TW201913493A - Method for improving accuracy of commodity recommendation being applied to an online shopping platform - Google Patents

Abstract

This invention relates to a method for improving the accuracy of commodity recommendation. The method is applied to an online shopping platform and includes the following steps: receiving the commodity information, commodity image files, and commodity labels inputted from a user side, wherein the commodity information, the commodity labels, and the commodity image files are correlated; recognizing one or a combination of any two of the inputted commodity information, the commodity labels, and the commodity image files to generate a recognition result; acquiring the to-be-selected commodity labels according to the recognition result; outputting the to-be-selected commodity labels to the user side; receiving at least one of the to-be-selected commodity labels selected at the user side as a recommended commodity label; correlating the recommended commodity label with the commodity image files; and making a commodity online shopping page according to the commodity information, the commodity image files, the commodity labels, and the recommended commodity label.

Description

 Method for improving the accuracy of product recommendation  

The invention relates to a method for improving the accuracy of product recommendation, in particular to a method for providing recommended label information for a product image file on an online shopping platform, and providing a product matching and recommendation method.

In recent years, online shopping has gradually become one of the main channels for people to shop. In order to increase the sales volume of goods, many merchants have also invested in various online shopping platforms for sales. And the same goods are sold by different merchants in each online shopping platform, and the same or different product image files may be uploaded separately to display the goods to be sold.

In addition, in the online shopping platform website, in order to allow the merchant to display the products according to the browsing (searching) requirements of the consumers, in addition to the product image files, the product title, the product description and the Product information such as product classification, and further tag information marked on the product image file. Use these product information and label information to guide consumers to browse the products published on the online shopping platform according to their needs (search conditions). Under this premise, it is of great significance for merchants or consumers to provide correct product image files and label information.

However, how to give the correct image information to the product image file and effectively use the different product image files of the same product, so that the merchant can provide different product image files of the same product more completely and correctly, and attract consumers to purchase goods. Or, from the consumer's point of view, how to obtain a variety of different product image files of the same product, in order to better understand the goods, there are still many challenges on the current online shopping platform.

In view of the prior art, the correct label information is not effectively applied to the product image file, and the different product image files of the same product are effectively utilized, and the object of the present invention is to create a product webpage in the online shopping platform of the merchant. , providing different product image files of the same product, and recommending suitable label information for each product image file.

According to the purpose of the present invention, a method for improving the accuracy of product recommendation is provided on an online shopping platform, and includes the following steps: receiving product information, product image files, and product labels input by the user terminal, wherein the product information and the product are The tag is associated with the product image file and uses a web browsing device that is used by the merchant or consumer. Identifying one of the input product information, the product label, and the product image file, or a combination of any two or more, to generate the identification result, and obtaining the product label of the same or similar product image file in the existing online shopping platform according to the identification result Selecting the product label, outputting the candidate product label to the user end, receiving at least one of the selected product labels selected by the user end as the recommended product label, and associating the recommended product label with the product image file, according to the product information, the product image file, The product label and the recommended product label are made into a product online shopping page. In this way, the online shopping platform is not limited to relying on the product label input by the merchant to be associated with the product image file, and can also enable the consumer to provide the product label and the product image file, so that the online shopping platform is provided by the multi-party product label, and Can reduce the error rate of product labels. In turn, when the consumer browses or retrieves the product on the online shopping platform, the online shopping page of the product can be output to the consumer more accurately, so as to improve the accuracy of the product recommendation.

Wherein, when the identification result does not obtain the same or similar product image file in the existing online shopping platform, the candidate product label is not provided, and the product online shopping page is created according to the product information, the product image file, and the product label.

The online shopping page of the product includes a seller's online shopping page of the seller and a consumer online shopping page of the consumer.

The seller product online shopping page is further integrated into a seller product integration page according to the combination of one or both of the same or similar product information, product image file, product label and recommended product label.

The buyer product online shopping page is further integrated into a buyer product integration page according to a combination of one or both of the same or similar product information, product image file, product label and recommended product label.

Wherein, according to each seller's online shopping page and each buyer's online shopping page, the online shopping page of each buyer's product for which the user wants to purchase the same product, and the online shopping page for the seller's product selling the same product are paired.

The seller product classification is further generated according to the same or similar product information, product status and product attributes according to each seller's online shopping page.

The buyer shopping classification is further generated according to the interest, shopping habits and estimated style integration provided by the consumers.

Among them, the recommended product webpage is further provided to the user according to the seller's product classification and the buyer's shopping classification.

Among them, product information includes product classification, brand name, product name, product specifications, and product descriptions.

Among them, the web browsing device can be a smart phone, a tablet, a desktop computer or a notebook computer.

S101~S106‧‧‧Step procedure

1 is a schematic flow chart of an embodiment of the present invention.

FIG. 2 is a schematic diagram of a posted post page with a merchant role as a merchant role according to an embodiment of the present invention.

FIG. 3 is a schematic diagram of a request order page in which a usage end is a consumer role according to an embodiment of the present invention.

FIG. 4 is a schematic diagram of a user-side product description page according to an embodiment of the present invention.

FIG. 5 is a schematic diagram of providing a recommended product image file on a product description page by using a terminal according to an embodiment of the present invention.

6 is a schematic diagram of a product online shopping page according to an embodiment of the present invention.

FIG. 7 is a schematic diagram of a product online shopping page related to FIG. 6.

Figure 8 is a schematic diagram of the seller's merchandise integration page of the present invention.

Figure 9 is a schematic diagram of the buyer's product integration page of the present invention.

Figure 10 is a schematic diagram of a recommended product webpage of the present invention.

The present invention will be further described in detail below with reference to the drawings and the embodiments of the invention. The invention is defined.

Referring to FIG. 1 , the present invention is a method for improving the accuracy of product recommendation, and is applied to an online shopping platform, and includes the following steps: S101: receiving product information, product image files, and product labels input by the user end, wherein The product information and the product label are associated with the product image file; S102: identifying one of the input product information, the product label, and the product image file, or a combination of any two or more to generate the identification result; S103: obtaining the obtained result according to the identification result There is a product label of the same or similar product image file in the online shopping platform as the candidate product label; S104: output the candidate product label to the use end; S105: receive at least one of the selected product labels selected by the user end as the recommended product label , the recommended product label is associated with the product image file; S106: the product online shopping page is created according to the product information, the product image file, the product label, and the recommended product label.

In an embodiment of the present invention, the use end is a member of the online shopping platform used, and each member may be a merchant selling goods or a consumer purchasing goods, and the merchant or consumer is a web browsing device. Log in to the online shopping platform.

In this embodiment, when the user is used as a merchant or a consumer, the web browsing device (such as a smart phone, a tablet, a desktop computer, or a notebook computer, etc.) can be used. Enter the product information, product image file and product label to be sold to the online shopping platform. And when the online shopping platform generates the identification result, it is first determined whether the input product image file matches the product label, and when the product image file matches the product label, the step of generating the identification result is performed; otherwise, the online shopping platform An error message is generated to be returned to the user, so that the user can edit (including adding, modifying, or deleting) the product label, and then perform the step of generating the identification result, thereby improving the correctness of the product image file and the product label.

In this embodiment, when the usage end is a consumer role, the use end browses the product online shopping page or completes the purchase on the product online shopping page, the online shopping platform provides the use end to edit the product image file again on the product online shopping page (add new Or delete the product label and re-edit it into a new product online shopping page.

In this way, the online shopping platform is not limited to relying on the product label input by the merchant to be associated with the product image file, and can also enable the consumer to provide the product label and the product image file, so that the online shopping platform is provided by the multi-party product label, and Can reduce the error rate of product labels. Further, when the consumer browses or retrieves the product on the online shopping platform, the online shopping platform can output the online shopping page of the product to the consumer more accurately, thereby improving the accuracy of the product recommendation.

In the present invention, the online shopping page of the product includes a seller's online shopping page of the merchant and a consumer online shopping page of the consumer. The online shopping platform is further integrated into a seller product integration page according to each seller product online shopping page based on one or a combination of the same or similar product information, product image file, product label and recommended product label. Moreover, the online shopping platform is further integrated into a buyer's product integration according to each buyer's product online shopping page based on one or a combination of the same or similar product information, product image file, product label and recommended product label. page.

According to the above, the user of the present invention can view the seller's product online shopping page or the buyer's product online shopping page of other users on the seller's product integration page or the buyer's product integration page, regardless of the role of the merchant or the consumer.

In the present invention, the online shopping platform further generates a seller product classification according to the same or similar product information, product status, and product attribute integration according to each seller product online shopping page, wherein the product status includes whether the merchant has the spot, the product arrives at the sales place time, and the product is out. Delivery time, product delivery method, whether the goods are shipped by the online shopping platform certification seller, whether the goods have anti-imitation certification...etc. Furthermore, the product attributes include, but are not limited to, the current level of the product in the market, the degree of purchase inquiry, whether the product is a luxury item, etc. In addition, the buyer's shopping classification is further generated according to the interests, shopping habits, and estimated styles provided by the consumers, and the seller's product classification and the buyer's shopping classification are provided to the user-recommended product webpage. Among them, the interest is based on the daily life interest association problem input by the consumer's use end on the online shopping platform, and the consumer's possible shopping habits or the category of the product that he or she wishes to browse are estimated, for example, the interest category that the mountaineering may wish to browse is Including sporting goods, clothing accessories, health foods, travel-related goods, etc. In addition, shopping habits include personal photos provided by consumers, payment methods, shipping methods, and interactions with other users to estimate whether users are impulsive shopping, pure shopping, shopping, shopping, shopping, shopping... Wait. Estimated style: Combine the above interests and shopping habits, the shopping platform estimates the consumer's estimated style, including but not limited to the age layer as a distinction, and then further distinguish technology, electrical appliances, fashion apparel, home apparel, fashion footwear , women's footwear, sneakers, sporting goods, child care, regional characteristics, etc., and then give a positive degree of division, including but not limited to active, passive, active female trends, passive female trends, active technology, etc. Estimated.

In order to further understand the present invention, the following is specifically described based on the actual online shopping platform screen. The online shopping platform is exemplified by the screen of the smart phone, but the actual implementation is not limited to this: please refer to FIG. 2 and FIG. As shown in the figure, when the user is used as the role of the merchant, the option to publish the post can be selected to publish the product to be sold. In the page where the post is posted, the user can input at least the product information, the product image file, and the product label. The product information includes product descriptions, travel information and travel destinations. When the role of the consumer is used, the option to place an order may be selected to publish the desired product. In the page where the order is placed, the user may at least enter the product label, wherein the product information includes the order content, the order amount, and the product map. Like files, etc. Please refer to FIG. 4 to 5, wherein the product description and the order content further include product classification, brand name, product name, product specification and product model, etc., and the online shopping platform can recommend existing online shopping according to the input product description or the order content. The associated product image file in the platform is given to the user.

Furthermore, please refer to FIG. 6 to 7. When the online shopping platform generates a product online shopping page, the related product online shopping page can be queried according to the product information, the product image file and the product label (in Figure 6 Select to view related posts as an example). In addition, as shown in Figures 8-9, the online shopping platform is presented to the seller's product integration page (in the figure 8 to select who is selling as an example) and the buyer. The product integration page (in Figure 9 is a selection of who wants to buy as an example). In addition, as shown in FIG. 10, the online shopping platform is provided to the user according to the recommended product webpage, and the recommendation is based on the interest input previously used by the user, and the usage end uses the shopping habit of the online platform and the estimated style integration to generate the buyer. The shopping classification is further provided to the recommended product page of the user according to the seller product classification and the buyer shopping classification.

In summary, the present invention allows a merchant and a consumer to edit a product label, thereby improving the correct rate of the product label, and also determining whether the product image file matches the product label after inputting the product image file and the product label. Inconsistent allows the user to edit, which in turn improves the correctness of the product image file and product label. In addition, the use of the seller's product integration page and the buyer's product integration page can enable businesses and consumers to clearly understand the market conditions and greatly improve the success rate of commodity transactions.

The detailed description of the preferred embodiments of the present invention is intended to be limited to the scope of the invention, and is not intended to limit the scope of the invention. The patent scope of this case.

Claims (8)

 A method for improving the accuracy of product recommendation is applied to an online shopping platform, and includes the following steps: receiving a product information input by a user terminal, a product image file, and a product label, wherein the product information, the product label, and the product label Corresponding to the product image file; identifying the input of the product information, the combination of the product label and the product image file, or a combination of any two or more to generate a recognition result; obtaining the existing online shopping platform according to the identification result The product label of the same or similar product image file is used as a candidate product label; the candidate product label is output to the user terminal; and at least one of the candidate product labels selected by the user terminal is received as the recommended product label, The recommended product label is associated with the product image file; and the product online shopping page is created according to the product information, the product image file, the product label, and the recommended product label.    The method for improving the accuracy of product recommendation according to claim 1, wherein the user is a member of the online shopping platform, and each member may be a merchant selling the product or a consumer purchasing the product, and The merchant or the consumer logs in to the online shopping platform as a web browsing device.    The method for improving the accuracy of product recommendation according to claim 2, wherein when the identification result does not obtain the same or similar product image file already existing in the online shopping platform, the candidate product label is not provided, and according to the The product information, the product image file, and the product label are made into the product online shopping page.    The method for improving the accuracy of product recommendation according to claim 3, wherein before the online shopping platform generates the identification result, it is first determined whether the input image file of the product matches the product label; when the product image file When the product label matches, the step of generating the identification result is performed; otherwise, the online shopping platform generates an error message to report to the user, so that the user can edit the product label, and then generate the identification result. step.    The method for improving the accuracy of product recommendation according to claim 3, wherein when the user is the role of the consumer, the user browses the online shopping page of the product or completes the purchase at the online shopping page of the product, the user may The product online shopping page edits the product label again for the product image file, and then re-edits the new product online shopping page.    The method for improving the accuracy of product recommendation according to claim 3, wherein the online shopping page of the product includes a seller's online shopping page of the merchant; the online shopping platform further relies on the same or similar product information according to each of the seller's online shopping pages. And a combination of the product image file, the product label, and the recommended product label, or a combination of the two, is integrated into a seller product integration page.    The method for improving the accuracy of product recommendation according to claim 4, wherein the online shopping page of the product comprises a buyer's online shopping page of the buyer; the online shopping platform further relies on the same or similar product information according to each of the buyer's online shopping pages. And a combination of the product image file, the product label, and the recommended product label, or a combination of the two, is integrated into a buyer merchandise integration page.    The method for improving the accuracy of product recommendation according to claim 7, wherein the online shopping platform further generates a seller product classification according to the same or similar product information, the product status and the product attribute integration according to each of the seller product online shopping pages. And further generating a buyer shopping classification according to the interest, shopping habits, and estimated style integration provided by each consumer's user end, and providing the recommended product webpage to the user according to the seller product classification and the buyer shopping classification.