以下特点,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:
根据所述商品映射模型中的第二文本转换模型,将所述第一商品信息中的商品详情信息转换为所述第二商品信息中的详情信息。
本发明实施例的商品映射方法,能够实现不同电商平台之间商品的自动映射,提高了跨电商平台的商品映射效率,降低了电商平台的运营成本。
为实现上述目的,本发明还提出了一种商品映射装置,包括:
第一获取模块,用于获取第一电商平台商品的商品信息,记为第一商品信息;
映射模块,用于根据预先训练好的商品映射模型,将所述第一获取模块获取的第一商品信息映射为第二商品信息,所述第二商品信息为第二电商平台商品的商品信息。
进一步地,上述装置还可具有以下特点,还包括:
第二获取模块,用于获取设定数量以上的第二商品信息和用户在所述第二电商平台
的搜索记录信息;
训练模块,用于利用所述第二获取模块获取的第二商品信息和搜索记录信息训练得到所述商品映射模型。
进一步地,上述装置还可具有以下特点,所述训练模块包括:
创建单元,所述搜索记录信息包括搜索短语和所述搜索短语对应的用户点击商品,创建单元用于将所述用户点击商品的商品类目作为与所述搜索短语关联的商品类目,创建搜索短语与商品类目的关联项,记为第一关联项;
分解关联单元,用于对所述搜索短语进行分解,得到至少一个分词,用每一个分词替换所述搜索短语,得到分词与商品类目关联项,记为第二关联项。
进一步地,上述装置还可具有以下特点,所述映射模块包括:
第一类目映射单元,用于根据所述分解关联单元得到的第二关联项,从所述第一商品信息中的商品标题信息预测出所述第二商品信息中的类目信息。
进一步地,上述装置还可具有以下特点,所述第一类目映射单元包括:
分词子单元,用于将所述第一商品信息中的商品类目信息和/或商品标题信息分解为至少一个分词,记为匹配分词;
查找子单元,用于查找包含所述匹配分词的第二关联项;
预测子单元,用于将查找到的第二关联项中包含的商品类目信息作为所述第二商品信息中的类目信息。
进一步地,上述装置还可具有以下特点,所述映射模块包括:
第二类目映射单元,用于将所述第一商品信息中的商品类目信息通过字面匹配直接映射为所述第二商品信息中的类目信息。
进一步地,上述装置还可具有以下特点,所述训练模块包括:
提取单元,用于从所述第二商品信息中提取商品标题信息和商品详情信息;
匹配单元,用于用所述第二电商平台的结构化属性数据全匹配提取出的所述商品标题信息和商品详情信息,得到无结构文本中所述第二电商平台的属性对标注数据;
训练单元,用于利用所述第二电商平台的属性对标注数据训练得到属性对抽取模型,所述属性对抽取模型用于从无结构文本中抽取出属性对信息。
进一步地,上述装置还可具有以下特点,所述映射模块包括:
第一属性对映射单元,用于根据所述属性对抽取模型,从所述第一商品信息的商品标题信息和商品详情信息中抽取所述第二商品信息的属性对信息。
进一步地,上述装置还可具有以下特点,所述映射模块包括:
第二属性对映射单元,用于将所述第一商品信息中的属性对信息通过字面匹配直接映射为所述第二商品信息中的属性对信息。
进一步地,上述装置还可具有以下特点,所述映射模块包括:
标题映射单元,用于根据所述商品映射模型中的第一文本转换模型,将所述第一商品信息中的商品标题信息转换为所述第二商品信息中的标题信息。
进一步地,上述装置还可具有以下特点,所述映射模块包括:
详情映射单元,用于根据所述商品映射模型中的第二文本转换模型,将所述第一商品信息中的商品详情信息转换为所述第二商品信息中的详情信息。
本发明实施例的商品映射装置,能够实现不同电商平台之间商品的自动映射,提高了跨电商平台的商品映射效率,降低了电商平台的运营成本。
为实现上述目的,本发明还提出了一种商品映射方法,包括:
获取第一电商平台商品的条码信息;
从第二电商平台查找具有所述条码信息的商品,记为匹配商品;
将所述匹配商品的商品信息作为所述第一电商平台的商品映射到所述第二电商平台的商品信息。
本发明实施例的商品映射方法,能够实现不同电商平台之间商品的自动映射,提高了跨电商平台的商品映射效率,降低了电商平台的运营成本。
为实现上述目的,本发明还提出了一种商品映射装置,包括:
条码获取模块,用于获取第一电商平台商品的条码信息;
条码匹配模块,用于从第二电商平台查找具有所述条码信息的商品,记为匹配商品;
条码映射模块,用于将所述匹配商品的商品信息作为所述第一电商平台的商品映射到所述第二电商平台的商品信息。
本发明实施例的商品映射装置,能够实现不同电商平台之间商品的自动映射,提高了跨电商平台的商品映射效率,降低了电商平台的运营成本。
为实现上述目的,本发明还提出了一种商品映射设备,包括前述任一项所述的商品映射装置。
本发明实施例的商品映射设备,能够实现不同电商平台之间商品的自动映射,提高了跨电商平台的商品映射效率,降低了电商平台的运营成本。
图1为本发明实施例一中商品映射方法的流程图;
图2为本发明实施例二中商品映射方法的流程图;
图3为本发明实施例三中商品映射装置的结构框图;
图4为本发明实施例四中商品映射方法的流程图;
图5为本发明实施例五中商品映射装置的结构框图;
图6为本发明实施例六中商品映射设备的结构框图;
图7为本发明实施例七中商品映射设备的结构框图;
图8为本发明实施例中的商品映射示意图。
以下结合附图对本发明的原理和特征进行描述,所举实施例只用于解释本发明,并非用于限定本发明的范围。对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,根据本发明精神所获得的所有实施例,都属于本发明的保护范围。
应当理解,尽管在本发明实施例中可能采用术语第一、第二、第三等来描述XXX,但这些XXX不应限于这些术语。这些术语仅用来将XXX彼此区分开。例如,在不脱离本发明实施例范围的情况下,第一XXX也可以被称为第二XXX,类似地,第二XXX也可以被称为第一XXX。
图1为本发明实施例一中商品映射方法的流程图。如图1所示,本实施例中,商品映射方法可以包括如下步骤:
步骤S101,获取第一电商平台商品的商品信息,记为第一商品信息;
其中,商品信息可以包括商品标题信息、商品类目信息、商品属性对信息、商品详情信息等等。在电商平台的商品展位上,可以获取该展位商品的具体商品信息。虽然各电商平台的商品描述格式完全不同,但是一般都包含商品标题信息、商品类目信息、商品详情信息这些最基本的信息,一些电商平台的商品信息中还包括商品属性对信息。
其中,第一电商平台是要进行商品映射的平台,本文中称为来源电商平台。下面的第二电商平台是商品要映射到的目标平台,本文中称为目标电商平台。
步骤S102,根据预先训练好的商品映射模型,将第一商品信息映射为第二商品信息,其中,第二商品信息为第二电商平台商品的商品信息。
商品映射模型是根据第二电商平台的商品数据训练得到的。依据商品映射模型,可
以将来源电商平台的商品信息映射到目标电商平台。
商品映射模型的具体训练方式的示例将在下面的实施例二中进行说明。
其中,商品的映射可以包括类目、属性对、标题、详情等几方面的映射。
对于商品类目的映射,本文中给出两种示例映射方式。商品类目的示例映射方式之一可以是:将第一商品信息中的商品类目信息通过字面匹配直接映射为第二商品信息中的类目信息。其中,字面匹配是指通过字符串字面的编辑距离所进行的匹配。比如,第一商品信息中的商品类目信息为“冬季连衣裙”,第二商品信息中的类目信息为“连衣裙”,则通过字面匹配,将第一商品信息中的“冬季连衣裙”直接映射为第二商品信息中的“连衣裙”。商品类目的示例映射方式之二将在下面的实施例二中进行说明。
对于商品属性对的映射,本文中也给出两种示例映射方式。商品属性对的示例映射方式之一可以是:将第一商品信息中的属性对信息通过字面匹配直接映射为第二商品信息中的属性对信息。例如,第一商品信息中的属性对信息为“粉色”,第二商品信息中的属性对信息为“粉红色”,通过字面匹配,直接将第一商品信息中的“粉色”映射为第二商品信息中的“粉红色”。商品属性对的示例映射方式之二将在下面的实施例二中进行说明。
对于商品标题的映射,映射方式可以是:根据商品映射模型中的第一文本转换模型,将第一商品信息中的商品标题信息转换为第二商品信息中的标题信息。其中,第一文本转换模型可以是机器翻译模型,比如将繁体字的商品标题信息翻译为简体字的商品标题信息。
对于商品详情的映射,映射方式可以是:根据商品映射模型中的第二文本转换模型,将第一商品信息中的商品详情信息转换为第二商品信息中的详情信息。其中,第二文本转换模型可以是机器翻译模型,比如将英文的商品标题信息翻译为简体中文的商品标题信息。其中,第二文本转换模型与前述的第一文本转换模型可以相同。
本发明实施例的商品映射方法,能够实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率。并且,本发明实施例的商品映射方法,由于不再需要投入大量的人力,因此还节省了人工费用,降低了电商平台的运营成本。
图2为本发明实施例二中商品映射方法的流程图。如图2所示,本实施例中,商品映射方法可以包括如下步骤:
步骤S201,获取设定数量以上的第二商品信息和用户在第二电商平台的搜索记录信
息;
其中,第二商品信息为第二电商平台商品的商品信息。
步骤S202,利用获取的第二商品信息和搜索记录信息训练得到商品映射模型;
在本发明实施例中,利用获取的第二商品信息和搜索记录信息可以训练得到商品类目的映射模型、商品属性对的映射模型等。下面分别举例说明。
(a)利用获取的第二商品信息和搜索记录信息可以训练得到商品类目的映射模型。
在本发明实施例中,利用获取的第二商品信息和搜索记录信息训练得到商品映射模型可以包括:
搜索记录信息可以包括搜索短语和该搜索短语对应的用户点击商品,将用户点击商品的商品类目作为与搜索短语关联的商品类目,创建搜索短语与商品类目的关联项,记为第一关联项;
对搜索短语进行分解,得到至少一个分词,用每一个分词替换搜索短语,得到分词与商品类目关联项,记为第二关联项。
这里,第二关联项即为训练得到的商品类目的映射模型,可以将第二关联项称为类目预测模型。
用户在电商平台输入的搜索短语一般是用户对要找的商品的特征的描述,该搜索短语之后的点击数据一般就是用户要找的商品。因此,用户点击的商品应该是与用户输入的搜索短语有关联的商品。依据用户的搜索短语和该搜索短语对应的用户点击商品训练类目预测模型,可以提高类目预测的准确率。
下面举例说明第二关联项的训练过程和原理。
例如,用户在第二电商平台输入搜索短语“冬季连衣裙”,在搜索后该用户点击了商品a、商品b和商品c,商品a、商品b和商品c的商品类目分别为类目1、类目2、类目3。则可以创建3个第一关联项,即:
冬季连衣裙---类目1;
冬季连衣裙---类目2;
冬季连衣裙---类目3。
然后,将搜索短语“冬季连衣裙”分解为两个分词“冬季”和“连衣裙”,用分词“冬季”替换搜索短语“冬季连衣裙”,得到3个第二关联项,即:
冬季---类目1;
冬季---类目2;
冬季---类目3。
用分词“连衣裙”替换搜索短语“冬季连衣裙”,也得到3个第二关联项,即:
连衣裙---类目1;
连衣裙---类目2;
连衣裙---类目3。
(b)利用获取的第二商品信息和搜索记录信息可以训练得到商品属性对的映射模型。
在本发明实施例中,利用获取的第二商品信息和搜索记录信息训练得到商品映射模型可以包括:
从第二商品信息中提取商品标题信息和商品详情信息;
用第二电商平台的结构化属性数据全匹配提取出的商品标题信息和商品详情信息,得到无结构文本中第二电商平台的属性对标注数据;
利用第二电商平台的属性对标注数据训练得到属性对抽取模型,该属性对抽取模型用于从无结构文本中抽取出属性对信息。
其中,结构化属性数据是指电商平台对商品的描述结构。例如,在淘宝网搜到的一款连衣裙商品中,该商品的“宝贝详情”中有如下内容:货号、风格、组合形式、裙长、袖长、领型、袖型、腰型等。这些内容就是淘宝网这个电商平台的结构化属性数据。
这里,属性对抽取模型即为训练得到的商品属性对的映射模型。
例如,第二商品信息中的商品标题信息为:海青蓝2016春装新款女装复古印花通勤A字裙圆领中袖宽松连衣裙,标注这个标题中的属性对,比如颜色:海清蓝,上市年份:2016,风格:复古,图案:印花,领型:圆领,袖长:中袖。积累大量的这种标注数据,用HMM(隐马尔科夫模型)或者CRF(条件随机场)等模型训练就可以得到属性对抽取模型。
步骤S203,获取第一电商平台商品的商品信息,记为第一商品信息;
步骤S204,根据预先训练好的商品映射模型,将第一商品信息映射为第二商品信息。
此处结合步骤S202所列举的训练示例来说明商品类目的示例映射方式之二和商品属性对的示例映射方式之二。
在本发明实施例中,商品类目的示例映射方式之二可以是:根据前述训练得到的第二关联项,从第一商品信息中的商品标题信息预测出第二商品信息中的类目信息。
在本发明实施例中,根据第二关联项,从第一商品信息中的商品标题信息预测出第
二商品信息中的类目信息,可以包括:将第一商品信息中的商品类目信息和/或商品标题信息分解为至少一个分词,记为匹配分词;查找包含该匹配分词的第二关联项;将查找到的第二关联项中包含的商品类目信息作为第二商品信息中的类目信息。
例如,来源商品的商品类目为“长袖连衣裙”,将“长袖连衣裙”分解为两个分词“长袖”和“连衣裙”,“长袖”和“连衣裙”为匹配分词。前述例子中有3个第二关联项(连衣裙---类目1;连衣裙---类目2;连衣裙---类目3)中包含“连衣裙”,因此,预测目标商品的商品类目为类目1、类目2、类目3。即这三个类目都是目标商品的商品类目。
在本发明实施例中,商品属性对的示例映射方式之二可以是:根据前述训练得到的属性对抽取模型,从第一商品信息的商品标题信息和商品详情信息中抽取第二商品信息的属性对信息。
本发明实施例的商品映射方法,能够实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率,并且,由于不再需要投入大量的人力,还节省了人工费用,降低了电商平台的运营成本。此外,本发明实施例的商品映射方法,利用目标电商平台的商品数据训练出的模型预测来源商品的相关商品信息,能够提高映射的准确度。
为了实施本发明前述实施例中的商品映射方法,本发明还提出了下面的商品映射装置实施例。
图3为本发明实施例三中商品映射装置的结构框图。如图3所示,本实施例中,商品映射装置300可以包括第一获取模块310和映射模块320。其中,第一获取模块310用于获取第一电商平台商品的商品信息,记为第一商品信息。映射模块320用于根据预先训练好的商品映射模型,将第一获取模块310获取的第一商品信息映射为第二商品信息,其中,第二商品信息为第二电商平台商品的商品信息。
在本发明实施例中,商品映射装置300还可以包括第二获取模块和训练模块。其中,第二获取模块用于获取设定数量以上的第二商品信息和用户在第二电商平台的搜索记录信息;训练模块用于利用所述第二获取模块获取的第二商品信息和搜索记录信息训练得到商品映射模型。
在本发明实施例中,训练模块可以包括创建单元和分解关联单元。搜索记录信息可以包括搜索短语和该搜索短语对应的用户点击商品,创建单元用于将用户点击商品的商品类目作为与搜索短语关联的商品类目,创建搜索短语与商品类目的关联项,记为第一
关联项。分解关联单元用于对搜索短语进行分解,得到至少一个分词,用每一个分词替换搜索短语,得到分词与商品类目关联项,记为第二关联项。
在本发明实施例中,在训练模块包括创建单元和分解关联单元的情况下,映射模块320可以包括第一类目映射单元。第一类目映射单元用于根据分解关联单元得到的第二关联项,从第一商品信息中的商品标题信息预测出第二商品信息中的类目信息。
在本发明实施例中,第一类目映射单元可以包括分词子单元、查找子单元和预测子单元。其中,分词子单元用于将第一商品信息中的商品类目信息和/或商品标题信息分解为至少一个分词,记为匹配分词。查找子单元用于查找包含匹配分词的第二关联项。预测子单元用于将查找到的第二关联项中包含的商品类目信息作为第二商品信息中的类目信息。
在本发明实施例中,映射模块320可以包括第二类目映射单元。第二类目映射单元用于将第一商品信息中的商品类目信息通过字面匹配直接映射为第二商品信息中的类目信息。
在本发明实施例中,训练模块可以包括提取单元、匹配单元和训练单元。提取单元用于从第二商品信息中提取商品标题信息和商品详情信息。匹配单元用于用第二电商平台的结构化属性数据全匹配提取出的商品标题信息和商品详情信息,得到无结构文本中所述第二电商平台的属性对标注数据。训练单元用于利用第二电商平台的属性对标注数据训练得到属性对抽取模型,该属性对抽取模型用于从无结构文本中抽取出属性对信息。
在本发明实施例中,在训练模块包括提取单元、匹配单元和训练单元的情况下,映射模块320可以包括第一属性对映射单元。第一属性对映射单元用于根据属性对抽取模型,从第一商品信息的商品标题信息和商品详情信息中抽取第二商品信息的属性对信息。
在本发明实施例中,映射模块320可以包括第二属性对映射单元。第二属性对映射单元用于将第一商品信息中的属性对信息通过字面匹配直接映射为第二商品信息中的属性对信息。
在本发明实施例中,映射模块320可以包括标题映射单元。标题映射单元用于根据商品映射模型中的第一文本转换模型,将第一商品信息中的商品标题信息转换为第二商品信息中的标题信息。
在本发明实施例中,映射模块320可以包括详情映射单元。详情映射单元用于根据
商品映射模型中的第二文本转换模型,将第一商品信息中的商品详情信息转换为第二商品信息中的详情信息。
由于本实施例中的商品映射装置能够执行本发明前述实施例中所述的商品映射方法,本实施例未详细描述的部分,可参考对本发明前述商品映射方法实施例的相关说明。
本发明实施例的商品映射装置,能够实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率,并且,由于不再需要投入大量的人力,还节省了人工费用,降低了电商平台的运营成本。此外,本发明实施例的商品映射装置,利用目标电商平台的商品数据训练出的模型预测来源商品的相关商品信息,能够提高映射的准确度。
图4为本发明实施例四中商品映射方法的流程图。如图4所示,本实施例中,商品映射方法可以包括如下步骤:
步骤S401,获取第一电商平台商品的条码信息;
同一种商品具有相同的条码信息,因此,如果两个电商平台的商品的条码信息相同,那么该两个商品就是同一种商品。这样,就可以利用相同条码信息,将一个电商平台上的商品A直接映射为另一个电商平台上与商品A具有相同条码信息的商品B。
步骤S402,从第二电商平台查找具有该条码信息的商品,记为匹配商品;
步骤S403,将该匹配商品的商品信息作为第一电商平台的商品映射到第二电商平台的商品信息。
本发明实施例的商品映射方法,能够利用商品的条码信息,实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率,并且,由于不再需要投入大量的人力,还节省了人工费用,降低了电商平台的运营成本。
为了实施本发明图4所示实施例中的商品映射方法,本发明还提出了图5所示的商品映射装置实施例。图5为本发明实施例五中商品映射装置的结构框图。如图5所示,本实施例中,商品映射装置500可以包括条码获取模块510、条码匹配模块520和条码映射模块530。条码获取模块510用于获取第一电商平台商品的条码信息。条码匹配模块520用于从第二电商平台查找具有该条码信息的商品,记为匹配商品。条码映射模块530用于将该匹配商品的商品信息作为第一电商平台的商品映射到第二电商平台的商品信息。
由于本实施例中的商品映射装置能够执行本发明前述图4所示实施例中所述的商品映射方法,本实施例未详细描述的部分,可参考对本发明前述图4所示商品映射方法实
施例的相关说明。
本发明实施例的商品映射装置,能够利用商品的条码信息,实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率,并且,由于不再需要投入大量的人力,还节省了人工费用,降低了电商平台的运营成本。
图6为本发明实施例六中商品映射设备的结构框图。如图6所示,本实施例中,商品映射设备600可以包括商品映射装置300。其中,商品映射装置300可以是本发明前述图1或图2所示实施例中的任一种商品映射装置。
其中,商品映射设备600可以是服务器、计算机等。
其中,商品映射装置300可以用于获取第一电商平台商品的商品信息,记为第一商品信息;根据预先训练好的商品映射模型,将获取的第一商品信息映射为第二商品信息,其中,第二商品信息为第二电商平台商品的商品信息。
本发明实施例的商品映射设备中包括商品映射装置,能够实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率,并且,由于不再需要投入大量的人力,还节省了人工费用,降低了电商平台的运营成本。
图7为本发明实施例七中商品映射设备的结构框图。如图7所示,本实施例中,商品映射设备700可以包括商品映射装置500。其中,商品映射装置500可以是本发明前述图5所示实施例中的任一种商品映射装置。
其中,商品映射设备700可以是服务器、计算机等。
其中,商品映射装置500可以用于获取第一电商平台商品的条码信息;从第二电商平台查找具有该条码信息的商品,记为匹配商品;将该匹配商品的商品信息作为第一电商平台的商品映射到第二电商平台的商品信息。
本发明实施例的商品映射设备中包括商品映射装置,能够利用商品的条码信息,实现不同电商平台之间商品的自动映射,不需要人工干预,大大提高了跨电商平台的商品映射效率,并且,由于不再需要投入大量的人力,还节省了人工费用,降低了电商平台的运营成本。
下面通过一个典型的商品映射示意图来说明商品映射过程。图8为本发明实施例中的商品映射示意图。如图8所示,第一电商平台商品的商品信息(即第一商品信息)经过商品映射模型后,被映射为第二电商平台商品的商品信息(即第二商品信息)。在第一商品信息映射为第二商品信息的过程中,可以采用本发明前述实施例中的任意一种商品映射方法。
以上所述仅为本发明的较佳实施例,并不用以限制本发明,凡在本发明的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本发明的保护范围之内。
Claims (26)
- 一种商品映射方法,其特征在于,包括:获取第一电商平台商品的商品信息,记为第一商品信息;根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,所述第二商品信息为第二电商平台商品的商品信息。
- 根据权利要求1所述的商品映射方法,其特征在于,还包括:获取设定数量以上的第二商品信息和用户在所述第二电商平台的搜索记录信息;利用获取的第二商品信息和搜索记录信息训练得到所述商品映射模型。
- 根据权利要求2所述的商品映射方法,其特征在于,所述利用获取的第二商品信息和搜索记录信息训练得到所述商品映射模型,包括:所述搜索记录信息包括搜索短语和所述搜索短语对应的用户点击商品,将所述用户点击商品的商品类目作为与所述搜索短语关联的商品类目,创建搜索短语与商品类目的关联项,记为第一关联项;对所述搜索短语进行分解,得到至少一个分词,用每一个分词替换所述搜索短语,得到分词与商品类目关联项,记为第二关联项。
- 根据权利要求3所述的商品映射方法,其特征在于,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:根据所述第二关联项,从所述第一商品信息中的商品标题信息预测出所述第二商品信息中的类目信息。
- 根据权利要求4所述的商品映射方法,其特征在于,所述根据所述第二关联项,从所述第一商品信息中的商品标题信息预测出所述第二商品信息中的类目信息,包括:将所述第一商品信息中的商品类目信息和/或商品标题信息分解为至少一个分词,记为匹配分词;查找包含所述匹配分词的第二关联项;将查找到的第二关联项中包含的商品类目信息作为所述第二商品信息中的类目信息。
- 根据权利要求1所述的商品映射方法,其特征在于,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:将所述第一商品信息中的商品类目信息通过字面匹配直接映射为所述第二商品信息中的类目信息。
- 根据权利要求2所述的商品映射方法,其特征在于,所述利用获取的第二商品信息和搜索记录信息训练得到所述商品映射模型,包括:从所述第二商品信息中提取商品标题信息和商品详情信息;用所述第二电商平台的结构化属性数据全匹配提取出的所述商品标题信息和商品详情信息,得到无结构文本中所述第二电商平台的属性对标注数据;利用所述第二电商平台的属性对标注数据训练得到属性对抽取模型,所述属性对抽取模型用于从无结构文本中抽取出属性对信息。
- 根据权利要求7所述的商品映射方法,其特征在于,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:根据所述属性对抽取模型,从所述第一商品信息的商品标题信息和商品详情信息中抽取所述第二商品信息的属性对信息。
- 根据权利要求1所述的商品映射方法,其特征在于,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:将所述第一商品信息中的属性对信息通过字面匹配直接映射为所述第二商品信息中的属性对信息。
- 根据权利要求1所述的商品映射方法,其特征在于,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:根据所述商品映射模型中的第一文本转换模型,将所述第一商品信息中的商品标题信息转换为所述第二商品信息中的标题信息。
- 根据权利要求1所述的商品映射方法,其特征在于,所述根据预先训练好的商品映射模型,将所述第一商品信息映射为第二商品信息,包括:根据所述商品映射模型中的第二文本转换模型,将所述第一商品信息中的商品详情信息转换为所述第二商品信息中的详情信息。
- 一种商品映射装置,其特征在于,包括:第一获取模块,用于获取第一电商平台商品的商品信息,记为第一商品信息;映射模块,用于根据预先训练好的商品映射模型,将所述第一获取模块获取的第一商品信息映射为第二商品信息,所述第二商品信息为第二电商平台商品的商品信息。
- 根据权利要求12所述的商品映射装置,其特征在于,还包括:第二获取模块,用于获取设定数量以上的第二商品信息和用户在所述第二电商平台的搜索记录信息;训练模块,用于利用所述第二获取模块获取的第二商品信息和搜索记录信息训练得到所述商品映射模型。
- 根据权利要求13所述的商品映射装置,其特征在于,所述训练模块包括:创建单元,所述搜索记录信息包括搜索短语和所述搜索短语对应的用户点击商品,创建单元用于将所述用户点击商品的商品类目作为与所述搜索短语关联的商品类目,创建搜索短语与商品类目的关联项,记为第一关联项;分解关联单元,用于对所述搜索短语进行分解,得到至少一个分词,用每一个分词替换所述搜索短语,得到分词与商品类目关联项,记为第二关联项。
- 根据权利要求14所述的商品映射装置,其特征在于,所述映射模块包括:第一类目映射单元,用于根据所述分解关联单元得到的第二关联项,从所述第一商品信息中的商品标题信息预测出所述第二商品信息中的类目信息。
- 根据权利要求15所述的商品映射装置,其特征在于,所述第一类目映射单元包括:分词子单元,用于将所述第一商品信息中的商品类目信息和/或商品标题信息分解为至少一个分词,记为匹配分词;查找子单元,用于查找包含所述匹配分词的第二关联项;预测子单元,用于将查找到的第二关联项中包含的商品类目信息作为所述第二商品信息中的类目信息。
- 根据权利要求12所述的商品映射装置,其特征在于,所述映射模块包括:第二类目映射单元,用于将所述第一商品信息中的商品类目信息通过字面匹配直接映射为所述第二商品信息中的类目信息。
- 根据权利要求13所述的商品映射装置,其特征在于,所述训练模块包括:提取单元,用于从所述第二商品信息中提取商品标题信息和商品详情信息;匹配单元,用于用所述第二电商平台的结构化属性数据全匹配提取出的所述商品标题信息和商品详情信息,得到无结构文本中所述第二电商平台的属性对标注数据;训练单元,用于利用所述第二电商平台的属性对标注数据训练得到属性对抽取模型,所述属性对抽取模型用于从无结构文本中抽取出属性对信息。
- 根据权利要求18所述的商品映射装置,其特征在于,所述映射模块包括:第一属性对映射单元,用于根据所述属性对抽取模型,从所述第一商品信息的商品标题信息和商品详情信息中抽取所述第二商品信息的属性对信息。
- 根据权利要求12所述的商品映射装置,其特征在于,所述映射模块包括:第二属性对映射单元,用于将所述第一商品信息中的属性对信息通过字面匹配直接映射为所述第二商品信息中的属性对信息。
- 根据权利要求12所述的商品映射装置,其特征在于,所述映射模块包括:标题映射单元,用于根据所述商品映射模型中的第一文本转换模型,将所述第一商品信息中的商品标题信息转换为所述第二商品信息中的标题信息。
- 根据权利要求12所述的商品映射装置,其特征在于,所述映射模块包括:详情映射单元,用于根据所述商品映射模型中的第二文本转换模型,将所述第一商品信息中的商品详情信息转换为所述第二商品信息中的详情信息。
- 一种商品映射方法,其特征在于,包括:获取第一电商平台商品的条码信息;从第二电商平台查找具有所述条码信息的商品,记为匹配商品;将所述匹配商品的商品信息作为所述第一电商平台的商品映射到所述第二电商平台的商品信息。
- 一种商品映射装置,其特征在于,包括:条码获取模块,用于获取第一电商平台商品的条码信息;条码匹配模块,用于从第二电商平台查找具有所述条码信息的商品,记为匹配商品;条码映射模块,用于将所述匹配商品的商品信息作为所述第一电商平台的商品映射到所述第二电商平台的商品信息。
- 一种商品映射设备,其特征在于,包括权利要求12至22任一项所述的商品映射装置。
- 一种商品映射设备,其特征在于,包括权利要求24所述的商品映射装置。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610180436.2 | 2016-03-25 | ||
| CN201610180436.2A CN107230123A (zh) | 2016-03-25 | 2016-03-25 | 商品映射方法、装置及设备 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017162074A1 true WO2017162074A1 (zh) | 2017-09-28 |
Family
ID=59899246
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2017/076708 Ceased WO2017162074A1 (zh) | 2016-03-25 | 2017-03-15 | 商品映射方法、装置及设备 |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN107230123A (zh) |
| TW (1) | TW201734913A (zh) |
| WO (1) | WO2017162074A1 (zh) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665358A (zh) * | 2018-08-03 | 2018-10-16 | 北京引领海逛科技有限公司 | 产品信息快速匹配多个平台的方法和装置 |
| CN111209725A (zh) * | 2018-11-19 | 2020-05-29 | 阿里巴巴集团控股有限公司 | 一种文本信息生成方法、装置以及计算设备 |
| CN111353838A (zh) * | 2018-12-21 | 2020-06-30 | 北京京东尚科信息技术有限公司 | 自动化校验商品类目的方法和装置 |
| CN112308059A (zh) * | 2020-10-29 | 2021-02-02 | 中智关爱通(上海)科技股份有限公司 | 用于电子商务的商品信息翻译方法、系统和可读存储介质 |
| CN113298609A (zh) * | 2021-01-27 | 2021-08-24 | 阿里巴巴集团控股有限公司 | 对象识别码处理方法、对象发布方法、装置、设备及介质 |
| CN113869986A (zh) * | 2021-09-30 | 2021-12-31 | 广州新丝路信息科技有限公司 | 一种跨渠道商品类目及商品信息的复制方法及装置 |
| WO2025113090A1 (zh) * | 2023-11-30 | 2025-06-05 | 杭州阿里巴巴海外互联网产业有限公司 | 商品结构化信息的确定方法、装置及设备 |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665349A (zh) * | 2018-05-09 | 2018-10-16 | 商派软件有限公司 | 一种商品全渠道铺货方法及系统 |
| CN109242601A (zh) * | 2018-08-08 | 2019-01-18 | 阿里巴巴集团控股有限公司 | 商品信息查询方法和系统 |
| CN109542916A (zh) * | 2018-10-15 | 2019-03-29 | 平安科技(深圳)有限公司 | 平台商品入驻方法、装置、计算机设备及存储介质 |
| CN110738553A (zh) * | 2019-10-18 | 2020-01-31 | 深圳市比量科技传媒有限公司 | 一种将不同商城的商品链接进行相互映射的方法及系统 |
| CN112990995A (zh) * | 2019-12-13 | 2021-06-18 | 阿里巴巴集团控股有限公司 | 商品对象的发布方法、装置及计算机系统 |
| CN114549121A (zh) * | 2022-01-29 | 2022-05-27 | 北京元年科技股份有限公司 | 商品的智能识别方法、装置、设备及计算机可读存储介质 |
| CN118503206A (zh) * | 2024-04-30 | 2024-08-16 | 杭州阿里巴巴海外互联网产业有限公司 | 商品信息的搬迁方法、装置、设备及计算机程序产品 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1753020A (zh) * | 2005-11-08 | 2006-03-29 | 北京基福信源信息技术有限公司 | 应用于互联网电子交易整合的系统及其方法 |
| CN104063802A (zh) * | 2013-03-19 | 2014-09-24 | 阿里巴巴集团控股有限公司 | 商品信息处理方法、装置及系统 |
| CN104063523A (zh) * | 2014-07-21 | 2014-09-24 | 焦点科技股份有限公司 | 一种电子商务搜索评分与排名的方法及系统 |
| CN104657396A (zh) * | 2013-11-25 | 2015-05-27 | 腾讯科技(深圳)有限公司 | 数据迁移方法及装置 |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1466087A (zh) * | 2002-06-28 | 2004-01-07 | 明日工作室股份有限公司 | 具有扫描功能的通讯装置 |
| CN101206752A (zh) * | 2007-12-25 | 2008-06-25 | 北京科文书业信息技术有限公司 | 电子商务网站相关商品推荐系统及其方法 |
| CN103310343A (zh) * | 2012-03-15 | 2013-09-18 | 阿里巴巴集团控股有限公司 | 商品信息发布方法和装置 |
| CN103425658B (zh) * | 2012-05-15 | 2017-06-09 | 阿里巴巴集团控股有限公司 | 商品信息的存储、搜索方法与移动终端和服务器 |
-
2016
- 2016-03-25 CN CN201610180436.2A patent/CN107230123A/zh active Pending
-
2017
- 2017-02-17 TW TW106105353A patent/TW201734913A/zh unknown
- 2017-03-15 WO PCT/CN2017/076708 patent/WO2017162074A1/zh not_active Ceased
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1753020A (zh) * | 2005-11-08 | 2006-03-29 | 北京基福信源信息技术有限公司 | 应用于互联网电子交易整合的系统及其方法 |
| CN104063802A (zh) * | 2013-03-19 | 2014-09-24 | 阿里巴巴集团控股有限公司 | 商品信息处理方法、装置及系统 |
| CN104657396A (zh) * | 2013-11-25 | 2015-05-27 | 腾讯科技(深圳)有限公司 | 数据迁移方法及装置 |
| CN104063523A (zh) * | 2014-07-21 | 2014-09-24 | 焦点科技股份有限公司 | 一种电子商务搜索评分与排名的方法及系统 |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665358A (zh) * | 2018-08-03 | 2018-10-16 | 北京引领海逛科技有限公司 | 产品信息快速匹配多个平台的方法和装置 |
| CN111209725A (zh) * | 2018-11-19 | 2020-05-29 | 阿里巴巴集团控股有限公司 | 一种文本信息生成方法、装置以及计算设备 |
| CN111209725B (zh) * | 2018-11-19 | 2023-04-25 | 阿里巴巴集团控股有限公司 | 一种文本信息生成方法、装置以及计算设备 |
| CN111353838A (zh) * | 2018-12-21 | 2020-06-30 | 北京京东尚科信息技术有限公司 | 自动化校验商品类目的方法和装置 |
| CN112308059A (zh) * | 2020-10-29 | 2021-02-02 | 中智关爱通(上海)科技股份有限公司 | 用于电子商务的商品信息翻译方法、系统和可读存储介质 |
| CN113298609A (zh) * | 2021-01-27 | 2021-08-24 | 阿里巴巴集团控股有限公司 | 对象识别码处理方法、对象发布方法、装置、设备及介质 |
| WO2022161231A1 (zh) * | 2021-01-27 | 2022-08-04 | 阿里巴巴集团控股有限公司 | 对象识别码处理方法、对象发布方法、装置、设备及介质 |
| CN113869986A (zh) * | 2021-09-30 | 2021-12-31 | 广州新丝路信息科技有限公司 | 一种跨渠道商品类目及商品信息的复制方法及装置 |
| WO2025113090A1 (zh) * | 2023-11-30 | 2025-06-05 | 杭州阿里巴巴海外互联网产业有限公司 | 商品结构化信息的确定方法、装置及设备 |
Also Published As
| Publication number | Publication date |
|---|---|
| TW201734913A (zh) | 2017-10-01 |
| CN107230123A (zh) | 2017-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017162074A1 (zh) | 商品映射方法、装置及设备 | |
| US10896392B2 (en) | Methods and systems for generating supply chain representations | |
| US20200210396A1 (en) | Image and Text Data Hierarchical Classifiers | |
| CN106776897B (zh) | 一种用户画像标签确定方法及装置 | |
| CN108664477B (zh) | 交易信息多语机器翻译子系统的翻译方法 | |
| US9892384B2 (en) | Extracting product purchase information from electronic messages | |
| CN107292365B (zh) | 商品标签的绑定方法、装置、设备及计算机可读存储介质 | |
| US9875486B2 (en) | Extracting product purchase information from electronic messages | |
| Montenegro et al. | Hybridization of the Kano model and business model canvas: aeronautical and metalworking industry in Bogota, Colombia | |
| CN107832338B (zh) | 一种识别核心产品词的方法和系统 | |
| JP2022133264A (ja) | 電子メッセージのフィルタリング | |
| CN111831818B (zh) | 文本信息识别方法、装置及存储介质 | |
| US20230214592A1 (en) | Methods and apparatus to generate textual data using machine learning processes | |
| CN104133913A (zh) | 一种基于视频分析与搜索聚合的城市商铺信息库自动构建系统及方法 | |
| CN119741062A (zh) | 基于知识图谱推理的涉外企业合规辅助决策方法及系统 | |
| CN110942392A (zh) | 一种业务数据处理方法、装置、设备和介质 | |
| Ye et al. | Cross cultural comparative study on emotional analysis of social media | |
| WO2016064679A1 (en) | Extracting product purchase information from electronic messages | |
| CN107609921A (zh) | 一种数据处理方法及服务器 | |
| Alamsyah et al. | Airfare fluctuation analysis with event and sentiment features by stacking ensemble model | |
| CN113127597B (zh) | 搜索信息的处理方法、装置及电子设备 | |
| CN104820713A (zh) | 一种基于用户历史数据获得工业产品名称同义词的方法 | |
| CN106845880A (zh) | 运单号归属的识别方法及装置 | |
| CN110413659B (zh) | 通用购物小票数据精准提取方法 | |
| US20250200624A1 (en) | Method and system for identifying a match with a product |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 17769351 Country of ref document: EP Kind code of ref document: A1 |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 17769351 Country of ref document: EP Kind code of ref document: A1 |