CN110020132B - Keyword recommendation method and device, computing equipment and storage medium - Google Patents
Keyword recommendation method and device, computing equipment and storage medium Download PDFInfo
- Publication number
- CN110020132B CN110020132B CN201711068460.8A CN201711068460A CN110020132B CN 110020132 B CN110020132 B CN 110020132B CN 201711068460 A CN201711068460 A CN 201711068460A CN 110020132 B CN110020132 B CN 110020132B
- Authority
- CN
- China
- Prior art keywords
- keyword
- text
- keywords
- recommended word
- click
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a keyword recommendation method, a keyword recommendation device, a computing device and a storage medium. The keyword recommendation method comprises the following steps: acquiring a keyword set; determining a frequency parameter value corresponding to each keyword in a keyword set; and sequencing all the keywords in the keyword set according to the frequency parameter value corresponding to each keyword, extracting the keywords related to the preset putting effect from the corresponding sequencing result, and taking the keywords as a recommended word set.
Description
Technical Field
The present application relates to the field of information delivery, and in particular, to a keyword recommendation method and apparatus, a computing device, and a storage medium.
Background
When a user accesses the network via the terminal device, the pages displayed by the terminal device are typically also presented with additional content (e.g., advertisements, news, etc.). The expression words of the additional content (such as the keywords therein) directly influence the interest level of the user in the additional content.
Disclosure of Invention
According to one aspect of the application, a keyword recommendation method is provided, which includes: acquiring a keyword set; determining a frequency parameter value corresponding to each keyword in a keyword set, wherein the frequency parameter value corresponding to each keyword is used for describing the use range of the keyword; and sequencing the keywords in the keyword set according to the frequency parameter value corresponding to each keyword, extracting the keywords related to the preset putting effect from the corresponding sequencing result, and taking the keywords as a recommended word set.
According to still another aspect of the present application, there is provided a method of presenting a recommended word, including: acquiring a plurality of recommended word sets generated according to the keyword recommendation method, wherein the attribute value of each recommended word set comprises an industry category and a delivery channel; responding to the operation of selecting the recommended word set, and providing a user interface for presenting attribute options corresponding to the recommended word sets, wherein the attribute options comprise industry categories and release channels; and in response to the selection operation of the attribute option in the user interface, determining and presenting a recommended word set corresponding to the selected attribute value so as to prompt the user to use keywords in the recommended word set when editing the text.
According to still another aspect of the present application, there is provided a keyword recommendation apparatus including: the device comprises an acquisition unit, a frequency determination unit and a screening unit. The acquisition unit is used for acquiring a keyword set. The frequency determining unit is used for determining a frequency parameter value corresponding to each keyword in the keyword set, and the frequency parameter value corresponding to each keyword is used for describing the use width of the keyword. And the screening unit is used for sequencing the keywords in the keyword set according to the frequency parameter value corresponding to each keyword, extracting the keywords related to the preset putting effect from the corresponding sequencing result and taking the keywords as a recommended word set.
In some embodiments, the obtaining unit further comprises a text obtaining module and a keyword generating module. The text acquisition module is used for acquiring a text set containing a plurality of texts, wherein each text is used for describing additional content which is used for being presented in a page, and acquiring a release effect record of the additional content corresponding to each text. The keyword generation module is used for performing word segmentation operation on each text in the text set respectively, acquiring keywords corresponding to each text and generating a keyword set corresponding to the text set. The keyword set comprises keywords corresponding to each text.
In some embodiments, the text acquisition module acquires a text collection comprising a plurality of texts according to the following manner: the method comprises the steps of obtaining a plurality of additional contents which belong to a preset industry category and are released to a preset release channel, and generating a text set, wherein the text set comprises texts corresponding to the additional contents in the additional contents.
In some embodiments, the frequency determining unit determines the frequency parameter value corresponding to each keyword in the keyword set according to the following manner: determining a text frequency of each keyword in the set of keywords, wherein the text frequency of each keyword represents a total number of texts containing the keyword in the set of texts; calculating an importance parameter value of each keyword in the keyword set to the keyword set; and performing weighted calculation on the text frequency and the importance parameter value of each keyword, and taking the corresponding calculation result as the frequency parameter value of the keyword.
In some embodiments, the frequency determination unit calculates the importance parameter value of each keyword in the keyword set for the keyword set according to the following manner: constructing a network graph comprising a plurality of nodes, wherein each node in the network graph represents a keyword in the keyword set, two nodes contained in any text in the text set in the network graph are connected, and the correlation parameter value of each pair of connected two nodes in the network graph corresponds to the number of texts containing the two nodes in the text set; and calculating the weight value of each node in the network graph according to a text arrangement algorithm, and taking the weight value as the importance parameter value of the node.
In some embodiments, the corresponding delivery effect record of each text includes an exposure record and a click record corresponding to the text, and the screening unit extracts a keyword related to the predetermined delivery effect from the corresponding ranking result as a recommended word set according to the following manner: extracting a plurality of keywords from the sequencing result according to the sequence of the frequency parameter values from high to low, and taking the keywords as a first screening set; calculating the exposure support degree of each keyword in the first screening set, wherein the exposure support degree represents the ratio of the exposure times corresponding to the texts containing the keyword in the text set to the total exposure times corresponding to all the texts in the text set; extracting keywords with exposure support higher than a first threshold value in the first screening set, and taking the extracted keywords as a second screening set; determining the click rate corresponding to each text according to the exposure record and the click record corresponding to each text in the text set; calculating the exposure confidence coefficient of each keyword in the second screening set, wherein the exposure confidence coefficient of each keyword represents the ratio of the exposure times of the text which contains the keyword in the text set and has the click rate higher than a second threshold value to the total exposure times of the text which contains the keyword in the text set; and sequencing the keywords in the second screening set according to the exposure confidence of each keyword in the second screening set, extracting a plurality of keywords from corresponding sequencing results according to the sequence of the exposure confidence from high to low, and taking the extracted keywords as the recommended word set.
In some embodiments, the screening unit is further configured to extract, according to the delivery effect record corresponding to each text in the text set, a keyword for reducing the delivery effect from the ranking result of the keyword set, and use the keyword as the non-recommended word set.
In some embodiments, the screening unit is further configured to extract a plurality of keywords that do not belong to the recommended word set from the ranking results corresponding to the second screening set, and use the extracted keywords as a non-recommended word set.
In some embodiments, the corresponding delivery effect record of each text includes a click record and a conversion record corresponding to the text, and the screening unit extracts a keyword related to the predetermined delivery effect from the corresponding ranking result as a recommended word set according to the following manner: extracting a plurality of keywords from the sequencing result according to the sequence of the frequency parameter values from high to low, and taking the keywords as a first screening set; calculating the click support degree of each keyword in the first screening set, wherein the click support degree represents the ratio of the click times corresponding to texts containing the keyword in the text set to the total click times corresponding to all texts in the text set; extracting key words with click support higher than a third threshold value in the first screening set, and taking the extracted key words as a third screening set; determining a conversion rate corresponding to each text according to the click record and the conversion record corresponding to each text in the text set; calculating the click confidence of each keyword in the third screening set, wherein the click confidence of each keyword represents the ratio of the number of clicks of the text which contains the keyword in the text set and has the conversion rate higher than a fourth threshold value to the total number of clicks of the text which contains the keyword in the text set; and sequencing the keywords in the third screening set according to the click confidence of each keyword in the third screening set, extracting a plurality of keywords from corresponding sequencing results according to the sequence of the click confidence from high to low, and taking the extracted keywords as the recommended word set.
According to still another aspect of the present application, there is provided an apparatus for presenting a recommended word, including: the system comprises an acquisition unit, a recommendation unit and a recommendation unit, wherein the acquisition unit is used for acquiring a plurality of recommendation word sets generated according to the keyword recommendation method of the application, and the attribute value of each recommendation word set comprises an industry category and a delivery channel; the selection unit responds to the operation of selecting the recommended word set and provides a user interface for presenting attribute options corresponding to the recommended word sets, wherein the attribute options comprise industry categories and delivery channels; and the presentation unit is used for responding to the selection operation of the attribute option in the user interface, determining and presenting the recommended word set corresponding to the selected attribute value so as to prompt the user to use the key words in the recommended word set when editing the text.
In some embodiments, the presentation unit is further configured to determine and present a set of non-recommended words corresponding to the selected attribute value, so as to prompt a user to avoid using keywords in the set of non-recommended words when editing the text.
According to yet another aspect of the present application, there is provided a computing device comprising: one or more processors, memory, and one or more programs. A program is stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing the keyword recommendation method of the present application.
According to still another aspect of the present application, there is provided a storage medium storing one or more programs. The one or more programs include instructions. The instructions, when executed by a computing device, cause the computing device to perform the keyword recommendation method of the present application.
In summary, according to the keyword recommendation scheme of the present application, by accurately ranking the frequency (i.e., the usage breadth) of each keyword in the keyword set (for example, ranking based on the weighting result of the text frequency and the importance parameter value), the first filtering set with high usage frequency can be obtained. On the basis, the scheme of the application can screen the keywords in the first screening set (for example, a method based on exposure support and exposure confidence or a method based on click support and click confidence), and can obtain high-quality recommended words. In this way, the scheme of the application can obtain the recommended word set capable of improving the interest degree (the interest degree can be expressed as the click probability or the conversion probability of the user on the advertisement, for example) of the user on the additional content (such as the advertisement, and the like).
Drawings
In order to more clearly illustrate the technical solutions in the examples of the present application, the drawings needed to be used in the description of the examples are briefly introduced below, and it is obvious that the drawings in the following description are only some examples of the present application, and it is obvious for a person skilled in the art to obtain other drawings based on these drawings without inventive exercise.
FIG. 1 illustrates a schematic diagram of an application scenario 100, in accordance with some embodiments of the present application;
FIG. 2 illustrates a flow diagram of a keyword recommendation method 200 according to some embodiments of the present application;
FIG. 3 illustrates a flow diagram of a method 300 of calculating a keyword frequency parameter value according to one embodiment of the present application;
FIG. 4 shows a schematic diagram of a network diagram 400 according to one embodiment of the present application;
FIG. 5 is a diagram illustrating a method 500 of obtaining a set of recommended words according to one embodiment of the application;
FIG. 6 is a diagram illustrating a method 600 of obtaining a set of recommended words according to one embodiment of the present application;
FIG. 7 illustrates a flow diagram of a keyword recommendation method 700 according to some embodiments of the present application;
FIG. 8 illustrates a flow diagram of a method 800 of presenting recommended words, according to some embodiments of the present application;
FIG. 9A illustrates a user page view of a selected set of recommended words according to one embodiment of the present application;
FIG. 9B illustrates a user interface diagram presenting a set of recommended words, according to one embodiment of the present application;
FIG. 10 illustrates a flow diagram of a method 1000 of presenting recommended words according to some embodiments of the present application;
FIG. 11A illustrates a schematic diagram of a keyword recommendation apparatus 1100 in accordance with some embodiments of the present application;
FIG. 11B shows a schematic diagram of an acquisition unit 1110 according to one embodiment of the present application;
FIG. 12 illustrates a schematic diagram of an apparatus 1200 for presenting recommended words according to some embodiments of the present application; and
FIG. 13 illustrates a block diagram of the components of a computing device.
Detailed Description
The technical solutions in the examples of the present application will be clearly and completely described below with reference to the drawings in the examples of the present application, and it is obvious that the described examples are only a part of the examples of the present application, and not all examples. All other examples, which can be obtained by a person skilled in the art without making any inventive step based on the examples in this application, are within the scope of protection of this application.
In some embodiments, when an application in the terminal device accesses the network (i.e., the user surfs the web), the terminal device may display the page accessed by the user. Here, the terminal device is, for example, various computing devices such as a desktop computer, a notebook computer, a tablet computer, a mobile phone, and a handheld game machine. The terminal device is applied to various software such as a browser, weChat, QQ, microblog and the like. In addition, the accessed page may also typically display additional content. The additional content is various contents such as news information and advertisement. Here, the additional content may be in the form of various media such as still pictures, moving pictures, text messages, and videos.
Fig. 1 illustrates a schematic diagram of an application scenario 100 according to some embodiments of the present application. As shown in fig. 1, the application scenario 100 includes a storage device 110, a computing device 120, and a computing device 130. Here, the storage device 110 may be, for example, a database server, but is not limited thereto. The storage device 110 stores text 111 and impression effect records 112. Text collection 111 may include a plurality of texts. Each text is used to describe an additional content for additional presentation in the page. In other words, the text is a literal representation of the additional content. Taking an advertisement as an example, one text may be an advertisement copy corresponding to one advertisement. In one embodiment, the impression effect record 112 includes an exposure record and a click record. The exposure record may include an exposure record of the additional content (e.g., an advertisement) corresponding to each text in text set 111. Here, the additional content may be considered to complete one exposure each time it is displayed in the page. The click record may include multiple records. Each record is used to record a user click event for an additional piece of content. In yet another embodiment, impression effect records 112 may include click records and conversion records. Here, each conversion record is used to record a conversion event. Each conversion event is, for example, the user clicking on an additional content and performing the desired action. The expected operation is, for example, registering an account, ordering a commodity, downloading and logging in an application to be promoted, and the like, which is not limited in the present application. Based on text 111 and impression effect record 112, computing device 120 may generate a set of recommended words 121. Here, the recommended word set 121 is used to be applied to a text corresponding to the additional content, for example, to an advertisement case, but is not limited thereto. In other words, the keywords in the recommended word set 121 are considered to improve the user's interest level in the additional content. The computing device 130 may obtain the set of recommended words 121 from the computing device 120 for the user to edit text (e.g., edit an advertising copy, etc.) from the set of recommended words 121. Here, the computing devices 120 and 130 may be implemented as various devices such as a hardware independent server, a distributed application, a personal computer, and the like, for example, and the present application is not limited thereto. The process of generating the set of recommended words is described below with reference to fig. 2.
FIG. 2 illustrates a flow diagram of a keyword recommendation method 200 according to some embodiments of the present application. The method 200 may be performed, for example, in an application that generates recommended words. The application that generates the recommended word may reside, for example, but is not limited to, in computing device 120. Here, the computing device 120 is, for example, an advertisement material server, a news material server, or the like.
The method 200 includes step S201. In step S201, a keyword set is acquired.
In one embodiment, each keyword in the set is used to edit additional content presented in the page. Specifically, in step S201, a text set including a plurality of texts is first acquired, and a delivery effect record of information corresponding to each text is acquired. Wherein each text is used to describe an additional content for additional presentation in the page.
In one embodiment, step S201 may obtain a text set and a delivery effect record from the storage device 110, for example. In yet another embodiment, step S201 may obtain a plurality of texts (e.g., advertisement copy) from a text server (e.g., an advertisement server) and treat them as a text set. In addition, step S201 may extract a release effect record corresponding to each text from the weblog. In yet another embodiment, considering that the sets of recommended words required for the additional contents of different business types (e.g., IT products, real estate, apparel, delicacies, etc.) and delivery channels (e.g., news client, video client, instant messenger application, etc.) are different, step S201 may obtain a plurality of additional contents (e.g., a plurality of advertisements) belonging to a predetermined industry category and used for delivery according to a predetermined delivery channel, and take the text corresponding to each information as the text set.
When determining the text set, step S201 may further perform word segmentation on each text in the text set, acquire a keyword corresponding to each text, and generate a keyword set corresponding to the text set, where the keyword set includes the keyword corresponding to each text. Here, the word segmentation method is, for example, a Chinese Lexical Analysis System (abbreviated as ICTCLAS) or a text mining algorithm (text miner) based on a multi-layer hidden horse model, and the present application does not limit this.
For the keyword set obtained in step S201, the method 200 may execute step S202 to determine a frequency parameter value corresponding to each keyword, where the frequency parameter value corresponding to each keyword is used to describe the usage breadth of the keyword. Here, the degree of use is, for example, a degree to which the keyword is commonly used in a text collection. In some embodiments, step S202 may be implemented as method 300 shown in fig. 3.
As shown in fig. 3, the method 300 includes step S301 and step S302. In step S301, the text frequency (document frequency) of each keyword in the keyword set is determined. Wherein the text frequency of each keyword represents the total number of texts containing the keyword in the text collection.
In step S302, an importance parameter value of each keyword in the keyword set to the keyword set is calculated. In one embodiment, step S302 may first construct a network graph including a plurality of nodes, where each node in the network graph represents a keyword in the set of keywords. Two nodes of any text in the text set in the network graph are connected. The relevance parameter value of each pair of two connected nodes in the network graph corresponds to the number of texts in the text set containing the two nodes. For example, fig. 4 shows a schematic diagram of a network diagram 400 according to one embodiment of the present application. As shown in fig. 4, the network graph 400 includes nodes w1 to w 9. c1 to c10 each represent a correlation parameter value. Here, each node in the network graph 400 represents a keyword in the set of keywords. For example, nodes w1 and w5 are connected, and the correlation parameter value of the two nodes is c 1. The value of c1 is proportional to the number of texts in the text set that contain both nodes w1 and w 5.
On the basis of the network graph, step S302 may calculate a weight value of each node in the network graph according to a text rank (text rank) algorithm, and use the weight value as an importance parameter value of the node. In one embodiment, step S302 may determine the importance parameter value of each node (i.e., each keyword) according to the following formula.
Wherein, V i 、V j And V k Each representing a node in the network graph. d is a damping coefficient, and the value range is [0.5,0.9 ] for example]。S(V i ) Representation node V i S (V) is the importance parameter value of j ) Representation node V j The importance parameter value of. In (V) i ) Representing a junction point V in a network graph i Set of connected nodes, out (V) j ) Representing a junction point V in a network graph j A collection of connected nodes. Omega jk Represents node V k And V j Value of the correlation parameter of (c), ω ji Representation node V i And V j The correlation parameter value of (2). Here, step S302 may iteratively calculate the importance parameter value of each keyword based on the above formula until the importance parameter value converges (e.g., the difference between the latest two calculated importance parameter values does not exceed the convergence threshold).
Upon determining the importance parameter value for each keyword in the set of keywords in step S302, the method 300 may perform step S303. In step S303, the text frequency and the importance parameter value of each keyword are weighted and calculated, and the corresponding calculation result is used as the frequency parameter value of the keyword. In summary, the method 300 can greatly avoid the situation that the frequency parameter values of a plurality of keywords are the same by performing weighted calculation on the text frequency and the importance parameter value of the keyword, so that the frequency parameter values of each keyword can be accurately distinguished.
When determining the frequency parameter value corresponding to each keyword in step S202, the method 200 may perform step S203. In step S203, the keywords in the keyword set are ranked according to the frequency parameter value corresponding to each keyword, and the keywords related to the predetermined delivery effect are extracted from the corresponding ranking result and are used as the recommended word set. According to one embodiment, in step S203, a plurality of keywords are extracted from the ranking results as a first filtering set. For example, step S203 may extract a keyword having a frequency parameter value higher than a frequency threshold from the ranking result as a first filtering set. For another example, step S203 may extract a plurality of keywords (for example, a first predetermined number of keywords) in the order of the frequency parameter values from high to low, and use them as a first filtering set. Here, step S203 may also extract a plurality of top-ranked keywords from the ranking result in other suitable manners, and use them as the first filtering set. On this basis, step S203 may extract keywords related to the predetermined delivery effect from the first filtering set, and use them as a recommended word set. In an embodiment, in step S203, a keyword related to a predetermined delivery effect may be extracted from the first filtering set according to the delivery effect record corresponding to each text in the text set, and the keyword is used as a recommended word set. Here, the purpose of acquiring the set of recommended words is to acquire a degree of interest (which may be expressed as, for example, a probability of clicking on an advertisement by a user, a conversion rate, or the like) that can increase the user's interest in additional content corresponding to the text (e.g., an advertisement corresponding to an advertisement document, or the like). The extraction operation performed in step S203 may be various ways for determining the keyword capable of increasing the interest level, which is not limited in the present application.
In some embodiments according to the present application, the delivery effect records include exposure records and click records. The exposure record and the click record may be extracted from, for example, a web log within a predetermined time range (e.g., the last month). In these embodiments, the process of determining the set of recommended words from the first filtered set in step S203 may be implemented as method 500.
As shown in fig. 5, the method 500 includes a step S501 of calculating an exposure support for each keyword in the first filtering set. Here, the exposure support degree indicates a ratio of the number of exposures corresponding to the text including the keyword in the text set to the total number of exposures corresponding to all the texts in the text set.
In step S502, keywords in the first filtering set whose exposure support degree is higher than the first threshold are extracted, and the extracted keywords are taken as the second filtering set. By screening keywords for exposure support, the method 500 may filter out keywords (i.e., keywords deemed to be statistically insignificant) with low exposure support (i.e., below a first threshold).
In step S503, a click rate corresponding to each text in the text set is determined according to the exposure record and the click record corresponding to the text. Here, the click rate corresponding to one text refers to a ratio of the number of click records corresponding to the text (i.e., the number of clicks of the additional content corresponding to the text by the user) to the number of exposure records corresponding to the text (i.e., the number of exposures of the additional content corresponding to the text). Note that the click rate data of the text may be acquired in step S201 when the text set is acquired. When the click rate corresponding to each text is acquired in step S201, step S503 may be omitted.
In step S504, the exposure confidence of each keyword in the second filtering set is calculated. The exposure confidence of each keyword represents the ratio of the exposure times of the text containing the keyword in the text set and having the click rate higher than the second threshold to the total exposure times of the text containing the keyword in the text set. Here, the second threshold value is, for example, 3%, but is not limited thereto. The effects of exposure support and exposure confidence are exemplified below. Assume that there are 2 keywords: first-order purchase and air pavilion. The first keyword is exposed for 1 million times, the exposure support degree is one thousandth, and the exposure confidence coefficient is 0.75. The second word is exposed 1 time, the exposure support degree is one million, and the exposure confidence degree is 1.0. The second keyword, although having a very high exposure confidence (i.e., 100%), has a low exposure support. Therefore, the clicking operation of the user on the word may be a wrong click (which may also be referred to as a accidental click), and the clicking action that is not performed by the word to arouse the interest of the user, that is, the second word has no statistical significance because the data amount of the exposure is too small.
In step S505, the keywords in the second filtering set are ranked according to the exposure confidence of each keyword in the second filtering set, and multiple keywords are extracted from the corresponding ranking result, and the extracted keywords are used as the recommended word set. For example, step S505 may extract a keyword having an exposure confidence higher than an exposure confidence threshold from the ranking result, to serve as the recommended word set. For another example, step S505 may extract a plurality of keywords (for example, a second predetermined number of keywords) in the order from high to low of the exposure confidence, and use them as the recommended word set. It should be appreciated that the higher the exposure confidence, the more interesting the keyword is to the user in the additional content (e.g., advertisements or news). Therefore, by sorting by exposure confidence, step S505 can acquire high-quality recommended words (i.e., keywords that can increase the degree of interest of the user).
In still other embodiments according to the present application, the impression effect record includes a click record and a conversion record. The click record and the conversion record may be extracted, for example, from the weblog within a predetermined time range (e.g., the last month). In these embodiments, the process of determining the set of recommended words from the first filtering set in step S203 may be implemented as the method 600.
In step S601, the click support degree of each keyword in the first filtering set is calculated. Here, the click support degree indicates a ratio of the number of clicks corresponding to the text including the keyword in the text set to the total number of clicks corresponding to all the texts in the text set.
In step S602, a keyword in the first filtering set whose click support degree is higher than a third threshold value is extracted, and the extracted keyword is taken as a third filtering set.
In step S603, a conversion rate corresponding to each text in the text set is determined according to the click record and the conversion record corresponding to the text. Here, the conversion rate corresponding to one text is a ratio of the number of conversion records corresponding to the text (i.e., the number of times that the user converts the additional content corresponding to the text) to the number of click records corresponding to the text (i.e., the number of times that the user clicks the additional content corresponding to the text).
In step S604, the click confidence of each keyword in the third filtering set is calculated. And the click confidence of each keyword represents the ratio of the number of clicks of the text which contains the keyword in the text set and has the conversion rate higher than a fourth threshold to the total number of clicks of the text which contains the keyword in the text set.
In step S605, the keywords in the third filtering set are sorted according to the click confidence of each keyword in the third filtering set, and a plurality of keywords are extracted from the corresponding sorting result, and the extracted keywords are used as the recommended word set. For example, step S605 may extract a keyword having a click confidence higher than a click confidence threshold from the ranking result, and use it as a recommended word set. For another example, in step S605, a plurality of keywords (for example, a third predetermined number of keywords) may be extracted in the order from high click confidence to low click confidence, and the extracted keywords may be used as the recommended word set.
In summary, according to the method 200 of the embodiment of the present application, by accurately ranking the frequency (i.e. the usage breadth) of each keyword in the keyword set (for example, ranking based on the text frequency and the weighting result of the importance parameter value), the first filtering set with high usage frequency can be obtained. On the basis, the method 200 may filter the keywords in the first filtering set (for example, the method 500 based on the exposure support and the exposure confidence or the method 600 based on the click support and the click confidence), and may obtain the high-quality recommended word. In this way, the method 200 of the present application may obtain a set of recommended words capable of improving a user's interest level in additional content (e.g., advertisement, etc.) (the interest level may be expressed as a user's click probability or conversion rate on the advertisement, etc.).
FIG. 7 illustrates a flow diagram of a keyword recommendation method 700 according to some embodiments of the application. Method 700 may be performed, for example, in computing device 120, but is not limited to such. As shown in fig. 7, the method 700 may include steps S701 to S703. Here, the embodiments of steps S701 to S703 are respectively consistent with steps S201 to S203, and are not described herein again.
In some embodiments, the method 700 may further include step S704, extracting a keyword for reducing the delivery effect from the ranking result according to the frequency parameter value, and using the keyword as the non-recommended word set. Specifically, in one embodiment, step S704 may extract a plurality of keywords in the ranking result that do not belong to the first filtering set, and use them as the non-recommended word set. In another embodiment, step S704 may extract a plurality of keywords that do not belong to the recommended word set from the ranking results corresponding to the second filtering set, and use the extracted keywords as the non-recommended word set. Here, each keyword in the non-recommended word set may be regarded as a term having a negative influence on increasing the degree of interest of the user in the additional content. Based on the non-recommended word set, when the user edits text (such as an advertisement case) some words which are not beneficial to improving the interest level can be avoided. In another embodiment, step S704 may extract a plurality of keywords that do not belong to the recommended word set from the ranking results corresponding to the third filtering set, and use the extracted keywords as the non-recommended word set.
FIG. 8 illustrates a flow diagram of a method 800 of presenting recommended words according to some embodiments of the application. Here, the method 800 may be performed, for example, in an application presenting recommended words. Here, the application presenting the recommended word may reside in the computing device 130, for example, but is not limited thereto.
As shown in fig. 8, in step S801, a plurality of recommended word sets are acquired. Here, the attribute value of each recommended word set includes an industry category and a delivery channel. Each set of recommended words may be obtained, for example, by the method 200 or the method 700 according to the present application.
In step S802, in response to an operation of selecting a recommended word set, a user interface presenting attribute options corresponding to a plurality of recommended word sets is provided. The attribute options include industry categories and delivery channels.
In step S803, in response to a selection operation of an attribute option in the user interface, a recommended word set corresponding to the selected attribute value is determined and presented so as to prompt the user to use a keyword in the recommended word set when editing text. For example, FIG. 9A illustrates a user page view of a selected set of recommended words according to one embodiment of the present application. The user can select the attribute values of the industry category and the release channel in the user interface. In this way, step S803 may determine a recommended word set to be presented according to a selection operation of the user. Here, the presented set of recommended words may be, for example, grouped and ordered by the pinyin initials corresponding to the keywords therein, but is not limited thereto. FIG. 9B illustrates a user interface diagram presenting a set of recommended words, according to one embodiment of the present application. As shown in fig. 9B, the area B is a text (e.g., an advertisement pattern) editing area, and the area a is a presentation area of the recommended word set. The user can edit the text in the area b according to the recommended word in the area a. Here, in addition to the grouping by keyword initials, the embodiment of the present application may display the recommended word set in the area a in other various applicable presentation manners.
FIG. 10 illustrates a flow diagram of a method 1000 of presenting recommended words according to some embodiments of the present application. Here, the method 1000 may be performed, for example, in an application presenting recommended words. Here, the application presenting the recommended word may reside in the computing device 130, for example, but is not limited thereto.
As shown in fig. 10, the method 1000 includes steps S1001 to S1003. Here, the embodiments of steps S1001 to S1003 correspond to steps S801 to S803, respectively, and are not described here again. In addition, the method 1000 further includes step S1004. In step S1004, a non-recommended word set corresponding to the selected attribute value is determined and presented so as to prompt the user to avoid using a keyword in the non-recommended word set when editing the text.
FIG. 11A illustrates a schematic diagram of a keyword recommendation device 1100 according to some embodiments of the present application. The apparatus 1100 may reside, for example, in the computing device 120. As shown in fig. 11A, the apparatus 1100 includes an acquisition unit 1110, a frequency determination unit 1120, and a filtering unit 1130.
The acquiring unit 1110 is configured to acquire a keyword set. Wherein each keyword in the set is used to edit additional content presented in the page. In one embodiment, the obtaining unit 1110 may be implemented as the structure shown in fig. 11B.
As shown in fig. 11B, the acquisition unit 1110 may include a text acquisition module 1111 and a keyword generation module 1112. The text acquisition module 1111 may acquire a text set including a plurality of texts. And each text is used for describing one additional content for presenting in the page, and acquiring a delivery effect record of the additional content corresponding to each text. In one embodiment, the text obtaining module 1111 may obtain a plurality of additional contents belonging to a predetermined industry category and delivered to a predetermined delivery channel, and generate a text set. The text set includes text corresponding to each of the plurality of additional contents.
The keyword generation module 1112 is configured to perform word segmentation on each text in the text set, acquire a keyword corresponding to each text, and generate a keyword set corresponding to the text set. The keyword set comprises keywords corresponding to each text.
The frequency determining unit 1120 is configured to determine a frequency parameter value corresponding to each keyword in the keyword set. The frequency parameter value corresponding to each keyword is used for describing the using popularity of the keyword, such as the using popularity in the text set. In some embodiments, the frequency determination unit 1120 may determine a text frequency for each keyword in the set of keywords. Wherein the text frequency of each keyword represents the total number of texts containing the keyword in the text collection. The frequency determining unit 1120 may further calculate an importance parameter value of each keyword in the keyword set to the keyword set. In one embodiment, the frequency determination unit 1120 may construct a network graph comprising a plurality of nodes. Each node in the network graph represents a keyword in the set of keywords. Two nodes of any text in the text set in the network graph are connected. The relevance parameter value of each pair of two connected nodes in the network graph corresponds to the number of texts in the text set containing the two nodes. According to the text arrangement algorithm, the frequency determination unit 1120 calculates the weight value of each node in the network graph, and uses the weight value as the importance parameter value of the node.
On this basis, the frequency determining unit 1120 may perform weighted calculation on the text frequency and the importance parameter value of each keyword, and use the corresponding calculation result as the frequency parameter value of the keyword.
The screening unit 1130 is configured to sort the keywords in the keyword set according to the frequency parameter value corresponding to each keyword, extract the keywords related to the predetermined delivery effect from the corresponding sorting result, and use the extracted keywords as the recommended word set. In one embodiment, the filtering unit 1130 may extract a plurality of keywords (e.g., a first predetermined number of keywords) from the corresponding ranking results in order of the frequency parameter values from high to low, and take the extracted keywords as a first filtering set. In another embodiment, the filtering unit 1130 may extract a keyword having a frequency parameter value higher than a frequency threshold from the sorting result, and use the keyword as the first filtering set.
On this basis, the screening unit 1130 may extract keywords related to a predetermined delivery effect from the first screening set according to the delivery effect record corresponding to each text in the text set, and use the keywords as a recommended word set. In some embodiments, the corresponding impression effect record for each text includes an exposure record and a click record corresponding to the text. The filtering unit 1130 may calculate an exposure support for each keyword in the first filtering set. The exposure support degree represents the ratio of the exposure times corresponding to the texts containing the keywords in the text set to the total exposure times corresponding to all the texts in the text set. The filtering unit 1130 may extract a keyword having an exposure support degree higher than a first threshold value from the first filtering set, and regard the extracted keyword as a second filtering set. Based on the exposure record and the click record corresponding to each text in the text set, the filtering unit 1130 may determine the click rate corresponding to the text. The screening unit 1130 may calculate an exposure confidence for each keyword in the second screening set. The exposure confidence of each keyword represents the ratio of the exposure times of the text containing the keyword in the text set and having the click rate higher than the second threshold to the total exposure times of the text containing the keyword in the text set. And sequencing the keywords in the second screening set according to the exposure confidence coefficient of each keyword in the second screening set, extracting a plurality of keywords from corresponding sequencing results, and taking the extracted keywords as a recommended word set. For example, the screening unit 1130 may extract a plurality of keywords (e.g., a second predetermined number of keywords) from the ranking results of the second screening set in order from high to low exposure confidence, and take the extracted keywords as the recommended word set. In another embodiment, the filtering unit 1130 may extract a keyword having a frequency parameter value higher than a frequency threshold from the ranking result, and use the keyword as the recommended word set.
In some embodiments, the screening unit 1130 may further extract, according to the delivery effect record corresponding to each text in the text set, a keyword for reducing the delivery effect from the sorting result of the keyword set, and use the keyword as the non-recommended word set. For example, the filtering unit 1130 may extract a plurality of keywords that do not belong to the first filtering set from the sorting results of the keyword sets and treat them as the non-recommended word set. For another example, the filtering unit 1130 may extract a plurality of keywords that do not belong to the recommended word set from the ranking results corresponding to the second filtering set (or the third filtering set), and use the extracted keywords as the non-recommended word set.
In some embodiments, the corresponding impression effect record for each text includes a click record and a conversion record corresponding to the text. The filtering unit 1130 is configured to calculate a click support degree for each keyword in the first filtering set. The click support degree represents the ratio of the number of clicks corresponding to the text containing the keyword in the text set to the total number of clicks corresponding to all the texts in the text set. The screening unit 1130 extracts a keyword having a click support degree higher than a third threshold value in the first screening set, and takes the extracted keyword as a third screening set. According to the click record and the conversion record corresponding to each text in the text set, the filtering unit 1130 may determine the conversion rate corresponding to the text. The screening unit 1130 may calculate the click confidence for each keyword in the third screening set. And the click confidence of each keyword represents the ratio of the number of clicks of the text which contains the keyword in the text set and has the conversion rate higher than a fourth threshold to the total number of clicks of the text which contains the keyword in the text set. The screening unit 1130 may sort the keywords in the third screening set according to the click confidence of each keyword in the third screening set, extract a plurality of keywords from the corresponding sorting result according to the order of the click confidence from high to low, and use the extracted keywords as the recommended word set. More specific implementations of the apparatus 1100 are consistent with the method 200 and will not be described further herein.
FIG. 12 illustrates a schematic diagram of an apparatus 1200 for presenting recommended words according to some embodiments of the present application. The apparatus 1200 may reside in the computing device 130, for example.
As shown in fig. 12, the apparatus 1200 includes an acquisition unit 1210, a selection unit 1220, and a presentation unit 1230. The obtaining unit 1210 is configured to obtain a plurality of recommended word sets generated according to the keyword recommendation method (e.g., method 200 or 700) of the present application. The attribute value of each recommended word set comprises an industry category and an issuing channel. The selecting unit 1220 provides a user interface presenting attribute options corresponding to a plurality of recommended word sets in response to an operation of selecting the recommended word set. The attribute options include industry categories and delivery channels. The presenting unit 1230 determines and presents a recommended word set corresponding to the selected attribute value in response to a selection operation of an attribute option in the user interface, so as to prompt the user to use a keyword in the recommended word set when editing text.
In some embodiments, the presenting unit 1230 is further configured to determine and present a set of non-recommended words corresponding to the selected attribute value, so as to prompt the user to avoid using keywords in the set of non-recommended words when editing the text. More specific embodiments of the apparatus 1200 are consistent with the method 800 and will not be described further herein.
FIG. 13 illustrates a block diagram of the components of a computing device. As shown in fig. 13, the computing device includes one or more processors (CPU or GPU) 1302, communication module 1304, memory 1306, user interface 1310, and a communication bus 1308 for interconnecting these components.
The processor 1302 may receive and transmit data via the communication module 1304 to enable network communications and/or local communications.
The user interface 1310 includes one or more output devices 1312 including one or more speakers and/or one or more visual displays. The user interface 1310 also includes one or more input devices 1314, including, for example, a keyboard, mouse, voice command input unit or microphone, touch screen display, touch sensitive tablet, gesture capture camera or other input buttons or controls, and the like.
The memory 1306 may be a high-speed random access memory such as DRAM, SRAM, DDR RAM, or other random access solid state memory devices; or non-volatile memory, such as one or more magnetic disk storage devices, optical disk storage devices, flash memory devices, or other non-volatile solid-state storage devices.
Memory 1306 stores a set of instructions executable by processor 1302, including:
an operating system 1316, including programs for handling various basic system services and for performing hardware related tasks;
the application 1318 includes various programs for implementing the keyword recommendation method (e.g., the method 200 or 700) or the method for presenting a recommended word (e.g., the method 800), and such programs can implement the processing flows in the examples described above, such as may include an application for generating a recommended word or an application for presenting a recommended word according to the present application. The application generating the recommended word may include the keyword recommendation apparatus 1100 shown in fig. 11. The application presenting the recommended word may include, for example, the apparatus 1200 presenting the recommended word shown in fig. 12.
In addition, each of the examples of the present application may be realized by a data processing program executed by a data processing apparatus such as a computer. It is clear that a data processing program constitutes the present application. Further, the data processing program, which is generally stored in one storage medium, is executed by directly reading the program out of the storage medium or by installing or copying the program into a storage device (such as a hard disk and/or a memory) of the data processing device. Such a storage medium therefore also constitutes the present invention. The storage medium may use any type of recording means, such as a paper storage medium (e.g., paper tape, etc.), a magnetic storage medium (e.g., a flexible disk, a hard disk, a flash memory, etc.), an optical storage medium (e.g., a CD-ROM, etc.), a magneto-optical storage medium (e.g., an MO, etc.), and the like.
The present application therefore also discloses a non-volatile storage medium having stored therein a data processing program for executing any one of the examples of the method described herein.
In addition, the method steps described in this application may be implemented by hardware, for example, logic gates, switches, application Specific Integrated Circuits (ASICs), programmable logic controllers, embedded microcontrollers, and the like, in addition to data processing programs. Such hardware capable of implementing the methods described herein may also constitute the present application.
The above description is only a preferred example of the present application and should not be taken as limiting the present application, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims (14)
1. A keyword recommendation method is characterized by comprising the following steps:
acquiring a keyword set, wherein the keyword set comprises keywords corresponding to each text in a text set;
determining a frequency parameter value corresponding to each keyword in the keyword set, including: determining a text frequency of each keyword in the set of keywords, wherein the text frequency of each keyword represents a total number of texts containing the keyword in the set of texts; calculating an importance parameter value of each keyword in the keyword set to the keyword set; carrying out weighted calculation on the text frequency and the importance parameter value of each keyword, and taking the corresponding calculation result as the frequency parameter value of the keyword; the frequency parameter value corresponding to each keyword is used for describing the use range of the keyword; and
and sequencing all the keywords in the keyword set according to the frequency parameter value corresponding to each keyword, extracting the keywords related to the preset putting effect from the corresponding sequencing result, and taking the keywords as a recommended word set.
2. The method of claim 1, wherein the obtaining a set of keywords comprises:
acquiring a text set containing a plurality of texts, wherein each text is used for describing additional content which is presented in a page, and acquiring a release effect record of the additional content corresponding to each text;
and performing word segmentation operation on each text in the text set respectively, acquiring a keyword corresponding to each text, and generating a keyword set corresponding to the text set.
3. The method of claim 2, wherein said obtaining a text collection comprising a plurality of texts comprises:
acquiring a plurality of additional contents which belong to a preset industry category and are released to a preset release channel, and generating the text set, wherein the text set comprises texts corresponding to the additional contents in the plurality of additional contents.
4. The method of claim 1, wherein the step of calculating the importance parameter value for each keyword in the set of keywords comprises:
constructing a network graph comprising a plurality of nodes, wherein each node in the network graph represents a keyword in the keyword set, two nodes contained in any text in the text set in the network graph are connected, and the correlation parameter value of each pair of connected two nodes in the network graph corresponds to the number of texts containing the two nodes in the text set;
and calculating the weight value of each node in the network graph according to a text arrangement algorithm, and taking the weight value as the importance parameter value of the node.
5. The method as claimed in claim 2, wherein the corresponding delivery effect record of each text comprises an exposure record and a click record corresponding to the text, and the extracting keywords related to the predetermined delivery effect from the corresponding ranking result as the set of recommended words comprises:
extracting a plurality of keywords from the sequencing result according to the sequence of the frequency parameter values from high to low, and taking the keywords as a first screening set;
calculating the exposure support degree of each keyword in the first screening set, wherein the exposure support degree represents the ratio of the exposure times corresponding to the texts containing the keyword in the text set to the total exposure times corresponding to all the texts in the text set;
extracting keywords with exposure support higher than a first threshold value in the first screening set, and taking the extracted keywords as a second screening set;
determining the click rate corresponding to each text according to the exposure record and the click record corresponding to each text in the text set;
calculating the exposure confidence coefficient of each keyword in the second screening set, wherein the exposure confidence coefficient of each keyword represents the ratio of the exposure times of the text which contains the keyword in the text set and has the click rate higher than a second threshold value to the total exposure times of the text which contains the keyword in the text set; and
and sequencing the keywords in the second screening set according to the exposure confidence of each keyword in the second screening set, extracting a plurality of keywords from corresponding sequencing results according to the sequence of the exposure confidence from high to low, and taking the extracted keywords as the recommended word set.
6. The method of claim 1, further comprising: and extracting keywords for reducing the putting effect from the sequencing result, and using the keywords as a non-recommended word set.
7. The method of claim 5, further comprising: and extracting a plurality of keywords which do not belong to the recommended word set from the sequencing result corresponding to the second screening set, and taking the extracted keywords as a non-recommended word set.
8. The method of claim 2, wherein the corresponding delivery effect record of each text comprises a click record and a conversion record corresponding to the text, and the extracting keywords related to the predetermined delivery effect from the corresponding sorting result as a recommendation set comprises:
extracting a plurality of keywords from the sequencing result according to the sequence of the frequency parameter values from high to low, and taking the keywords as a first screening set;
calculating the click support degree of each keyword in the first screening set, wherein the click support degree represents the ratio of the click times corresponding to texts containing the keyword in the text set to the total click times corresponding to all texts in the text set;
extracting key words with click support higher than a third threshold value in the first screening set, and taking the extracted key words as a third screening set;
determining a conversion rate corresponding to each text according to the click record and the conversion record corresponding to each text in the text set;
calculating the click confidence of each keyword in the third screening set, wherein the click confidence of each keyword represents the ratio of the number of clicks of the text which contains the keyword and has a conversion rate higher than a fourth threshold to the total number of clicks of the text which contains the keyword in the text set; and
and sequencing the keywords in the third screening set according to the click confidence of each keyword in the third screening set, extracting a plurality of keywords from corresponding sequencing results according to the sequence of the click confidence from high to low, and taking the extracted keywords as the recommended word set.
9. A method of presenting recommended words, comprising:
acquiring a plurality of recommended word sets generated according to the method of claim 1, wherein the attribute value of each recommended word set comprises an industry category and a delivery channel;
responding to the operation of selecting the recommended word set, and providing a user interface for presenting attribute options corresponding to the recommended word sets, wherein the attribute options comprise industry categories and delivery channels; and
and responding to the selection operation of the attribute option in the user interface, determining and presenting a recommended word set corresponding to the selected attribute value so as to prompt the user to use the keywords in the recommended word set when editing the text.
10. The method of claim 9, further comprising: and determining and presenting a non-recommended word set corresponding to the selected attribute value so as to prompt a user to avoid using keywords in the non-recommended word set when editing the text.
11. A keyword recommendation device, comprising:
the device comprises an acquisition unit, a processing unit and a display unit, wherein the acquisition unit is used for acquiring a keyword set, and the keyword set comprises keywords corresponding to each text in a text set;
a frequency determining unit, configured to determine a frequency parameter value corresponding to each keyword in the keyword set, including: determining a text frequency of each keyword in the set of keywords, wherein the text frequency of each keyword represents a total number of texts containing the keyword in the set of texts; calculating an importance parameter value of each keyword in the keyword set to the keyword set; performing weighted calculation on the text frequency and the importance parameter value of each keyword, and taking the corresponding calculation result as the frequency parameter value of the keyword; the frequency parameter value corresponding to each keyword is used for describing the use popularity of the keyword;
and the screening unit is used for sorting the keywords in the keyword set according to the frequency parameter value corresponding to each keyword, extracting the keywords related to the preset putting effect from the corresponding sorting result and taking the keywords as a recommended word set.
12. An apparatus for presenting a recommended word, comprising:
an obtaining unit, configured to obtain a plurality of recommended word sets generated according to the method of claim 1, where an attribute value of each recommended word set includes an industry category and an issuing channel;
the selection unit responds to the operation of selecting the recommended word set and provides a user interface for presenting attribute options corresponding to the recommended word sets, wherein the attribute options comprise industry categories and delivery channels; and
and the presentation unit is used for responding to the selection operation of the attribute option in the user interface, determining and presenting the recommended word set corresponding to the selected attribute value so as to prompt the user to use the key words in the recommended word set when editing the text.
13. A computing device, comprising:
one or more processors;
a memory; and
one or more programs stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing the method of any of claims 1-10.
14. A storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform the method of any of claims 1-10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711068460.8A CN110020132B (en) | 2017-11-03 | 2017-11-03 | Keyword recommendation method and device, computing equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711068460.8A CN110020132B (en) | 2017-11-03 | 2017-11-03 | Keyword recommendation method and device, computing equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110020132A CN110020132A (en) | 2019-07-16 |
| CN110020132B true CN110020132B (en) | 2023-04-11 |
Family
ID=67186749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711068460.8A Active CN110020132B (en) | 2017-11-03 | 2017-11-03 | Keyword recommendation method and device, computing equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110020132B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110472039A (en) * | 2019-08-21 | 2019-11-19 | 上海络昕信息科技有限公司 | A kind of generation method of synopsis, system and relevant apparatus |
| CN111311351B (en) * | 2020-01-15 | 2023-09-01 | 上海风秩科技有限公司 | Document recommendation method, device, computer equipment and readable storage medium |
| CN112417130B (en) * | 2020-11-19 | 2023-06-16 | 贝壳技术有限公司 | Word screening method, device, computer readable storage medium and electronic equipment |
| CN114168843B (en) * | 2021-11-05 | 2024-10-18 | 杭州阿里巴巴海外互联网产业有限公司 | Search term recommendation method, device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103885971A (en) * | 2012-12-20 | 2014-06-25 | 阿里巴巴集团控股有限公司 | Data pushing method and data pushing device |
| CN103970800A (en) * | 2013-02-05 | 2014-08-06 | 腾讯科技（深圳）有限公司 | Method and system for extracting and processing webpage related keywords |
| CN104063387A (en) * | 2013-03-19 | 2014-09-24 | 三星电子（中国）研发中心 | Device and method abstracting keywords in text |
| CN104636334A (en) * | 2013-11-06 | 2015-05-20 | 阿里巴巴集团控股有限公司 | Keyword recommending method and device |
| CN104699696A (en) * | 2013-12-05 | 2015-06-10 | 深圳市腾讯计算机系统有限公司 | File recommendation method and device |
| CN105701108A (en) * | 2014-11-26 | 2016-06-22 | 阿里巴巴集团控股有限公司 | Information recommendation method, information recommendation device and server |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110145076A1 (en) * | 2005-09-14 | 2011-06-16 | Jorey Ramer | Mobile Campaign Creation |
| US20090240498A1 (en) * | 2008-03-19 | 2009-09-24 | Microsoft Corporation | Similiarity measures for short segments of text |
| JP2013058110A (en) * | 2011-09-08 | 2013-03-28 | Toshiba Corp | Information processing apparatus, information processing method and program |
| WO2016070028A1 (en) * | 2014-10-31 | 2016-05-06 | Informite Inc. | Systems and methods for keyword research and analysis for paid search |
-
2017
- 2017-11-03 CN CN201711068460.8A patent/CN110020132B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103885971A (en) * | 2012-12-20 | 2014-06-25 | 阿里巴巴集团控股有限公司 | Data pushing method and data pushing device |
| CN103970800A (en) * | 2013-02-05 | 2014-08-06 | 腾讯科技（深圳）有限公司 | Method and system for extracting and processing webpage related keywords |
| CN104063387A (en) * | 2013-03-19 | 2014-09-24 | 三星电子（中国）研发中心 | Device and method abstracting keywords in text |
| CN104636334A (en) * | 2013-11-06 | 2015-05-20 | 阿里巴巴集团控股有限公司 | Keyword recommending method and device |
| CN104699696A (en) * | 2013-12-05 | 2015-06-10 | 深圳市腾讯计算机系统有限公司 | File recommendation method and device |
| CN105701108A (en) * | 2014-11-26 | 2016-06-22 | 阿里巴巴集团控股有限公司 | Information recommendation method, information recommendation device and server |
Non-Patent Citations (2)
| Title |
|---|
| Wei-Hao Hwang 等."Personalized Internet Advertisement Recommendation Service Based on Keyword Similarity".《2015 IEEE 39th Annual International Computers, Software & Applications Conference》.2015,第29-33页. * |
| 朱泽德 等."一种基于LDA 模型的关键词抽取方法".《中南大学学报(自然科学版)》.2015,第2142-2148页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110020132A (en) | 2019-07-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11797622B2 (en) | Text-to-media indexes on online social networks | |
| US11829404B2 (en) | Functional image archiving | |
| JP6290952B2 (en) | Grammar model for structured search queries | |
| CN105247507B (en) | Method, system and storage medium for the influence power score for determining brand | |
| CN110020132B (en) | Keyword recommendation method and device, computing equipment and storage medium | |
| CN107784092A (en) | A kind of method, server and computer-readable medium for recommending hot word | |
| CN107851092A (en) | Personal entity models | |
| CN109791680A (en) | Key frame of video on online social networks is shown | |
| JP2019519019A (en) | Method, apparatus and device for identifying text type | |
| CN102779114A (en) | Unstructured data support generated by utilizing automatic rules | |
| US10021061B1 (en) | Message presentation management in a social networking environment | |
| WO2022262487A1 (en) | Form generation method, apparatus and device, and medium | |
| US20150350139A1 (en) | System and method for following topics in an electronic textual conversation | |
| US20120046937A1 (en) | Semantic classification of variable data campaign information | |
| CN113688310A (en) | Content recommendation method, device, equipment and storage medium | |
| CN106663123B (en) | Comment-centric news reader | |
| US10642914B2 (en) | Automatically generating multi-image posts on online social networks | |
| WO2019133164A1 (en) | System and method for recommending features for content presentations | |
| CN106993226A (en) | A kind of method and terminal of recommendation video | |
| US20170316807A1 (en) | Systems and methods for creating whiteboard animation videos | |
| EP3306555A1 (en) | Diversifying media search results on online social networks | |
| JP2019175212A (en) | Information display device, article page generation device, information processing device, information display system, and program | |
| CN110020120A (en) | Feature word treatment method, device and storage medium in content delivery system | |
| CN110147488B (en) | Page content processing method, processing device, computing equipment and storage medium | |
| Preotiuc-Pietro | Temporal models of streaming social media data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |