E-commerce Site Promotion Configuration Method and Its Device, Equipment, Medium, Product

Abstract

The present application discloses an e-commerce site promotion configuration method and its device, equipment, medium, and product, the method comprising: constructing a search text according to the product words and attribute words of the product title in the product display page; obtaining statistical data matching the search text, the statistical data comprising a plurality of candidate long-tail words and their historical search statistical indicators; determining the only candidate long-tail word semantically matching the product title as the target long-tail word according to the statistical indicators; configuring the target long-tail word to the page title of the product display page. The present application can determine the long-tail keyword based on the product title corresponding to the product display page, automatically configure it as the page title of the product display page, realize search engine keyword optimization, utilize the long-tail effect, and exert the effect of high search ranking through the long-tail keywords of a large number of single product display pages, so as to achieve the effect of improving the search engine drainage ability of the entire independent site of the e-commerce platform.

Description

Technical Field

[0001] The present application relates to the field of e-commerce information technology, and in particular to an e-commerce site promotion configuration method and its corresponding device, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the rapid development of cross-border e-commerce and independent station-based e-commerce websites, it is becoming increasingly difficult for products to be found and ranked high by search engines. Therefore, search engine optimization (SEO) for independent e-commerce sites is becoming increasingly important.

[0003] The webpage title is a highly summarized language of the content provided by a webpage, and is one of the main bases for search engine retrieval. Designing a good webpage title can greatly improve the ranking of the webpage in the search results. For e-commerce websites, it means designing a multi-dimensional and scientific SEO webpage title based on search volume, competitiveness, product content, etc., which can greatly increase the probability of the product being retrieved, thereby increasing exposure and conversion rate.

[0004] Currently, the generation of SEO web page titles for product pages mainly relies on the seller's staff, who set them by using various query tools and experience. This is very inefficient and the effect is difficult to guarantee. Summary of the invention

[0005] The primary purpose of the present application is to solve at least one of the above problems and to provide an e-commerce site promotion configuration method and its corresponding device, computer equipment, computer-readable storage medium, and computer program product.

[0006] In order to meet the various objectives of this application, this application adopts the following technical solutions:

[0007] An e-commerce site promotion configuration method provided to meet one of the purposes of this application includes the following steps:

[0008] Construct search text based on product words and attribute words in product titles on product display pages;

[0009] Obtaining statistical data matching the search text, the statistical data including a plurality of candidate long-tail words and their historical search statistical indicators;

[0010] Determine, according to the statistical indicators, a unique candidate long-tail word that semantically matches the product title as a target long-tail word;

[0011] The target long-tail word is configured in the page title of the product display page.

[0012] In some embodiments of the refinement, constructing a search text based on the product terms and attribute terms in the product title on the product display page includes the following steps:

[0013] Obtain the product title that has been entered on the product display page;

[0014] Perform word segmentation and part-of-speech recognition on the product title to obtain a word segmentation set composed of multiple word segments, which includes both word segments belonging to product terms and word segments belonging to attribute terms;

[0015] Extract keywords from the word segmentation set in association with the product title, and determine the search weight corresponding to each word segment. The search weight represents the potential search value of its corresponding word segment;

[0016] Determine the unique product term with the highest search weight and a predetermined number of multiple attribute terms, and splice and construct them into a search text.

[0017] In some embodiments of the refinement, obtaining statistical data that matches the search text. The statistical data includes multiple candidate long-tail keywords and their historical search statistical indicators, including the following steps:

[0018] Call a search interface to obtain statistical data of candidate long-tail keywords that match the search text. The statistical data includes keywords and their statistical indicators statistically formed based on the historical search behavior data of a large number of users. The statistical indicators include the average search volume of the corresponding keywords and their competitiveness adopted by different websites. The keywords are long-tail keywords.

[0019] In some embodiments of the refinement, determining the only candidate long-tail keyword that semantically matches the product title as the target long-tail keyword includes the following steps:

[0020] Perform data cleaning on the statistical data according to the statistical indicators to obtain valid candidate long-tail keywords;

[0021] Quantitatively determine the semantic similarity between each valid candidate long-tail keyword and the product title;

[0022] Use the average search volume as the matching weight of the semantic similarity, and calculate the comprehensive score of each valid candidate long-tail keyword;

[0023] Determine the valid candidate long-tail keyword with the highest comprehensive score as the target long-tail keyword.

[0024] In some specific embodiments, performing data cleaning on the statistical data according to the statistical indicators to obtain valid candidate long-tail keywords includes any one or any combination of the following steps:

[0025] Delete the candidate long-tail keywords in the statistical data whose number of words is less than a preset value;

[0026] Delete the candidate long-tail words within a preset time range in the said statistical data;

[0027] Delete the candidate long-tail words in the said statistical data whose competitiveness is higher than a preset level;

[0028] Delete the candidate long-tail words in the said statistical data whose average search volume is higher than a preset threshold.

[0029] In some specific embodiments, quantitatively determining the semantic similarity between each of the effective candidate long-tail words and the product title includes the following steps:

[0030] Encode to obtain the embedding vectors of the effective candidate long-tail words and the product title;

[0031] Use a pre-trained text feature extraction model to extract the high-level semantic information of the embedding vectors of the effective candidate long-tail words and the product title, and obtain their respective semantic feature vectors;

[0032] Use a preset data distance algorithm to calculate the data distance between the semantic feature vector of the product title and the semantic feature vectors of each effective candidate long-tail word as the corresponding semantic similarity.

[0033] In some further embodiments, configuring the target long-tail word into the page title of the product display page includes the following steps:

[0034] Display a search optimization page corresponding to the product display page to show the page title input box;

[0035] Configure the target long-tail word as the content data of the page title input box;

[0036] Respond to the user's submission instruction and publish the product display page and the search optimization page.

[0037] An e-commerce site promotion configuration device provided for one of the purposes of this application includes a search construction module, an index acquisition module, a target determination module, and a search optimization module. Among them, the search construction module is used to construct search text according to the product words and attribute words in the product title of the product display page; the index acquisition module is used to obtain statistical data matching the search text, and the statistical data includes multiple candidate long-tail words and their historical search statistical indicators; the target determination module is used to determine a unique candidate long-tail word that semantically matches the product title as the target long-tail word according to the statistical indicators; the search optimization module is used to configure the target long-tail word into the page title of the product display page.

[0038] In some of the in-depth embodiments, the search construction module includes: a title acquisition unit configured to acquire the product title entered in the product display page; a word segmentation and recognition unit configured to perform word segmentation and part-of-speech recognition on the product title to obtain a word segmentation set composed of multiple word segments, where the word segmentation set includes both word segments belonging to product words and word segments belonging to attribute words; a weight quantification unit configured to perform keyword extraction on the word segmentation set by associating with the product title, and determine the search weight corresponding to each word segment, where the search weight represents the potential search value of its corresponding word segment; and a search expression unit configured to determine the unique product word with the highest search weight and a predetermined number of multiple attribute words, and splice and construct them into a search text.

[0039] In some of the in-depth embodiments, the indicator acquisition module includes: calling a search interface to obtain statistical data of candidate long-tail words that match the search text, where the statistical data includes keywords and their statistical indicators statistically formed based on the historical search behavior data of a large number of users, and the statistical indicators include the average search volume of the corresponding keywords and their competition degrees adopted by different websites, and the keywords are long-tail words.

[0040] In some of the in-depth embodiments, the target determination module includes: a data cleaning unit configured to perform data cleaning on the statistical data according to the statistical indicators to obtain valid candidate long-tail words; a similarity quantification unit configured to quantify and determine the semantic similarity between each valid candidate long-tail word and the product title; a scoring quantification unit configured to use the average search volume as the matching weight of the semantic similarity and calculate the comprehensive score of each valid candidate long-tail word; and a target selection unit configured to determine the valid candidate long-tail word with the highest comprehensive score as the target long-tail word.

[0041] In some of the specific embodiments, the data cleaning unit includes any one or more of the following sub-modules: a word count cleaning sub-module configured to delete candidate long-tail words with a word count less than a preset value in the statistical data; a time cleaning sub-module configured to delete candidate long-tail words within a preset time range in the statistical data; a competition degree cleaning sub-module configured to delete candidate long-tail words with a competition degree higher than a preset level in the statistical data; and a search volume cleaning sub-module configured to delete candidate long-tail words with an average search volume higher than a preset threshold in the statistical data.

[0042] In some specific embodiments, the similarity quantification unit includes: a vector encoding subunit for encoding and obtaining the embedded vectors of the effective candidate long-tail words and the product title; a semantic extraction subunit for using a pre-trained text feature extraction model to extract the high-level semantic information of the embedded vectors of the effective candidate long-tail words and the product title to obtain their respective semantic feature vectors; and a similarity calculation subunit for using a preset data distance algorithm to calculate the data distance between the semantic feature vector of the product title and the semantic feature vectors of each effective candidate long-tail word as the corresponding semantic similarity.

[0043] In some further embodiments, the search optimization module includes: a page display unit for displaying a search optimization page corresponding to the product display page to display a page title input box; an automatic editing unit for configuring the target long-tail word as the content data of the page title input box; and an optimization publishing unit for responding to a user submission instruction to publish the product display page and the search optimization page.

[0044] A computer device provided for one of the purposes of the present application includes a central processing unit and a memory, and the central processing unit is used to call and run a computer program stored in the memory to execute the steps of the e-commerce site promotion configuration method described in the present application.

[0045] A computer-readable storage medium provided for another purpose of the present application stores a computer program implemented according to the e-commerce site promotion configuration method in the form of computer-readable instructions. When the computer program is called and run by a computer, it executes the steps included in the method.

[0046] A computer program product provided for another purpose of the present application includes a computer program / instructions. When the computer program / instructions are executed by a processor, they implement the steps of the method described in any embodiment of the present application.

[0047] Compared with the prior art, the technical solution of the present application at least includes the following technical advantages:

[0048] First, the present application constructs a search text for the product words and attribute words in the product title used for the product display page, retrieves candidate long-tail words and their statistical indicators with the search text, and then determines the optimal candidate long-tail word as the target long-tail word according to the statistical indicators, and uses the target long-tail word as the page title of the product display page to achieve search engine optimization of the product display page in a per-product personalized manner, serving the production of product display pages for e-commerce independent sites, without manual participation, and improving the efficiency of configuring the information required for a large number of product display pages.

[0049] Secondly, the long-tail words in this application, as the name suggests, are keywords with long-tail effect, also known as long-tail keywords. Although their search volume is relatively small, they have the advantage of strong targeting. For e-commerce platforms that have a large number of products and each product corresponds to a product display page, configuring the preferred long-tail keywords that match the product title as the page title of the product display page will help increase the probability of each product display page ranking high in the search engine's search results, thereby increasing the overall traffic of the entire independent site of the e-commerce platform.

[0050] In addition, when determining the target long-tail word for the page title of the product display page, the present application not only considers the semantic matching relationship between the candidate long-tail word and the product title, but also considers the historical search statistical indicators of the candidate long-tail word itself. The statistical indicator itself is the popularity information that characterizes the long-tail word in the historical process of being searched, which helps to make the best selection of candidate long-tail words. The long-tail word determined in this way, after combining the semantics of the product title as a reference, is not only highly consistent with the product title in semantics, but also more effective. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0052] Figure 1 A flowchart of a typical embodiment of the e-commerce site promotion configuration method of the present application;

[0053] Figure 2 A flowchart of a process of constructing a search text according to a product title in an embodiment of the present application;

[0054] Figure 3 A graphical user interface provided as an example for this application, with a product display page and a search engine optimization page displayed on both sides at the same time;

[0055] Figure 4 A flowchart of a specific process of determining a target long-tail word in an embodiment of the present application;

[0056] Figure 5 A flowchart of a process of cleaning statistical data in an embodiment of the present application;

[0057] Figure 6 A flowchart of a process for calculating the semantic similarity between a product title and a valid candidate long-tail word in an embodiment of the present application;

[0058] Figure 7 A schematic diagram of the process of configuring a page title to complete the publishing of a product display page in an embodiment of the present application;

[0059] Figure 8 It is a principle block diagram of the e-commerce site promotion configuration device of this application;

[0060] Figure 9 It is a schematic structural diagram of a computer device adopted by this application. Specific implementation manners

[0061] The embodiments of the present application will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be construed as a limitation to the present application.

[0062] Those skilled in the art of this technology can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the", and "said" used herein may also include the plural forms. It should be further understood that the term "including" used in the specification of the present application means the presence of the described features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or their groups. It should be understood that when we say an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0063] Those skilled in the art of this technology can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as the general understanding of those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as here.

[0064] Those skilled in the art can understand that the "client", "terminal", and "terminal device" used herein include both devices with a wireless signal receiver that only has the ability to receive and no ability to transmit, and devices with both receiving and transmitting hardware that can perform two-way communication on a two-way communication link. Such devices can include: cellular or other communication devices such as personal computers, tablet computers, etc., which have a single-line display or a multi-line display or a cellular or other communication device without a multi-line display; PCS (Personal Communications Service), which can combine voice, data processing, fax, and / or data communication capabilities; PDA (Personal Digital Assistant), which can include a radio frequency receiver, a pager, Internet / intranet access, a web browser, a notepad, a calendar, and / or a GPS (Global Positioning System) receiver; conventional laptop and / or palm computers or other devices, which are conventional laptop and / or palm computers or other devices with and / or including a radio frequency receiver. The "client", "terminal", and "terminal device" used herein can be portable, transportable, installed in a vehicle (air, sea, and / or land), or suitable for and / or configured to run locally, and / or run in a distributed manner at any other location on the earth and / or in space. The "client", "terminal", and "terminal device" used herein can also be a communication terminal, an Internet access terminal, a music / video playback terminal, such as a PDA, a MID (Mobile Internet Device), and / or a mobile phone with music / video playback functions, or can also be devices such as a smart TV, a set-top box, etc.

[0065] The hardware referred to by names such as "server", "client", and "service node" in this application is essentially an electronic device with the equivalent capabilities of a personal computer, and is a hardware device with the necessary components disclosed by the von Neumann principle, including a central processing unit (including an arithmetic unit and a controller), a memory, an input device, and an output device. The computer program is stored in its memory, and the central processing unit loads the program stored in the external memory into the memory for execution, executes the instructions in the program, and interacts with the input / output devices to complete specific functions.

[0066] It should be noted that the concept of "server" as referred to in this application can similarly be extended to the case applicable to a server cluster. According to the network deployment principle understood by those skilled in the art, the various servers should be logically divided. Physically, these servers can either be independent of each other but can be invoked through interfaces, or integrated into a single physical computer or a set of computer clusters. Those skilled in the art should understand this flexibility and should not be restricted by this when implementing the network deployment method of this application.

[0067] One or several technical features of this application, unless explicitly specified, can either be deployed on the server and accessed by the client remotely invoking the online service interface provided by the server, or directly deployed and run on the client for access.

[0068] The neural network models cited or possibly cited in this application, unless explicitly specified, can either be deployed on a remote server and remotely invoked by the client, or deployed on a client with sufficient device capabilities for direct invocation. In some embodiments, when it runs on the client, its corresponding intelligence can be obtained through transfer learning to reduce the requirements for the client's hardware operation resources and avoid over-occupying the client's hardware operation resources.

[0069] All kinds of data involved in this application, unless explicitly specified, can either be remotely stored on the server or stored on the local terminal device, as long as it is suitable for being invoked by the technical solution of this application.

[0070] Those skilled in the art should be aware that although the various methods of this application are described based on the same concept and thus show commonality with each other, unless otherwise specified, these methods can be executed independently. Similarly, for each embodiment disclosed in this application, they are all proposed based on the same inventive concept. Therefore, for concepts with the same expression, as well as concepts that are only appropriately transformed for convenience although the concept expressions are different, they should be equivalently understood.

[0071] For each embodiment to be disclosed in this application, unless explicitly pointed out as mutually exclusive, the relevant technical features involved in each embodiment can be cross-combined to flexibly construct new embodiments, as long as this combination does not deviate from the creative spirit of this application and can meet the requirements in the prior art or solve certain deficiencies in the prior art. Those skilled in the art should be aware of this flexibility.

[0072] A method for promoting and configuring an e-commerce site of the present application can be programmed as a computer program product and implemented by running on a client or a server. For example, in the application scenario of an e-commerce platform including live e-commerce in the present application, it is generally deployed on the server for implementation. Thus, by accessing the interface opened after the computer program product runs, human-computer interaction can be performed with the process of the computer program product through a graphical user interface to execute this method.

[0073] Please refer to Figure 1 , in a typical embodiment of the method for promoting and configuring an e-commerce site of the present application, it includes the following steps:

[0074] Step S1100: Construct a search text based on the product words and attribute words in the product title on the product display page:

[0075] When an independent site of an e-commerce platform needs to publish a certain product, product information needs to be entered, usually including but not limited to information such as the product title, product abstract, product description, etc. Subsequently, a product display page corresponding to these product information is generated for terminal users to call and browse. In order to facilitate improving the probability of the product being searched, it is also allowed to configure search engine optimization parameters together so that the corresponding product display page is more likely to be included and displayed by the search engine.

[0076] When the content data of the corresponding product title is input in the product title input box of the product display page, the product title can be obtained. Accordingly, various traditional word segmentation methods can be used to segment the product title and determine its corresponding part of speech, and a word segmentation set is obtained accordingly. The words in the word segmentation set usually belong to nouns in terms of part of speech, and some belong to adjectives and / or adverbs. Among the nouns, there are generally product words indicating the product content, while adjectives and / or adverbs are usually used to describe a certain aspect of the product's attributes, so they are also called attribute words. Thus, the product words and attribute words in the product title are determined.

[0077] In order to obtain candidate long-tail words based on the product title, each product word can be arbitrarily concatenated with all attribute words to form one or more search texts, and this search text can be submitted to a search interface for execution to obtain the candidate long-tail words. When constructing the search text, the product words and attribute words can also be filtered and optimized to reduce the number of search texts and improve the operation speed. Similarly, a unique concatenation rule can also be set for the concatenation method of product words and attribute words, so that the product words and attribute words are uniquely concatenated according to this concatenation rule.

[0078] Step S1200: Obtain statistical data matching the search text, where the statistical data includes multiple candidate long-tail words and their historical search statistical indicators:

[0079] According to the search text, the corresponding candidate long-tail word and the statistical index formed by the historical search of the candidate long-tail word can be obtained in any one of a variety of ways.

[0080] In one method, the candidate long-tail words and their corresponding statistical indicators can be directly obtained through the search interface provided by the search engine, and the calling interface is implemented to search for semantically matching long-tail words from a preset long-tail word ranking table according to the search text as the candidate long-tail words, and provide corresponding statistical indicators of the candidate long-tail words. Generally speaking, the search engine is the target search engine that is expected to search for the product display page.

[0081] In another way, the e-commerce platform can obtain the long-tail word ranking list through self-statistics or other traditional data retrieval means, and then call the search interface provided by the e-commerce platform to obtain the candidate long-tail words and their corresponding statistical indicators from the customized long-tail word ranking list.

[0082] The long-tail words, i.e. long-tail keywords, generally refer to search strings with a small average search volume per unit time and containing multiple words. Long-tail words have a long-tail effect. Although their average unit search volume is small, they have the advantage of accurate hits. When a search engine user enters a long-tail word, the page with the long-tail word as the page title will be able to obtain a high priority ranking effect. Since each independent site of the e-commerce platform often has a large number of product display pages, by configuring long-tail words as the page titles required for search engine optimization in a large number of product display pages, the advantage of a single page being accurately searched and ranked high can be used to trigger a scale effect, thereby increasing the user traffic of the entire independent site and achieving the effect of promoting the entire independent site.

[0083] The statistical indicators generally include the average search volume and competitiveness of each long-tail keyword. Generally speaking, each search engine will analyze the user's search expression on its own, and count the usage frequency of the keywords in the search expression by users and the usage frequency of the corresponding keywords by website pages. Accordingly, from the user side, the average search volume of the keywords used by users can be counted periodically, for example, on a monthly basis; from the website side, the competitiveness of the keywords used by pages can be counted periodically, for example, on a monthly basis. The average search volume can be numerical data, and the competitiveness can be ordinal data, such as options {1, 2, 3} representing high, medium, and low, depending on the quantization habits of each search engine for relevant data. The average search volume and competitiveness of each keyword constitute the corresponding statistical indicator of each keyword. In other embodiments, the statistical indicator may further include other information such as the usage time corresponding to the long-tail keyword. Organizing each keyword and its average search volume and competitiveness into mapping relation data can constitute a keyword ranking table. Among them, keywords containing more than a preset number of words, for example, more than two words, are used as long-tail keywords, and the long-tail keyword ranking table can be constructed by screening from the keyword ranking table.

[0084] Thus, it can be seen that the long-tail keyword ranking table actually stores statistical data statistically formed from long-tail keywords corresponding to the historical search behavior data of a large number of users. The statistical data includes each long-tail keyword and its corresponding various statistical indicators. The statistical indicators include the average search volume of the corresponding long-tail keyword and its competitiveness adopted by different websites.

[0085] According to the principle of constructing the long-tail keyword ranking table by the search engine, the e-commerce platform can also obtain the corresponding data by itself and construct the long-tail keyword ranking table by itself, or directly call the long-tail keyword ranking published by the search engine and store it locally for calling.

[0086] When obtaining the long-tail keyword that matches the search text from the long-tail keyword ranking table as a candidate long-tail keyword, the corresponding statistical indicator of the candidate long-tail keyword is obtained at the same time. Among them, the search text and the long-tail keyword can be matched based on rules, including fuzzy matching or exact matching, or semantic matching, etc. Those skilled in the art can flexibly apply according to the principles disclosed here. Thus, the matched candidate long-tail keywords and their corresponding statistical indicators constitute a candidate long-tail keyword subset.

[0087] Step S1300: Determine the only candidate long-tail keyword that semantically matches the product title according to the statistical indicator as the target long-tail keyword:

[0088] Since only one page title is required for each product display page, it is necessary to determine a candidate long-tail word from the subset of candidate long-tail words determined in the previous step as the target long-tail word. To this end, the corresponding statistical indicators of each candidate long-tail word can be referred to, including the average search volume and / or competitiveness therein, and the semantic matching degree between the candidate long-tail word and the product title, so as to optimize each candidate long-tail word.

[0089] For example, the average search volume and the semantic similarity between the candidate long-tail word and the product title can be jointly used as the main sorting field and the secondary sorting field respectively to perform multi-field reverse sorting, and then the candidate long-tail word ranked first in the reverse sorting is determined as the target long-tail word used as the page title.

[0090] Another example is that the competitiveness and the semantic similarity between the candidate long-tail word and the product title can be jointly used as the main sorting field and the secondary sorting field respectively to perform multi-field reverse sorting, and then the candidate long-tail word with lower competitiveness and the highest similarity is determined as the target long-tail word used as the page title.

[0091] Another example is that the average search volume, the competitiveness, and the semantic similarity between the candidate long-tail word and the product title can be comprehensively sorted in multiple fields to determine the candidate long-tail word with higher average search volume, lower competitiveness, and higher similarity, and thus the candidate long-tail word ranked first in the comprehensive ranking is used as the target long-tail word for the page title.

[0092] In addition, other methods can also be flexibly used, such as the method of semantic determination based on a deep learning model pre-trained to a convergent state, to determine the only candidate long-tail word that is highly similar to the product title, has a high average search volume, and / or has a low competitiveness as the target long-tail word.

[0093] It can be seen from this that with the help of the statistical indicators, multiple methods can be used to determine the only target long-tail word for the product title. This target long-tail word has the characteristics of being semantically matched with the product title and being associated with one or more of the statistical indicators, thus realizing the optimization of the candidate long-tail words.

[0094] Step S1400: Configure the target long-tail word into the page title of the product display page:

[0095] After the target long-tail word is determined, it can be configured as the page title of the product display page. After the product display page is published, each search engine searches and indexes the product display page according to its own implementation logic. Subsequently, searching using the target long-tail word in the search engine can more accurately hit the product display page, thus achieving the purpose of page promotion.

[0096] Through the typical embodiments of the present application and its corresponding multiple alternative embodiments, it can be seen that compared with the prior art, the technical solution of the present application has at least the following technical advantages:

[0097] First, the application constructs a search text for product words and attribute words in the product titles used on the product display page, retrieves candidate long-tail words and their statistical indicators through the search text, and then determines the optimal candidate long-tail word as the target long-tail word based on the statistical indicators, and uses the target long-tail word as the page title of the product display page, thereby achieving search engine optimization of the product display page in a personalized manner for each product, serving the production of product display pages for e-commerce independent sites without the need for human intervention, thereby improving the efficiency of configuring the information required for massive product display pages.

[0098] Secondly, the long-tail words in this application, as the name suggests, are keywords with long-tail effect, also known as long-tail keywords. Although their search volume is relatively small, they have the advantage of strong targeting. For e-commerce platforms that have a large number of products and each product corresponds to a product display page, configuring the preferred long-tail keywords that match the product title as the page title of the product display page will help increase the probability of each product display page ranking high in the search engine's search results, thereby increasing the overall traffic of the entire independent site of the e-commerce platform.

[0099] In addition, when determining the target long-tail word for the page title of the product display page, the present application not only considers the semantic matching relationship between the candidate long-tail word and the product title, but also considers the historical search statistical indicators of the candidate long-tail word itself. The statistical indicator itself is the popularity information that characterizes the long-tail word in the historical process of being searched, which helps to make the best selection of candidate long-tail words. The long-tail word determined in this way, after combining the semantics of the product title as a reference, is not only highly consistent with the product title in semantics, but also more effective.

[0100] See also Figure 2 In some of the further embodiments, the step S1100, constructing a search text according to product words and attribute words in the product title in the product display page, includes the following steps:

[0101] Step S1110: Obtain the product title entered in the product display page:

[0102] like Figure 3 As shown, when the user is editing the product display page, a search engine optimization area is displayed. The user enters the product title he sets in the product title input box on the left side of the figure, and the background obtains the text data of the product title.

[0103] Step S1120: Segment the product title and identify the part-of-speech of each word, obtaining a set of word segments composed of multiple word segments, which includes both word segments belonging to product words and word segments belonging to attribute words:

[0104] To segment the product title, any traditional statistical word segmentation algorithm such as N-Gram can be used to segment the product title first, obtaining a corresponding set of word segments. In this set of word segments, there are usually multiple product words and multiple attribute words. Product words are usually used to describe or refer to the product name, mostly nouns; attribute words are usually used to describe the product attributes, mostly adjectives or adverbs.

[0105] To perform part-of-speech analysis on each word segment in the set of word segments, a neural network model pre-trained to a convergent state can be used. Recommended models can adopt basic network architectures such as LSTM+CRF, BERT+CRF, etc., which can be flexibly selected by those skilled in the art. Such neural network models can be fine-tuned by those skilled in the art on the basis of pre-training until they converge, so that they can learn to perform word segmentation and part-of-speech analysis according to the embedding vectors of the word segment set of the input product title, and then determine the product words and attribute words in the product title.

[0106] Step S1130: Extract keywords from the set of word segments in association with the product title, and determine the search weight corresponding to each word segment. The search weight represents the potential search value of its corresponding word segment:

[0107] As mentioned above, in the set of word segments, the number of product words and attribute words may both be multiple. However, the syntax required to construct the search text generally follows the usage habits of natural language, so it needs to be as concise as possible. For this purpose, for example, a single representative product word can be combined with multiple representative attribute words to construct the search text. Therefore, it is necessary to quantitatively determine the search weights of the product word and the attribute words, that is, to determine the potential search value corresponding to each word segment in the set of word segments.

[0108] In one embodiment, the TF-IDF algorithm can be used to determine the TF-IDF value of each word segment in the set of word segments as the search weight. In another embodiment, the TextRank algorithm can also be used to construct a word segment graph for each word segment, and the weight corresponding to each word segment is determined according to the word segment graph as the search weight. In addition, those skilled in the art can also flexibly adopt other alternative methods according to the principles disclosed herein, as long as the potential search value of each word segment can be quantified.

[0109] Step S1140: Determine the unique product word with the highest search weight and a predetermined number of multiple attribute words, and splice and construct them into a search text:

[0110] According to the search weight, each word in the word set is sorted in reverse order, and the product words and attribute words therein are arranged in order according to their corresponding search weights. Based on this, the only product word with the highest search weight can be determined as the core product word of the product title. At the same time, according to a preset number, for example, 3, a corresponding number of attribute words with a high ranking are selected, and the core product word is spliced with the multiple attribute words selected, for example, spliced in order according to preset rules, so that the attribute words are in front and the core product words are in the back, thereby constructing a search text. Of course, the splicing order can also be adjusted to obtain multiple search texts, and those skilled in the art can implement this flexibly.

[0111] This embodiment performs word segmentation and part-of-speech recognition on the product title, and then calculates the search weight of each word segmentation that represents the potential search value, selects core product words and multiple preferred attribute words according to the search weight, and constructs a search text based on the core product words and the multiple preferred attribute words. The search text has a higher potential search value, and preliminary screening of candidate long-tail words is performed based on this, which can improve matching accuracy.

[0112] See also Figure 4 In some of the further embodiments, the step S1300, determining the only candidate long-tail word semantically matching the product title as the target long-tail word according to the statistical indicator, includes the following steps:

[0113] Step S1310: Clean the statistical data according to the statistical indicators to obtain valid candidate long-tail words:

[0114] As mentioned above, the statistical data includes statistical indicators, which include average search volume and / or competitiveness, and may even include other information. Therefore, the statistical indicators may be cleaned according to the constraints on the average search volume and / or competitiveness of the candidate long-tail words in the statistical data, thereby filtering out some invalid candidate long-tail words, and the remaining ones may be confirmed as valid long-tail words. The constraints may be flexibly set by those skilled in the art according to the principles disclosed herein.

[0115] Step S1320: quantify and determine the semantic similarity between each valid candidate long-tail word and the product title:

[0116] In order to examine the semantic closeness between each valid candidate long-tail word and the product title, the semantic similarity between the two can be calculated. When calculating the semantic similarity, the deep semantic information of the valid candidate long-tail word and the product title can be determined respectively, and on this basis, the data distance between each candidate long-tail word and the product title is determined by using a data distance algorithm, and then the data distance between each candidate long-tail word and the product title is expressed as a value corresponding to the semantic similarity through normalization.

[0117] Step S1330: Use the average search volume as the matching weight of the semantic similarity, and calculate the comprehensive score of each effective candidate long-tail word:

[0118] To facilitate the selection of the best among each effective candidate long-tail word, the comprehensive score of each effective candidate long-tail word can be determined first according to its statistical indicators and its similarity to the product title. An exemplary formula is as follows:

[0119] final_score = ln(2.001 - (search_score / max_search_score)) * sim_score

[0120] Among them, search_score is the average search volume corresponding to the current effective candidate long-tail word, max_search_score is the maximum average search volume among all effective candidate long-tail words. Take the logarithm of the ratio of 2.001 minus the two and multiply it by the corresponding semantic similarity to achieve weight matching, and ensure that the data is more stable, distinguish different differences, and quantitatively determine the corresponding comprehensive score final_score.

[0121] It should be understood that the above formula is only for example to illustrate the principle of using the average search volume of effective candidate long-tail words to weight and quantify their semantic similarity to the product title. Those skilled in the art can flexibly construct formulas according to the principle disclosed here as long as the same purpose is achieved. In this regard, it should be regarded as not exceeding the scope reflected by the creative spirit of this application.

[0122] After determining the corresponding comprehensive score for each candidate long-tail word, through the comprehensive score, it comprehensively reflects the average search popularity of the candidate long-tail word and its closeness to the product title, unifying the advantages and disadvantages of the effective candidate long-tail words into the same dimension. Therefore, the best among each effective candidate long-tail word can be selected according to the size of the comprehensive score.

[0123] Step S1340: Determine the effective candidate long-tail word with the highest comprehensive score as the target long-tail word:

[0124] To obtain the best effective candidate long-tail word as the target long-tail word, the effective candidate long-tail words can be sorted in reverse order according to the comprehensive score. Then, the effective candidate long-tail word ranked first, that is, the effective candidate long-tail word with the highest comprehensive score, can be determined as the target long-tail word required for generating the page title of this application.

[0125] After obtaining valid candidate long-tail keywords by cleaning the statistical data matched according to the search text in this embodiment, the comprehensive score corresponding to each valid candidate long-tail keyword is determined by applying a preset formula using the closeness between the valid candidate long-tail keyword and the product title and the average search volume in its own statistical indicators for reflecting the average search popularity, so as to quantify the quality of each valid candidate long-tail keyword. Then, the valid candidate long-tail keyword with the highest comprehensive score is selected as the target long-tail keyword required for the page title, realizing the optimal matching of the valid candidate long-tail keywords.

[0126] The target long-tail keyword determined in this embodiment is not only closer in semantics to the product title, but also has relatively better access popularity. Since the valid candidate long-tail keyword itself is the result of cleaning the statistical data and filters out some extreme cases, the determined target long-tail keyword is comprehensively optimal, which can ensure the advantage in long-tail keyword search after the product display page is published, thus assisting in the improvement of the overall site traffic.

[0127] Please refer to Figure 5 , in some specific embodiments, in order to achieve high-quality data cleaning effect and screen out valid candidate long-tail keywords, step S1310, cleaning the statistical data according to the statistical indicators to obtain valid candidate long-tail keywords, includes any one or more of the following steps:

[0128] Step S1311, deleting the candidate long-tail keywords with the number of words less than the preset value in the statistical data:

[0129] As the name implies, long-tail keywords are generally sentences with more than two or three words. For example, long-tail keywords can be set to contain no less than three words. Accordingly, the statistical data matched with the search text is filtered, and the candidate long-tail keywords with the number of words less than three are deleted from it, thus restricting the definition range of long-tail keywords.

[0130] Step S1312, deleting the candidate long-tail keywords within the preset time range in the statistical data:

[0131] Sometimes, the statistical data obtained by the search interface contains candidate long-tail keywords generated within a long time range. However, some products have sales timeliness. Accordingly, when obtaining the statistical data, the time information corresponding to each candidate long-tail keyword is obtained together, and then the candidate long-tail keywords are filtered according to the preset time range, and the candidate long-tail keywords falling within the preset time range are deleted from the statistical data, thus realizing the data cleaning of the expired candidate long-tail keywords.

[0132] Step S1313, deleting the candidate long-tail keywords with the competition degree higher than the preset level in the statistical data:

[0133] The competitiveness is generally represented in the form of "high, medium, low" or in numerical form, depending on the specific situation. In any case, since a candidate long-tail word has a high competitiveness, it means that a large number of websites and pages use the candidate long-tail word. If such a candidate long-tail word is still used, it is easy to cause the product display page to be difficult to rank high even if it is searched. Therefore, this feature can be used to delete the candidate long-tail words with a higher competitiveness from the statistical data to moderately avoid competition.

[0134] Step S1314: Delete the candidate long-tail words whose average search volume in the statistical data is higher than a preset threshold:

[0135] If the average search volume of candidate long-tail words is too high, it means that such candidate long-tail words are more popular. It also means that a large number of websites and pages may use such candidate long-tail words, and there is also a competition problem. Even if such candidate long-tail words are used in product display pages, it is difficult to improve their search rankings. According to this principle, a preset threshold corresponding to the average search volume can be set in advance, and candidate long-tail words with an average search volume higher than the preset threshold can be deleted from the statistical data to appropriately avoid competition.

[0136] The various embodiments disclosed herein may be used in any combination by those skilled in the art, for example, all embodiments corresponding to step S1311 to step S1314 may be adopted in full. In any case, those skilled in the art may flexibly apply the various embodiments disclosed herein based on different purposes to achieve the cleaning of statistical data matched with the search text, and ensure the validity of the valid candidate long-tail words relied upon when subsequently determining the target long-tail words, thereby ensuring the quality of the target long-tail words finally determined.

[0137] See also Figure 6 In some specific embodiments, step S1320, quantifying and determining the semantic similarity between each of the valid candidate long-tail words and the product title, includes the following steps:

[0138] Step S1321: Encode to obtain the embedding vector of the valid candidate long-tail word and the product title:

[0139] In order to calculate the semantic similarity between the valid candidate long-tail words and the product title, after conventional word segmentation, the product title and each of the valid candidate long-tail words are converted into embedding vectors according to the word list to complete the encoding.

[0140] Step S1322: Use a pre-trained text feature extraction model to extract high-level semantic information of the embedding vectors of the valid candidate long-tail words and the product title to obtain their respective semantic feature vectors:

[0141] Furthermore, a text feature extraction model is used to perform representation learning on the embedding vectors of the product title and the effective candidate long-tail words participating in the calculation, extract their deep semantic information, and obtain corresponding semantic feature vectors, that is, sentence vectors. The text feature extraction model can be a pre-trained model or a model fine-tuned and trained by those skilled in the art based on the pre-trained model, as long as it can extract corresponding semantic feature vectors from the embedding vectors of the text.

[0142] The text feature extraction model can be any one of the common basic models based on LSTM, Bert, Sentence-Transformer, etc., and those skilled in the art can flexibly select the model.

[0143] Step S1323: Use a preset data distance algorithm to calculate the data distance between the semantic feature vector of the product title and the tone feature vectors of each effective candidate long-tail word as the corresponding semantic similarity:

[0144] After determining the semantic feature vectors of the product title and each effective candidate long-tail word, a preset data distance algorithm can be used to calculate the similarity between the semantic feature vectors of the product title and each effective candidate long-tail word. The data distance algorithm can be any one of the traditional algorithms for calculating the distance between data points, such as the cosine similarity algorithm, Euclidean distance algorithm, Minkowski distance algorithm, Pearson correlation coefficient algorithm, Jaccard coefficient algorithm, etc.

[0145] Taking the cosine similarity algorithm as an example, its corresponding formula is as follows:

[0146]

[0147] Where A is the product title, B is a single effective candidate long-tail word, and n is the total number of elements in the semantic feature vector.

[0148] After calculating the data distance between the product title and each effective candidate long-tail word by applying the preset data distance algorithm, it can be normalized to the same dimension according to actual needs, so that the higher the value, the higher the semantic similarity, thereby obtaining the values corresponding to the semantic similarities of each effective candidate long-tail word, and subsequent target long-tail words can be preferably selected based on this.

[0149] In this embodiment, by means of the text feature model of deep learning, the semantic similarity between the effective candidate long-tail words and the product title is calculated. The obtained semantic similarity can better represent the semantic closeness between the effective candidate long-tail words and the product title. Subsequently, a better selection is made from the effective candidate long-tail words based on this, and the result will be more accurate.

[0150] Please refer to Figure 7, in a further exemplary embodiment, step S1400, configuring the target long-tail keyword into the page title of the product display page, includes the following steps:

[0151] Step S1410, displaying a search optimization page corresponding to the product display page to show a page title input box:

[0152] Please continue to refer to Figure 3 as shown in Figure 3 In the interface shown, a search optimization page is pre-displayed corresponding to the product display page, which includes a page title input box for inputting the target long-tail keyword determined in any of the above embodiments of the present application.

[0153] Step S1420, configuring the target long-tail keyword as the content data of the page title input box:

[0154] After the target long-tail keyword is obtained through any of the above embodiments in the present application, it can be automatically filled into the page title input box of the search optimization page and configured as the content data of the page title. The operating user can also further edit it or directly default to the automatically generated target long-tail keyword.

[0155] Step S1430, in response to the user submission instruction, publishing the product display page and the search optimization page:

[0156] When the operating user completes the entry of the data in the product display page and the search optimization page, the user submission instruction can be triggered by operating the submission control, and the product display page and the search optimization page are submitted to the background of the independent site to realize the publication of the product display page. At the same time, the data entered in the search optimization page, including the page title, is submitted. Thus, the publication process of a product display page is completed.

[0157] Each product display page can determine its corresponding page title in this way, thereby leveraging the long-tail effect of keywords to drive traffic to the e-commerce independent site.

[0158] This embodiment exemplarily shows the usage scenario of the target long-tail keyword automatically generated by the present application. It can be seen that with the help of the technical solution of the present application, when publishing each product on the independent site, there is no need to overly focus on the issue of search engine optimization. Just use the target long-tail keyword automatically generated based on the given product title as the page title of the product, which greatly improves the processing efficiency of the product publication link on the independent site and ensures that the page titles automatically generated for a large number of product display pages can assist in increasing the overall traffic of the independent site under the action of the long-tail effect.

[0159] Please refer to Figure 8, An e-commerce site promotion configuration device provided to meet one of the purposes of the present application is a functional embodiment of the e-commerce site promotion configuration method of the present application. The device includes a search construction module 1100, an index acquisition module 1200, a target determination module 1300, and a search optimization module 1400. Among them, the search construction module 1100 is used to construct a search text according to the product words and attribute words in the product title on the product display page; the index acquisition module 1200 is used to obtain statistical data matching the search text, and the statistical data includes a plurality of candidate long-tail words and their historical search statistical indicators; the target determination module 1300 is used to determine a single candidate long-tail word that semantically matches the product title as the target long-tail word according to the statistical indicators; the search optimization module 1400 is used to configure the target long-tail word into the page title of the product display page.

[0160] In some deepened embodiments, the search construction module 1100 includes: a title acquisition unit for acquiring the product title already input on the product display page; a word segmentation and recognition unit for performing word segmentation and part-of-speech recognition on the product title to obtain a word segmentation set composed of a plurality of word segments, which includes both word segments belonging to product words and word segments belonging to attribute words; a weight quantification unit for associating the product title to extract keywords from the word segmentation set and determining the search weight corresponding to each word segment, and the search weight represents the potential search value of its corresponding word segment; a search expression unit for determining the only product word with the highest search weight and a predetermined number of attribute words, and splicing and constructing them into a search text.

[0161] In some deepened embodiments, the index acquisition module 1200 includes: calling a search interface to obtain statistical data of candidate long-tail words matching the search text, and the statistical data includes keywords and their statistical indicators statistically formed according to the historical search behavior data of a large number of users. The statistical indicators include the average search volume of the corresponding keywords and their competitiveness adopted by different websites, and the keywords are long-tail words.

[0162] In some deepened embodiments, the target determination module 1300 includes: a data cleaning unit for cleaning the statistical data according to the statistical indicators to obtain valid candidate long-tail words; a similarity quantification unit for quantifying and determining the semantic similarity between each valid candidate long-tail word and the product title; a scoring quantification unit for using the average search volume as the matching weight of the semantic similarity and calculating the comprehensive score of each valid candidate long-tail word; a target selection unit for determining the valid candidate long-tail word with the highest comprehensive score as the target long-tail word.

[0163] In some specific embodiments, the data cleaning unit includes any one or more of the following sub-modules: a word count cleaning sub-module for deleting candidate long-tail words with a word count less than a preset value in the statistical data; a time cleaning sub-module for deleting candidate long-tail words within a preset time range in the statistical data; a competitiveness cleaning sub-module for deleting candidate long-tail words with a competitiveness higher than a preset level in the statistical data; and a search volume cleaning sub-module for deleting candidate long-tail words with an average search volume higher than a preset threshold in the statistical data.

[0164] In some specific embodiments, the similarity quantification unit includes: a vector encoding sub-unit for encoding to obtain the embedding vectors of the effective candidate long-tail words and the product title; a semantic extraction sub-unit for using a pre-trained text feature extraction model to extract the high-level semantic information of the embedding vectors of the effective candidate long-tail words and the product title to obtain their respective semantic feature vectors; and a similarity calculation sub-unit for using a preset data distance algorithm to calculate the data distance between the semantic feature vector of the product title and the semantic feature vectors of each effective candidate long-tail word as the corresponding semantic similarity.

[0165] In some deepened embodiments, the search optimization module 1400 includes: a page display unit for displaying a search optimization page corresponding to the product display page to display a page title input box; an automatic editing unit for configuring the target long-tail word as the content data of the page title input box; and an optimization publishing unit for responding to a user submission instruction to publish the product display page and the search optimization page.

[0166] To solve the above technical problems, an embodiment of the present application also provides a computer device. As Figure 9 shown, it is a schematic internal structure diagram of the computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected through a system bus. Among them, the computer-readable storage medium of the computer device stores an operating system, a database, and computer-readable instructions. The database can store a control information sequence. When the computer-readable instructions are executed by the processor, the processor can implement an e-commerce site promotion configuration method. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. The memory of the computer device can store computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor can execute the e-commerce site promotion configuration method of the present application. The network interface of the computer device is used to connect and communicate with the terminal. Those skilled in the art can understand, Figure 9The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0167] In this embodiment, the processor is used to execute Figure 8 The memory stores the program code and various data required to execute the above modules or submodules. The network interface is used to transmit data between user terminals or servers. The memory in this embodiment stores the program code and data required to execute all modules / submodules in the e-commerce site promotion configuration device of this application, and the server can call the program code and data of the server to execute the functions of all submodules.

[0168] The present application also provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the e-commerce site promotion configuration method of any embodiment of the present application.

[0169] The present application also provides a computer program product, including a computer program / instruction, which implements the steps of the method described in any embodiment of the present application when executed by one or more processors.

[0170] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments of the present application can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a computer-readable storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0171] To sum up, the present application can determine the long-tail keywords based on the product title of the product corresponding to the product display page, automatically configure it as the page title of the product display page, realize search engine keyword optimization, utilize the long-tail effect, and exert the effect of top search ranking through the long-tail keywords of a large number of single product display pages, so as to achieve the effect of improving the search engine traffic-generating capacity of the entire independent site of the e-commerce platform.

[0172] Those skilled in the art can understand that the various operations, methods, steps, measures, and solutions in the processes discussed in this application can be alternated, changed, combined, or deleted. Further, other steps, measures, and solutions in the various operations, methods, and processes discussed in this application can also be alternated, changed, rearranged, decomposed, combined, or deleted. Further, those in the prior art having steps, measures, and solutions in the various operations, methods, and processes disclosed in this application can also be alternated, changed, rearranged, decomposed, combined, or deleted.

[0173] The above are only partial embodiments of this application. It should be noted that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A method for promoting and configuring an e-commerce site, characterized in that, It includes the following steps: Construct a search text based on the product words and attribute words in the product title on the product display page; Obtain statistical data that matches the search text, where the statistical data includes multiple candidate long-tail words and their historical search statistical metrics; Determine a single candidate long-tail word that semantically matches the product title as the target long-tail word according to the statistical metrics, including: cleaning the statistical data according to the statistical metrics to obtain valid candidate long-tail words; quantifying and determining the semantic similarity between each valid candidate long-tail word and the product title; using the average search volume as the matching weight of the semantic similarity, and calculating the comprehensive score of each valid candidate long-tail word; determining the valid candidate long-tail word with the highest comprehensive score as the target long-tail word; Configure the target long-tail word into the page title of the product display page; Among them, the quantifying and determining the semantic similarity between each valid candidate long-tail word and the product title includes: encoding to obtain the embedding vectors of the valid candidate long-tail words and the product title; using a pre-trained text feature extraction model to extract the high-level semantic information of the embedding vectors of the valid candidate long-tail words and the product title to obtain their respective semantic feature vectors; using a preset data distance algorithm to calculate the data distance between the semantic feature vector of the product title and the semantic feature vectors of each valid candidate long-tail word as the corresponding semantic similarity.

2. The e-commerce site promotion configuration method according to claim 1, wherein Constructing a search text based on the product words and attribute words in the product title on the product display page includes the following steps: Obtain the product title already entered on the product display page; Perform word segmentation and part-of-speech recognition on the product title to obtain a word segmentation set composed of multiple word segments, which includes both word segments belonging to product words and word segments belonging to attribute words; Extract keywords from the word segmentation set by associating with the product title, and determine the search weight corresponding to each word segment, where the search weight represents the potential search value of its corresponding word segment; Determine the only product word with the highest search weight and a predetermined number of attribute words, and splice and construct them into a search text.

3. The e-commerce site promotion configuration method according to claim 1, characterized in that Obtain statistical data that matches the search text, where the statistical data includes multiple candidate long-tail words and their historical search statistical metrics, including the following steps: Call the search interface to obtain the statistical data of candidate long-tail words that match the search text. The statistical data includes keywords and their statistical metrics statistically formed based on the historical search behavior data of a large number of users. The statistical metrics include the average search volume of the corresponding keywords and their competitiveness adopted by different websites. The keywords are long-tail words.

4. The e-commerce site promotion configuration method according to claim 1, wherein Clean the statistical data according to the statistical metrics to obtain valid candidate long-tail words, including any one or any combination of the following steps: Delete the candidate long-tail words in the statistical data with the number of words less than the preset value; Delete the candidate long-tail words in the statistical data within the preset time range; Delete the candidate long-tail words in the statistical data with a competitiveness higher than the preset level; Delete the candidate long-tail words in the statistical data with an average search volume higher than the preset threshold.

5. The e-commerce site promotion configuration method according to any one of claims 1 to 4, characterized in that, Configuring the target long-tail word into the page title of the product display page includes the following steps: Display a search optimization page corresponding to the product display page to show a page title input box; Configure the target long-tail keyword as the content data of the page title input box; In response to the user submitting an instruction, publish the product display page and the search optimization page.

6. An e-commerce site promotion configuration device, characterized in that, It includes: A search construction module for constructing a search text according to the product word and attribute word of the product title in the product display page; An index acquisition module for acquiring statistical data matching the search text, where the statistical data includes multiple candidate long-tail keywords and their historical search statistical indicators; A target determination module for determining a single candidate long-tail keyword that semantically matches the product title as the target long-tail keyword according to the statistical indicators, including: a data cleaning unit for cleaning the statistical data according to the statistical indicators to obtain valid candidate long-tail keywords; a similarity quantification unit for quantifying and determining the semantic similarity between each valid candidate long-tail keyword and the product title; a scoring quantification unit for using the average search volume as the matching weight of the semantic similarity and calculating the comprehensive score of each valid candidate long-tail keyword; a target selection unit for determining the valid candidate long-tail keyword with the highest comprehensive score as the target long-tail keyword; A search optimization module for configuring the target long-tail keyword into the page title of the product display page; Among them, the similarity quantification unit includes: a vector encoding subunit for encoding and obtaining the embedding vectors of the valid candidate long-tail keyword and the product title; a semantic extraction subunit for using a pre-trained text feature extraction model to extract the high-level semantic information of the embedding vectors of the valid candidate long-tail keyword and the product title to obtain their respective semantic feature vectors; a similarity calculation subunit for using a preset data distance algorithm to calculate the data distance between the semantic feature vector of the product title and the tone feature vectors of each valid candidate long-tail keyword as the corresponding semantic similarity.

7. The e-commerce site promotion configuration device according to claim 6, wherein, The search construction module includes: A title acquisition unit for acquiring the product title already input in the product display page; A word segmentation and recognition unit for performing word segmentation and part-of-speech recognition on the product title to obtain a word segmentation set composed of multiple word segments, which includes both word segments belonging to product words and word segments belonging to attribute words; A weight quantification unit for associating the product title to extract keywords from the word segmentation set and determining the search weight corresponding to each word segment, where the search weight represents the potential search value of its corresponding word segment; A search expression unit for determining the only product word with the highest search weight and a predetermined number of multiple attribute words, and splicing and constructing them into a search text.

8. The e-commerce site promotion configuration device according to claim 6 or 7, characterized in that The index acquisition module includes: calling a search interface to obtain statistical data of candidate long-tail keywords matching the search text, where the statistical data includes keywords and their statistical indicators statistically formed according to the historical search behavior data of a large number of users, and the statistical indicators include the average search volume of the corresponding keywords and their competitiveness adopted by different websites, and the keywords are long-tail keywords.

9. A computer device, comprising a central processing unit and a memory, characterized in that, The central processing unit is configured to call and run a computer program stored in the memory to execute the steps of the method according to any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that, It stores, in the form of computer-readable instructions, a computer program implemented according to the method according to any one of claims 1 to 5. When the computer program is called and run by a computer, it executes the steps included in the corresponding method.

Images