Disclosure of Invention
The application provides a method for optimizing webpage content, which aims to solve at least one problem of optimizing content by a search engine in the prior art. The application further provides a device for optimizing the webpage content and the electronic equipment.
The application provides a method for optimizing webpage content, which comprises the following steps:
providing a content material library, wherein the content material library comprises description keywords of commodity objects and label information of the description keywords, and the label information of the description keywords comprises one or more of application scenes, application groups, application areas, application time, application languages, application types and application industries;
determining a target webpage to be optimized, wherein the target webpage comprises at least one target object;
acquiring characteristic information of the target object;
determining description keywords of which the label information is matched with the feature information from the content material library;
and generating a key point description text related to the target object based on the description key words matched with the characteristic information, and adding the key point description text into the target webpage so as to optimize the content of the target webpage.
Optionally, the providing a content material library includes:
acquiring original content related to a commodity object, wherein the original content comprises one or more of webpage content, pictures, videos and knowledge base content related to the commodity object;
acquiring a first description keyword of the commodity object from the original content;
taking the first description keyword as a description keyword of the commodity object;
marking each description keyword to generate label information of the description keyword.
Optionally, before marking each of the description keywords and generating tag information of the description keyword, the method further includes:
performing text translation on the first description keyword to obtain a second description keyword;
and taking the second description keyword as the description keyword of the commodity object.
Optionally, the text translation of the first description keyword to obtain a second description keyword includes at least one of the following steps:
performing feature analysis on the first description keywords, and determining second description keywords which accord with the description features on the basis of the extracted description features, wherein the second description keywords comprise one or more of alternative names, proper nouns and event names of the first description keywords;
and translating the first description keywords according to a second language, and taking the translated text as a second description keyword.
Optionally, the obtaining the first description keyword of the commodity object from the original content includes at least one of the following steps:
extracting a first description keyword of the commodity object from a webpage or a picture by a character recognition technology;
extracting a first description keyword of the commodity object from a video through a video understanding technology;
and extracting a first description keyword of the commodity object from the webpage text.
Optionally, the extracting, by using a video understanding technology, the first description keyword of the commodity object from the video includes:
identifying interested areas of each frame of image in the video, and determining corresponding interested areas;
carrying out object recognition on the region of interest, determining a commodity object and commodity characteristics as corresponding first description keywords, wherein the commodity characteristics comprise: one or more of color, morphology, material; and/or performing action recognition according to the region of interest, and determining a corresponding commodity object and a first description keyword;
and analyzing the associated knowledge of the first description keywords, determining associated knowledge information, and adding the associated knowledge information into the first description keywords of the commodity object.
Optionally, the method further includes:
providing a content template, wherein the content template provides different content templates according to the types, industries and language expressions of different commodity objects;
generating a key point description text related to the target object based on the description key words matched with the feature information, wherein the key point description text comprises the following steps:
matching a content template according to the label information of the description keyword matched with the characteristic information;
and synthesizing by adopting the content template and the description keywords to generate a key point description text related to the target object.
Optionally, the method further includes:
collecting a release feedback result corresponding to the optimized target webpage, wherein the release feedback result retrieves result information and/or click information;
and updating the content material library according to the release feedback result.
Optionally, the determining the target webpage to be optimized includes at least one of the following steps:
detecting the webpage and determining retrieval key information; taking the webpage as a target webpage to be optimized under the condition that the retrieval key information meets the retrieval requirement;
and receiving an optimization instruction, and determining a target webpage to be optimized according to the optimization instruction.
The embodiment of the application also provides a method for creating a content material library, which comprises the following steps:
acquiring original content related to a commodity object, wherein the original content comprises one or more of webpage content, pictures, videos and knowledge base content related to the commodity object;
acquiring a first description keyword of the commodity object from the original content;
taking the first description keyword as a description keyword of the commodity object;
marking each description keyword to generate label information of the description keyword, wherein the label information of the description keyword comprises one or more of an applicable scene, an applicable crowd, an applicable area, an applicable time, an applicable language, an applicable type and an applicable industry;
and generating a content material library to determine description keywords of which the tag information is matched with the feature information of the target object of the target webpage, generating a key point description text related to the target object based on the matched description keywords, adding the key point description text into the target webpage, and optimizing the content of the target webpage.
Optionally, before marking each of the description keywords and generating tag information of the description keyword, the method further includes:
performing text translation on the first description keyword to obtain a second description keyword;
and taking the second description keyword as the description keyword of the commodity object.
Optionally, the obtaining the first description keyword of the commodity object from the original content includes at least one of the following steps:
extracting a first description keyword of the commodity object from a webpage or a picture by a character recognition technology;
extracting a first description keyword of the commodity object from a video through a video understanding technology;
and extracting a first description keyword of the commodity object from the webpage text.
An electronic device, comprising: a processor; and a memory having executable code stored thereon that, when executed by the processor, performs a method as described in embodiments of the present application.
The present application also provides a computer-readable storage medium having stored therein instructions, which when run on a computer, cause the computer to perform the various methods described above.
The present application also provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform the various methods described above.
Compared with the prior art, the method has the following advantages:
the optimization method of the webpage content, provided by the embodiment of the application, can provide a content material library, the content material library comprises description keywords of commodity objects and label information of the description keywords, the label information of the description keywords comprises one or more of application scenes, application crowds, application areas, application time, application languages, application types and application industries, so that a rich data base of wind can be provided for webpage optimization, the quality of the optimized content is improved, then a target webpage to be optimized is determined, the target webpage comprises at least one target object, the characteristic information of the target object is obtained, the label information and the description keywords matched with the characteristic information are determined from the content material library, the key point description texts related to the target object are generated based on the description keywords matched with the characteristic information, the key point description texts are added into the target webpage to optimize the target webpage content, the description keywords can be automatically matched, the key point description texts optimized of the webpage are generated, the webpage content is richer, the search engine is convenient to retrieve, and the optimization efficiency and the optimization quality are effectively improved.
Detailed Description
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is capable of implementation in many different ways than those herein set forth and of similar import by those skilled in the art without departing from the spirit of this application and is therefore not limited to the specific implementations disclosed below.
In the application, a method and a device for optimizing webpage content and electronic equipment are provided. Each of the schemes is described in detail in the following examples.
First embodiment
Please refer to fig. 1, which is a flowchart illustrating steps of an embodiment of a method for optimizing web page content according to the present application. In this embodiment, the method may include the steps of:
step S101: providing a content material library, wherein the content material library comprises description keywords of commodity objects and label information of the description keywords, and the label information of the description keywords comprises one or more of application scenes, application groups, application areas, application time, application languages, application types and application industries.
According to the embodiment of the application, a content material library can be preset and used for providing various information for describing the commodity object, so that the description of the commodity object can be enriched. The content material library comprises description keywords of the commodity object and label information of the description keywords, the description keywords are keywords for describing characteristics of the commodity object, the described characteristics can be characteristics of various dimensions, such as scene dimensions and the like, therefore, the description keywords are also corresponding to the label information, and the dimensions of the characteristics are represented through the label information. The description keywords may describe various types of characteristics of the commodity object, such as keywords of various aspects of the name, appearance, usage, material, and the like of the commodity object.
The label information describing the keywords comprises one or more of applicable scenes, applicable groups, applicable areas, applicable time, applicable languages, applicable types and applicable industries. The applicable scene is used for indicating scene information applicable to the keyword in a scene dimension, such as being applicable to scenes of sports, commuting and the like. The applicable population is used for indicating user population information applicable to the keyword in user dimension, such as students, ladies, men and the like. The applicable region is used for indicating the region information applicable to the keyword in a region dimension, and the region can be a region in a geographic dimension, such as china, the united kingdom and the like. The applicable time is used for indicating the time information applicable to the keyword in a time dimension, such as winter, summer, 202X new year, and the like, and the label of the time dimension can be in various levels, such as quarterly, annual, monthly, hour level, and the like. The applicable language is used for indicating the language information applicable to the keyword in the language dimension, such as Chinese, japanese, english, and the like. The applicable type is used for indicating the type of the webpage plate, such as title, text, detail page, and the like, which the keyword is applicable to in the dimension of the webpage. The method is suitable for industry, and is used for indicating the industry information suitable for the keyword in the industry dimension, such as the industry information and the like of a commodity object, for example, the industry information and the like is electronic industry, food industry, household industry and the like, and can also comprise descending sub-industries of the industry, such as industries of furniture, lamps, security protection, bedding and the like. The tag information of the content material is not limited to the above-described tags (application scene, application crowd, application area, application time, application language, application type, application industry). In specific implementation, corresponding label information can be set for the content material based on various characteristic information of the commodity, and the method provided by the embodiment of the application has no limitation on the material label information.
In an example of a content material library, description keywords of commodity objects and tag information thereof are shown in table 1 below,
TABLE 1
In the embodiment of the present application, the content material library may be created by the following steps: acquiring original content related to a commodity object, wherein the original content comprises one or more of webpage content, pictures, videos and knowledge base content related to the commodity object; acquiring a first description keyword of the commodity object from the original content; taking the first description keyword as a description keyword of the commodity object; marking each description keyword to generate label information of the description keyword.
The original content of the commodity object can be obtained in various ways, the original content is various contents related to the commodity object, such as webpage contents of an introduction page, a detail page, an advertisement page and the like of the commodity object, further such as pictures, videos and the like corresponding to the commodity object, further such as knowledge base contents related to the commodity object, such as a classification knowledge base of the commodity object and the like, and data of various description and loading commodity objects can be obtained as the original content. The source of the original content may include an e-commerce website, various websites for delivering advertisements of commodity objects, various web page data obtained based on a search engine, web page snapshots, and the like. The knowledge base of the commodity object can be formed according to various characteristics of the commodity object, such as field characteristics, background characteristics, use characteristics, attribute characteristics and the like.
The obtained original content can be analyzed, and the keyword data describing the commodity object can be obtained by analyzing according to different data types by adopting various technologies such as text recognition, image recognition, video understanding and the like and used as a first keyword of the commodity object. In an optional embodiment, obtaining the first description keyword of the commodity object from the original content includes at least one of the following steps: extracting a first description keyword of the commodity object from a webpage or a picture by a character recognition technology; extracting a first description keyword of the commodity object from a video through a video understanding technology; and extracting a first description keyword of the commodity object from the webpage text.
For the characteristics of the text type, text data in a webpage, text data in an image and a video picture and the like, identification can be carried out based on a character identification technology, and a first description keyword is extracted. The text data in the image and the video picture may be processed by OCR (Optical Character Recognition), so as to obtain corresponding text data. And voice recognition can be carried out on the audio data of the video data to obtain corresponding text data. The obtained text data can be processed through text analysis to determine corresponding keywords, for example, semantic analysis and text segmentation are performed on the text data to determine corresponding keywords as first description keywords.
The commodity object and the corresponding keyword can be identified in the video data based on video understanding technology for the video data, and the first description keyword is used as the first description keyword. The recognition understanding technology is a technology for automatically recognizing and analyzing the content in the video through an automatic analysis means. In the embodiment of the application, one or more commodity objects in the video data can be automatically identified through various video analysis means, the description information of each commodity object is identified, and the first description keyword is analyzed according to the description information.
Wherein, the extracting the first description keyword of the commodity object from the video through the video understanding technology comprises: identifying interested areas of each frame of image in the video, and determining corresponding interested areas; identifying a commodity object and a first description keyword thereof based on the region of interest; and analyzing the associated knowledge of the first description keywords, determining associated knowledge information, and adding the associated knowledge information into the first description keywords of the commodity object. Identifying a commodity object and a first description keyword thereof based on the region of interest, wherein the identifying comprises: carrying out object recognition on the region of interest, determining a commodity object and commodity characteristics as corresponding first description keywords, wherein the commodity characteristics comprise: one or more of color, morphology, material; and/or performing action recognition according to the region of interest, and determining a corresponding commodity object and a first description keyword;
the Region Of Interest ROI (Region Of Interest) can be identified for the video and the image, the Region Of Interest is determined by processing the image, for example, the pixel characteristics Of the image are determined by gray processing, the Region Of Interest is identified, and the Region can be marked by a square frame, a circle, an ellipse, an irregular polygon, and the like.
The region may then be subjected to recognition processing to determine a corresponding commodity object and description information of the commodity object, and a keyword may be determined as a first description keyword based on the description. The commodity object can be determined based on target identification, the appearance style of the commodity object can be identified, the first description keyword can be determined, action identification can be performed by combining the characteristics of the video, so that the first description keyword which is the commodity object is determined, or gesture identification is performed on the commodity object and the associated object thereof based on the video or the image, the first description keyword is determined, and the like.
In one example, performing object recognition on a video and an image, extracting a commodity object as a recognized target object from each frame image and each image in the video, and recognizing appearance feature information of the commodity object, for example, recognizing through edge detection or an algorithm model, so as to obtain commodity features of the commodity object includes: color, form, material. The shape features may be features of appearance styles of the commodity objects, such as features representing shapes, commodity types, styles and the like, and the material features may be features related to materials of the commodity objects, such as materials of clothing commodity objects and the like.
In another example, motion recognition and tracking are performed based on each frame image in the video, a commodity object and motion features are recognized, semantic analysis, motion analysis and other processing are performed based on the motion features, and a first description keyword of the commodity object is determined. For example, the content of playing the motion sensing game through the motion sensing game device is identified and demonstrated through the video, the commodity object can be correspondingly identified as the motion sensing game device, and the first description keyword can be the motion sensing game, interest, reality and the like. If the action of stepping on one luggage box by the user is identified, the commodity object can be determined to be the luggage box, and the first description keyword is determined to be load-bearing, firm and durable and the like corresponding to the action characteristic.
In the embodiment of the present application, the above manners may also be combined to determine the commodity object and the first description keyword thereof, for example, after the region of interest is determined, the commodity object may be determined through target recognition, and the first description keyword is determined by combining action recognition.
After the commodity object and the first description keyword are recognized, associative knowledge analysis can be further performed based on the first description keyword, the associative knowledge analysis is an associative way for associative knowledge, that is, the associative analysis is performed based on the first description keyword, wherein the associative word library and the like can be analyzed, the associative word library includes words associated with various types of keywords, for example, various associated adjectives are set as associated words for the name of the commodity object, and the associated words of trousers include: and the commodity characteristics corresponding to the image recognition can be collected for the vocabulary, and the vocabulary can be determined as the related vocabulary serving as the first description keyword. An algorithm model can also be created through algorithms such as machine learning and neural networks, and corresponding associated words can be determined as first description keywords based on the algorithm model.
Therefore, the characteristics of the images and the videos can be analyzed to determine the commodity object and the first key description information thereof, the identification steps can establish an algorithm model through corresponding algorithms such as machine learning, neural networks and the like, and the corresponding first description keyword is determined based on the algorithm model.
In some other examples, the first description keyword may also be subjected to text translation to obtain a second description keyword; and taking the second description keyword as the description keyword of the commodity object. The text translation can be performed based on the first description keyword, the text translation can be a process of text-based translation, indirect translation and the like, the translation can be a translation between words in different languages, the indirect translation can be understood as a translation between words in multiple languages, and can also be understood as an extension of the keyword itself. Therefore, the second description keywords in different languages can be determined based on the first description keywords, and expansion can also be performed based on the characteristics of the first description keywords, such as determination of similar words, associated information and the like to obtain the second description keywords. Therefore, in an optional embodiment, the text translating the first description keyword to obtain the second description keyword includes at least one of the following steps: performing feature analysis on the first description keywords, and determining second description keywords which accord with the description features on the basis of the extracted description features, wherein the second description keywords comprise one or more of alternative names, proper nouns and event names of the first description keywords; and translating the first description keywords according to a second language, and taking the translated text as a second description keyword. The first description keyword is subjected to feature analysis, and the feature analysis can also be determined based on the result of the original content analysis, and other feature analysis processing is obtained and executed, so that description features corresponding to the first description keyword are obtained, for example, the first description keyword is eleven, description features such as dates and festivals are determined based on the analysis, the name or event of the first description keyword is correspondingly determined to be national celebration festival, and if the first description keyword is hand warmer, the description features such as hand warmer and warm keeping are determined based on the feature analysis, and the name or event of the first description keyword is correspondingly determined to be hot water bag, warm warmer and the like. The term "another name" refers to another name of the first description keyword, such as different names familiar to users in different regions, different ways of adjectives, such as chufa and hoof, a proper noun is a kind of noun representing a specific person or object, such as place names, holiday names, etc., and an event name representing a name corresponding to some events, such as a promotion activity of an e-commerce website, etc., and accordingly can be used as an extended second description keyword. In other examples, the first keyword may also be subjected to a translation process in a second language, such as the first description keyword "charger baby", and "power bank" or "portable charge" is obtained by english translation as the second description keyword.
Through the steps, the commodity object and the first description keyword, the second description keyword and the like thereof can be determined and used as the description keyword of the commodity object. The attributes of the description keywords in the dimensions can then be analyzed and tagged. For example, for "power bank", it may be determined that its usage scenario is all, the applicable population is over 14 years old, the applicable region is all, the applicable time is all, the applicable language is all, the applicable type is the title, the applicable industry is electronic, and the like. The labels that are all applicable may also be omitted, or labeled as labels such as all, 1, true, etc., and the labels that are all inapplicable may be labeled as labels such as null, 0, false, etc., to obtain the label information of each description keyword. Then, the description keywords and the label information thereof can be recorded to obtain the corresponding content material library.
Step S103: determining a target webpage to be optimized, wherein the target webpage comprises at least one target object.
And acquiring a target webpage to be optimized, wherein the target webpage can be a webpage in an e-commerce website, such as a webpage which is already on-line, or a webpage which is not published in the creation of a seller user and a merchant user, and the like. For example, the web pages in the e-commerce website are detected, the key information of the web pages is determined, and if the key information does not meet the search requirement, for example, the number of the keywords is less than the number threshold, the web pages can be used as the target web pages to be optimized. As another example of the received optimization instruction, a web page address is determined based on the optimization instruction, thereby determining a target web page to be optimized. For example, in the process of creating a webpage, a seller user and a merchant user determine that the webpage needs to be optimized, and can send an optimization instruction and the like. The target webpage comprises at least one target object, and the target object is an object capable of being optimized, such as the name of a commodity object, the descriptive information of adjectives and the like.
As shown in fig. 2, a merchant with a native language of chinese issues a commodity on an english e-commerce platform, and because the level of english of the merchant is limited, only the following information is issued: the picture of the commodity, the simple title of the commodity (such as 'red hooodies') and the price, and information such as detailed description of the commodity and warranty is not published. When a user searches for a product by a keyword (for example, "Cactus Jack hat"), even if the target product object itself meets the query condition, the distribution information is too simple to include "Cactus Jack", and therefore the product object cannot be searched. Accordingly, the webpage can be determined as a target webpage to be optimized, wherein the target object can comprise a banner text, a commodity picture and the like.
Step S105: and acquiring the characteristic information of the target object.
The target web page to be optimized may be analyzed to determine feature information of each target object, for example, text analysis may be performed on a title text to identify description information such as nouns, adjectives, and phrases as feature information, and for example, a commodity picture may be identified to analyze corresponding commodity features such as one or more of colors, forms, and materials. As in the example of fig. 2, the article feature information is extracted from the distribution information of the article, and accordingly the following feature information can be extracted from the picture and the simple article title: red, printed letter patterns, long sleeves, hats, cotton, neutral, clothing, etc.
Step S107: and determining the description keywords of which the label information is matched with the characteristic information from the content material library.
The content material library may be matched with the feature information of the target object, for example, the description keyword and the tag information thereof are queried in the content material library based on the feature information of the target object, and the description keyword with the tag information matched with the feature information is determined. In the embodiment of the application, the feature information of the identified target object may correspond to a tag, such as one or more of an application scenario, an application crowd, an application area, application time, an application language, an application type, and an application industry, and may be correspondingly matched with a content material library to determine a description keyword for matching the tag information with the feature information.
For example, the tag information of the material "Cactus Jack" includes "scene { all }; population { all }; region { all }; time { all }; language { all }; type { title, detailed description }, industry { clothing, shoes and hats } ", and the label information may further include" color { all }; sleeve length { all }; pattern { letter, drawing }; the design { cap up, sports pants, down jacket, \ 8230 }; face { cotton } ", and the like. Since the label information of the material "Cactus Jack" and the feature information of fig. 5 match people, if the crowd label thereof includes the feature "neutral", the industry label includes the feature "clothing", the color label includes the feature "red", the sleeve length label includes the feature "long sleeve", the pattern label includes the feature "letter", the style label includes the feature "cap up", and the fabric label includes the feature "cotton", the material "Cactus Jack" is a description keyword matched with the feature information of the commodity shown in fig. 2.
For another example, the tag information of the material "XX meeting franchise" includes "scene { XX meeting }; population { all }; region { all }; time {202110 to 202206}; language { all }; type { title, detailed description }, industry { clothing, footwear, backpack, toy \8230 } ", and the tag information may further include" color { blue and white phase }; pattern { letters and numbers, logo }; styles { hooded jacket, pullover, down jacket, \8230;); sleeve length { all }; face { all } ". Although the crowd label of the material "XX will franchise the good" includes the feature "neutral", the industry label includes the feature "clothing", the sleeve length label includes the feature "long sleeve", the pattern label includes the feature "letter", the style label includes the feature "cap up", and the face label includes the feature "cotton", since the color label does not include the feature "red", the material "XX will franchise the good" is not a description keyword matching the feature information of the good shown in fig. 2.
In the embodiment of the application, one description keyword can correspond to a plurality of commodity objects, and one commodity object can be matched with a plurality of description keywords. So that one or more description keywords can be matched from the content corpus based on the characteristic information of the target object.
Step S109: and generating a key point description text related to the target object based on the description key words matched with the characteristic information, and adding the key point description text into the target webpage so as to optimize the content of the target webpage.
After the description keywords matched with the feature information are determined, the key point description text related to the target object can be generated based on the description keywords, and the key point description text is added into the target webpage, so that the content of the target webpage is optimized, the description of the webpage on the commodity object is enriched, and the user search is facilitated. The natural language model can be trained, description keywords are input into the natural language model, and corresponding key point description texts are output. Some templates and description keywords can be set to synthesize the key point description text.
In order to generate contents quickly, some content templates can be prefabricated, and the content templates provide different content templates according to types, industries and language expressions of different commodity objects. The content template can be generated in advance according to types (topics, detail pages and the like), industries (clothes, electronics, homes and the like), languages (language frame sequence, expression emphasis points) and the like by combining with crawler rules of a search engine. In one example, the content template may be similar to a sentence gap filling style, and based on the word order and the expression emphasis of the corresponding language, description sentences for different industries under different types are designed, and partial blank positions corresponding to the description keywords in the sentences, such as reserved blank positions in the positions of expression color, material, application range, and the like, are formed into the content template.
Generating a key point description text related to the target object based on the description key words matched with the feature information, wherein the key point description text comprises the following steps: matching a content template according to the label information of the description keyword matched with the characteristic information; and synthesizing by adopting the content template and the description keywords to generate a key point description text related to the target object. Matching a content template by adopting label information of the description keywords matched with the characteristic information, such as matching a corresponding content template based on language, type, industry and the like, then adding the description keywords in corresponding blank positions of the content template, and also adding label information of the description keywords to obtain key point description sentences related to corresponding target objects, and adding the key point description sentences into the target webpage to complete the optimization of the target webpage.
In subsequent retrieval, query and other processes, as the key point description sentences are added in the target webpage in an optimized manner, more texts capable of being matched with the retrieval key words exist, the retrieval success rate is improved, the probability of webpage clicking is correspondingly improved, and the exposure is increased.
In the embodiment of the application, after the webpage optimization is completed and released, releasing feedback results corresponding to the optimized target webpage can be collected, wherein the releasing feedback results comprise retrieval result information and/or click information; and updating the content material library according to the release feedback result. The releasing feedback result is a releasing result after webpage optimization is completed and releasing is carried out, and the releasing feedback result can be collected periodically. The retrieval result information is delivery result information related to retrieval, such as the number of times that a web page is retrieved, the number of times that the web page is added to the retrieval result, position or weight information in the retrieval result, and the like, so that whether the web page can be effectively retrieved in the retrieval process can be fed back. The click information is information of a web page clicked, such as the number of times, probability, etc. clicked in the search result. The optimization effect information, the retrieval probability, the success rate, the click probability, the exposure and other data of the webpage can be analyzed based on the release feedback result, the description keywords with higher success rate can be determined by combining other webpages in the corresponding retrieval result of the webpage, and the matching degree can be improved by updating the content material library based on the description keywords. And text translation can be performed based on the description keywords with higher success rate, so that content homogenization is avoided.
On the basis of the above embodiment, the embodiment of the application also provides a method for creating the content material library and a method for optimizing the webpage based on the content material library. As in one optimization example shown in fig. 3. The ideogram content of the commodity object webpage can be obtained through technologies such as OCR recognition of texts, video/picture understanding technology, text crawling and the like. And then, translating and processing the text to obtain a description keyword. And multi-dimensional label information can be determined for the description keywords, and a corresponding content material library is generated. The content material library adopts subject words, key point adjectives, short sentences and the like of commodity objects as description keywords. The key point adjectives are adjectives of key features described by the commodity objects, and if the key points are in capacity and the adjectives are large, the large capacity is formed correspondingly. The steps of the method of creating a library of content material are provided as in fig. 4.
And determining types, industries and language expressions of different commodity objects by combining crawler rules of a search engine to provide different content templates, determining characteristic information based on the title, the picture and other original contents of the target webpage aiming at the target webpage to be optimized, matching a content material library by adopting the characteristic information, and determining description keywords of which the label information is matched with the characteristic information. And synthesizing the description keywords and the content template to optimize the webpage. In one example shown in fig. 5, matching and expansion optimization are performed on an original web page containing only titles and pictures, so that corresponding description keywords can be obtained, and an expanded portion of the web page can be obtained by synthesizing the corresponding description keywords with a content template. The optimization method can be implemented by the steps shown in fig. 6.
The optimized webpage can be launched to websites such as an e-commerce website, the launch feedback result corresponding to the optimized target webpage can be collected periodically subsequently, the content material library is updated based on the launch feedback result, and the webpage is continuously optimized. After the system is put in, the effect can be tracked and analyzed, dynamic searching and matching can be carried out, and the homogenization problem can be avoided.
Step S401, original content related to a commodity object is obtained, wherein the original content includes one or more of webpage content, pictures, videos and knowledge base content related to the commodity object.
Step S403, obtaining a first description keyword of the commodity object from the original content.
Obtaining a first description keyword of the commodity object from the original content, wherein the first description keyword comprises at least one of the following steps: extracting a first description keyword of the commodity object from a webpage or a picture through a character recognition technology; extracting a first description keyword of the commodity object from a video through a video understanding technology; and extracting a first description keyword of the commodity object from the webpage text.
The extracting, by a video understanding technology, the first description keyword of the commodity object from the video includes: identifying interested regions of each frame of image in the video, and determining corresponding interested regions; carrying out object recognition on the region of interest, determining a commodity object and commodity characteristics as corresponding first description keywords, wherein the commodity characteristics comprise: one or more of color, morphology, material; and/or performing action recognition according to the region of interest, and determining a corresponding commodity object and a first description keyword; and analyzing the associated knowledge of the first description keywords, determining associated knowledge information, and adding the associated knowledge information into the first description keywords of the commodity object.
Step S405, perform text translation on the first description keyword to obtain a second description keyword.
The text translation of the first description keyword to obtain a second description keyword includes at least one of the following steps: performing feature analysis on the first description keywords, and determining second description keywords which accord with the description features on the basis of the extracted description features, wherein the second description keywords comprise one or more of alternative names, proper nouns and event names of the first description keywords; and translating the first description keywords according to a second language, and taking the translated text as a second description keyword.
Step S407, using the first description keyword and the second description keyword as description keywords of the commodity object.
Step S409, marking each description keyword, generating label information of the description keyword, and creating a corresponding content element library.
According to the embodiment of the application, the content can be mined through a technical means, various factors such as areas, scenes, events, crowds and languages are considered, the content material library is constructed through expansion and the like, the description of the webpage content is enriched, and a data basis is provided for optimizing the webpage content.
By adopting the processing mode, a material library can be formed according to the auxiliary webpage content related to the target object, and the key point description language in the material library is utilized to optimize the content of the data object webpage.
Step S601, determining a target webpage to be optimized, wherein the target webpage comprises at least one target object.
Step S603, obtaining feature information of the target object.
Step S605, determining the description keyword of which the tag information matches with the feature information from the content material library.
Step S607, matching a content template according to the tag information of the description keyword matched with the feature information.
And step S609, synthesizing the content template and the description keywords to generate a key point description text related to the target object.
Step S611, adding the key point description text to the target webpage to optimize the content of the target webpage.
And content synthesis is carried out by combining the content template, and the crawling characteristics of each search engine are matched, so that the web page is easier to search, and the exposure is improved.
On the basis of the foregoing embodiments, the present application embodiment provides a process flow of optimizing existing webpages of a website, as shown in fig. 7:
step S701, detecting a web page, and determining retrieval key information.
Step S703, determining whether the retrieval key information meets the retrieval requirement.
The retrieval key information is information related to retrieval, such as the number of keywords contained in a webpage, and the retrieval success rate of the webpage history. And setting a threshold value based on the corresponding retrieval key information, wherein if the threshold value is met, the retrieval requirement is met, otherwise, the retrieval requirement is not met.
If yes, the feedback step S701 continues the detection. If not, the step S705 is executed with the webpage as the target webpage.
Step S705, acquiring feature information of the target object.
In step S707, a description keyword whose tag information matches the feature information is determined from the content material library.
Step S709, based on the description keyword matched with the feature information, generating a keypoint description text related to the target object.
Matching a content template according to the label information of the description keywords matched with the characteristic information; and synthesizing by adopting the content template and the description keywords to generate a key point description text related to the target object.
Step S711, adding the key point description text to the target webpage to optimize the content of the target webpage.
Therefore, the web pages in the website can be automatically detected and optimized.
On the basis of the foregoing embodiments, the present application embodiment provides a process flow of optimizing unpinned webpages, as shown in fig. 8:
step S801, receiving an optimization instruction, and determining a target webpage to be optimized according to the optimization instruction, wherein the target webpage comprises at least one target object.
Step S803, feature information of the target object is acquired.
Step S805, determining a description keyword whose tag information matches with the feature information from the content material library.
Step S807, generating a key point description text related to the target object based on the description key word matched with the feature information.
Matching a content template according to the label information of the description keywords matched with the characteristic information; and synthesizing by adopting the content template and the description keywords to generate a key point description text related to the target object.
Step S809, adding the key point description text to the target webpage to optimize the content of the target webpage.
Compared with a manual mode, the method and the device can automatically generate the content material library, automatically optimize the webpage, have high content generation efficiency and low cost, and simultaneously automatically select and synthesize vocabularies and phrases according to different factors (such as country, crowd, google and other crawlers and recording characteristics and the like), improve matching degree and avoid content homogenization.
Second embodiment
In the foregoing embodiment, a method for optimizing web page content is provided, and correspondingly, an apparatus for optimizing web page content is also provided. The device corresponds to the embodiment of the method. Parts of this embodiment that are the same as the first embodiment are not repeated, and please refer to corresponding parts in the first embodiment.
The application provides a webpage content optimizing device, can be applied to in the electronic equipment of server, the device includes:
the system comprises a material library providing module, a content material library providing module and a content database processing module, wherein the content material library comprises description keywords of commodity objects and label information of the description keywords, and the label information of the description keywords comprises one or more of application scenes, application groups, application areas, application time, application languages, application types and application industries;
the webpage determining module is used for determining a target webpage to be optimized, wherein the target webpage comprises at least one target object;
the characteristic matching module is used for acquiring characteristic information of the target object; determining description keywords of which the label information is matched with the feature information from the content material library;
and the webpage optimization module is used for generating a key point description text related to the target object based on the description key words matched with the characteristic information, and adding the key point description text into the target webpage so as to optimize the content of the target webpage.
In summary, a content material library is provided, the content material library includes description keywords of commodity objects and tag information of the description keywords, the tag information of the description keywords includes one or more of an application scene, an application crowd, an application area, application time, an application language, an application type and an application industry, so that a rich data base of wind can be provided for webpage optimization, the quality of optimized content is improved, then a target webpage to be optimized is determined, the target webpage includes at least one target object, the feature information of the target object is obtained, the description keywords matching the tag information and the feature information are determined from the content material library, a key point description text related to the target object is generated based on the description keywords matching the feature information, the key point description text is added into the target webpage to optimize the target webpage content, the description keywords can be automatically matched, and the key point description text for webpage optimization is generated, so that the webpage content is richer, retrieval is facilitated, and the efficiency and the quality of search engine are effectively improved. (ii) a Therefore, the optimization efficiency and the optimization quality of the search engine can be effectively improved.
The material library providing module is used for acquiring original contents related to the commodity object, wherein the original contents comprise one or more of webpage contents, pictures, videos and knowledge library contents related to the commodity object; acquiring a first description keyword of the commodity object from the original content; taking the first description keyword as a description keyword of the commodity object; marking each description keyword to generate label information of the description keyword.
The material library providing module is used for performing text translation on the first description keywords to obtain second description keywords; and taking the second description keyword as the description keyword of the commodity object.
The material library providing module is used for carrying out feature analysis on the first description keywords and determining second description keywords which accord with the description features on the basis of the extracted description features, wherein the second description keywords comprise one or more of alternative names, proper nouns and event names of the first description keywords; and translating the first description keywords according to a second language, and taking the translated text as second description keywords.
The material library providing module is used for extracting a first description keyword of the commodity object from a webpage or a picture through a character recognition technology; extracting a first description keyword of the commodity object from a video through a video understanding technology; and extracting a first description keyword of the commodity object from the webpage text.
The material library providing module is used for identifying interested regions of each frame of image in the video and determining corresponding interested regions; carrying out object recognition on the region of interest, determining a commodity object and commodity characteristics as corresponding first description keywords, wherein the commodity characteristics comprise: one or more of color, morphology, material; and/or performing action recognition according to the region of interest, and determining a corresponding commodity object and a first description keyword; and analyzing the associated knowledge of the first description keywords, determining associated knowledge information, and adding the associated knowledge information into the first description keywords of the commodity object.
Further comprising:
the template providing module is used for providing a content template, and the content template provides different content templates according to the types, industries and language expressions of different commodity objects;
the webpage optimization module is used for matching a content template according to the label information of the description keyword matched with the characteristic information; and synthesizing by adopting the content template and the description keywords to generate a key point description text related to the target object.
Further comprising:
the feedback updating module is used for collecting the release feedback result corresponding to the optimized target webpage, and the release feedback result retrieves result information and/or click information; and updating the content material library according to the release feedback result.
The webpage determining module is used for detecting the webpage and determining retrieval key information; and under the condition that the retrieval key information meets the retrieval requirement, taking the webpage as a target webpage to be optimized, wherein the target webpage comprises at least one target object.
The target webpage determining module is used for receiving an optimization instruction, determining a target webpage to be optimized according to the optimization instruction, detecting the webpage by at least one target object in the target webpage, and determining retrieval key information
The application provides a device for creating a content material library, which can be applied to electronic equipment of a server side, and comprises:
the system comprises a content acquisition module, a database acquisition module and a display module, wherein the content acquisition module is used for acquiring original content related to a commodity object, and the original content comprises one or more of webpage content, pictures, videos and knowledge base content related to the commodity object;
the keyword determining module is used for acquiring a first description keyword of the commodity object from the original content; taking the first description keyword as a description keyword of the commodity object;
the label determining module is used for marking each description keyword to generate label information of the description keyword, wherein the label information of the description keyword comprises one or more of an applicable scene, an applicable crowd, an applicable area, an applicable time, an applicable language, an applicable type and an applicable industry; and generating a content material library to determine description keywords of which the tag information is matched with the feature information of the target object of the target webpage, generating a key point description text related to the target object based on the matched description keywords, adding the key point description text into the target webpage, and optimizing the content of the target webpage.
The keyword determining module is used for extracting a first description keyword of the commodity object from a webpage or a picture through a character recognition technology; extracting a first description keyword of the commodity object from a video through a video understanding technology; and extracting a first description keyword of the commodity object from the webpage text.
The keyword determining module is used for identifying the region of interest of each frame of image in the video and determining the corresponding region of interest; carrying out object recognition on the region of interest, determining a commodity object and commodity characteristics as corresponding first description keywords, wherein the commodity characteristics comprise: one or more of color, morphology, material; and/or performing action recognition according to the region of interest, and determining a corresponding commodity object and a first description keyword; and analyzing the associated knowledge of the first description keywords, determining associated knowledge information, and adding the associated knowledge information into the first description keywords of the commodity object.
The keyword determining module is used for extracting a first description keyword of the commodity object from a webpage or a picture through a character recognition technology; extracting a first description keyword of the commodity object from a video through a video understanding technology; and extracting a first description keyword of the commodity object from the webpage text.
The keyword determining module is used for identifying the region of interest of each frame of image in the video and determining the corresponding region of interest; carrying out object recognition on the region of interest, determining a commodity object and commodity characteristics as corresponding first description keywords, wherein the commodity characteristics comprise: one or more of color, morphology, material; and/or performing action recognition according to the region of interest, and determining a corresponding commodity object and a first description keyword; and analyzing the associated knowledge of the first description keywords, determining associated knowledge information, and adding the associated knowledge information into the first description keywords of the commodity object.
The optimized webpage can be launched to websites such as an e-commerce website, the launch feedback result corresponding to the optimized target webpage can be collected periodically subsequently, the content material library is updated based on the launch feedback result, and the webpage is continuously optimized. After the system is put in, the effect can be tracked and analyzed, dynamic searching and matching can be carried out, and the homogenization problem can be avoided.
According to the embodiment of the application, the content can be mined through a technical means, various factors such as areas, scenes, events, crowds and languages are considered, the content material library is constructed through expansion and the like, the description of the webpage content is enriched, and a data basis is provided for optimizing the webpage content.
By adopting the processing mode, a material library can be formed according to the auxiliary webpage content related to the target object, and the key point description language in the material library is utilized to optimize the content of the data object webpage.
Third embodiment
The application also provides an electronic device. Since the apparatus embodiments are substantially similar to the method embodiments, they are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for relevant points. The device embodiments described below are merely illustrative.
An electronic device of the present embodiment includes: a processor and a memory; a memory for storing a program for implementing the web content optimization method, the apparatus being powered on and the program for the method being executed by the processor.
The present application also provides a computer-readable storage medium having stored therein instructions, which when run on a computer, cause the computer to perform the various methods described above.
The present application also provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform the various methods described above.
Although the present application has been described with reference to the preferred embodiments, it is not intended to limit the present application, and those skilled in the art can make variations and modifications without departing from the spirit and scope of the present application, therefore, the scope of the present application should be determined by the claims that follow.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
1. Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include non-transitory computer readable media (transient media), such as modulated data signals and carrier waves.
2. As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.