CN108776901B - Advertisement recommendation method and system based on search terms - Google Patents
Advertisement recommendation method and system based on search terms Download PDFInfo
- Publication number
- CN108776901B CN108776901B CN201810392779.4A CN201810392779A CN108776901B CN 108776901 B CN108776901 B CN 108776901B CN 201810392779 A CN201810392779 A CN 201810392779A CN 108776901 B CN108776901 B CN 108776901B
- Authority
- CN
- China
- Prior art keywords
- recommended
- advertisement
- current search
- search word
- word
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the field of advertisement putting, in particular to an advertisement recommendation method and system based on search terms. The method comprises the following steps: carrying out expansion processing on a current search word input by a user to obtain an expansion phrase of the current search word; performing expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended; calculating to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the expanded word group of the current search word and the expanded word group of each advertisement to be recommended; and selecting at least one recommended advertisement from the advertisements to be recommended to be displayed on the corresponding page of the current search word according to the comprehensive score of each advertisement to be recommended for the current search word. The invention can recommend the advertisement with strong correlation degree with the search word, and realize accurate delivery.
Description
Technical Field
The invention relates to the field of advertisement putting, in particular to an advertisement recommendation method and system based on search terms.
Background
In the prior art, advertisements are screened by matching search terms themselves; that is, the relevance is determined according to the matching degree of the search terms and the advertisement text, and then the advertisements are sequenced according to the relevance, and finally the recommended advertisements are obtained.
The prior art has the following problems:
1. search terms are typically short, extensive, and informative, resulting in fewer recalled ads.
2. The degree of word matching does not reflect the degree of relevance of the advertisement to the search; for example, the search term is "internet congress" and contains "interconnection"; an advertisement is a product of a router, and the product description also has "interconnection". The correlation between the two is really low. But according to word matching, inputting the internet meeting will correspondingly recommend the advertisement of the router. This may result in the recommended advertisement being not strongly correlated with the search term itself, which may affect the accurate placement of the advertisement.
3. If the words do not match, traffic may be wasted.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the defects of the prior art and provide a method and a system for recommending advertisements based on search terms, which can recommend advertisements with strong correlation degree with the search terms and achieve accurate delivery.
In order to achieve the above technical object, in one aspect, the advertisement recommendation method based on search terms provided by the present invention includes:
carrying out expansion processing on a current search word input by a user to obtain an expansion phrase of the current search word;
performing expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
calculating to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the expanded word group of the current search word and the expanded word group of each advertisement to be recommended;
and selecting at least one recommended advertisement from the advertisements to be recommended to be displayed on the corresponding page of the current search word according to the comprehensive score of each advertisement to be recommended for the current search word.
In another aspect, the present invention provides a search term-based advertisement recommendation system, including:
the expansion unit is used for carrying out expansion processing on the current search word input by the user to obtain an expansion phrase of the current search word;
the expansion unit is used for performing expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
the computing unit is used for computing to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the extended phrase of the current search word and the extended phrase of each advertisement to be recommended;
and the recommending unit is used for selecting at least one recommended advertisement from the advertisements to be recommended according to the comprehensive score of each advertisement to be recommended for the current search term and displaying the recommended advertisement on a corresponding page of the current search term.
In the invention, the search word and the advertisement to be recommended are respectively expanded and expanded to obtain an expanded phrase and an expanded phrase; then, calculating the comprehensive scores of the search words and the advertisements to be recommended through the expanded phrases and the expanded phrases; thereby selecting a recommended advertisement by synthesizing the scores. Therefore, after the search terms are expanded, more advertisements can be matched, so that more advertisements can be recommended; the recommendation advertisements and the search terms have stronger relevance, so that the advertisements can be accurately put.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a system configuration according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a specific work flow in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1.1: as shown in fig. 1, the advertisement recommendation method based on search terms according to the present invention includes:
101. carrying out expansion processing on a current search word input by a user to obtain an expansion phrase of the current search word;
102. performing expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
103. calculating to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the expanded word group of the current search word and the expanded word group of each advertisement to be recommended;
104. and selecting at least one recommended advertisement from the advertisements to be recommended to be displayed on the corresponding page of the current search word according to the comprehensive score of each advertisement to be recommended for the current search word.
Example 1.2: as embodiment 1.1 follows, the expanding process is performed on the current search word input by the user to obtain an expanded phrase of the current search word, and the method further includes:
judging that the current search word input by the user is a hot search word; the hot search word is a search word with the word frequency exceeding the preset word frequency.
That is to say, when the current search word input by the user is judged to be the hot search word, the step of performing expansion processing on the current search word input by the user to obtain an expanded phrase of the current search word is executed.
Example 1.3: next to embodiment 1.1 or 1.2, the expanding process is performed on the current search term input by the user to obtain an expanded phrase of the current search term, and then the expanding process further includes:
sequentially performing word segmentation, single word removal and stop word removal on the extended word group of the current search word;
the method comprises the following steps of carrying out expansion processing on text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended, and then:
and sequentially performing word segmentation, word removal and stop word removal on the extended word group corresponding to each text keyword of the advertisement to be recommended.
In the above technical solution, the expanding process is performed on the current search term input by the user to obtain an expanded phrase of the current search term, and specifically includes:
obtaining an expansion word of a current search word, wherein the expansion word comprises but is not limited to: related words, impression words, synonyms, co-occurrence words, associated words and log expanded words;
and combining the current search word and the corresponding expansion word into an expansion phrase of the current search word.
The related words are words input by the same user within a preset time period after the current search word is input;
the impression words are words having public sentiment relation with the current search words;
the co-occurrence word is a word input in the search box simultaneously with the current search word;
the associated words are words which have an association relation with the current search word in the word sequence;
the log expansion words are words with higher word frequency in the context, which are used for checking the context of the current search word through the search log.
In the above technical solution, the calculating, according to the extended phrase of the current search word and the extended phrase of each advertisement to be recommended, a comprehensive score of each advertisement to be recommended for the current search word includes:
respectively coding the extended phrases of the current search word and the extended phrases of each advertisement to be recommended to obtain the vector of the current search word and the vector of each advertisement to be recommended;
respectively calculating the cosine similarity of the vector of each advertisement to be recommended to the vector of the current search word, so as to obtain a word similarity score of each advertisement to be recommended to the current search word;
acquiring each attribute label of an extended phrase of a current search word;
judging each attribute label of the extended phrase of each advertisement to be recommended;
matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended respectively, and calculating to obtain an attribute similarity score of each advertisement to be recommended for the current search word;
and adding or weighted adding the word similarity score and the attribute similarity score of each advertisement to be recommended aiming at the current search word to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word.
Further, the attribute tag includes: a subject label, a type label, and an industry label;
the obtaining of each attribute label of the extended phrase of the current search word specifically includes:
respectively establishing a main body model, a type model and an industry model according to historical data;
respectively inputting the extended phrases of the current search word into a theme model, a type model and an industry model, and outputting to obtain a corresponding theme label, a type label and an industry label;
the method for calculating the attribute similarity score of each advertisement to be recommended aiming at the current search word comprises the following steps of matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended respectively, and calculating the attribute similarity score of each advertisement to be recommended aiming at the current search word, wherein the method specifically comprises the following steps:
comparing the subject label of the extended phrase of the current search word with the subject label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the first preset score is added into the theme score of the current advertisement to be recommended aiming at the current search word;
comparing the type label of the extended phrase of the current search word with the type label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the second preset score is added to the type score of the current advertisement to be recommended aiming at the current search word;
comparing the industry label of the extended phrase of the current search word with the industry label of the extended phrase of each advertisement to be recommended;
if the industry label of the expanded phrase of the current search word is the same as the industry label of the expanded phrase of the current advertisement to be recommended, the third preset score is added to the industry score of the current advertisement to be recommended for the current search word;
and adding or weighting and adding the theme score, the type score and the industry score of the current advertisement to be recommended aiming at the current search word to obtain the attribute similarity score of the current advertisement to be recommended aiming at the current search word.
In the above technical solution, according to the comprehensive score of each advertisement to be recommended for the current search term, at least one recommended advertisement is selected from the advertisements to be recommended and displayed on the corresponding page of the current search term, which specifically includes:
sequencing each advertisement to be recommended from high to low according to the comprehensive score of the advertisement to be recommended for the current search term;
taking the first n (n is more than 0) advertisements to be recommended as recommended advertisements;
and displaying the recommended advertisements on the corresponding pages of the current search terms according to the sequence.
Example 2.1: as shown in fig. 2, the advertisement recommendation system based on search terms according to the present invention includes:
the expansion unit 11 is configured to perform expansion processing on a current search term input by a user to obtain an expanded phrase of the current search term;
the expansion unit 12 is configured to perform expansion processing on the text keyword of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
the calculation unit 13 is configured to calculate, according to the extended phrase of the current search word and the extended phrase of each advertisement to be recommended, a comprehensive score of each advertisement to be recommended for the current search word;
and the recommending unit 14 is configured to select at least one recommended advertisement from the advertisements to be recommended according to the comprehensive score of each advertisement to be recommended for the current search term, and display the at least one recommended advertisement on a corresponding page of the current search term.
Example 2.2: as a possible configuration, following embodiment 2.1, the system further comprises:
the judging unit 15 is used for triggering the expanding unit 11 when judging that the current search word input by the user is a hot search word; the hot search word is a search word with the word frequency exceeding the preset word frequency.
Example 2.3: as another possible configuration, following embodiment 2.1 or 2.2, the system further comprises:
the first filtering unit 16 is used for sequentially performing word segmentation, word removal and stop word removal on the extended word group of the current search word;
and the second filtering unit 17 is configured to perform word segmentation, word removal and stop word removal on the extended word group corresponding to each text keyword of the advertisement to be recommended in sequence.
In the foregoing technical solution, the extension unit 11 is specifically configured to:
obtaining an expansion word of a current search word, wherein the expansion word comprises but is not limited to: related words, impression words, synonyms, co-occurrence words, associated words and log expanded words;
and combining the current search word and the corresponding expansion word into an expansion phrase of the current search word.
In the above technical solution, the calculating unit 13 includes:
the encoding module is used for respectively encoding the extended phrases of the current search word and the extended phrases of each advertisement to be recommended to obtain the vector of the current search word and the vector of each advertisement to be recommended;
the calculation module is used for calculating the cosine similarity of the vector of each advertisement to be recommended to the vector of the current search word respectively so as to obtain a word similarity score of each advertisement to be recommended to the current search word;
the acquisition module is used for acquiring each attribute label of the extended phrase of the current search word;
the judging module is used for judging each attribute label of the extended phrase of each advertisement to be recommended;
the matching module is used for respectively matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended and calculating to obtain an attribute similarity score of each advertisement to be recommended for the current search word;
and the adding module is used for adding or weighting and adding the word similarity score and the attribute similarity score of each advertisement to be recommended aiming at the current search word to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word.
Further, the attribute tag includes: a subject label, a type label, and an industry label;
the acquisition module is specifically configured to:
respectively establishing a main body model, a type model and an industry model according to historical data;
respectively inputting the extended phrases of the current search word into a theme model, a type model and an industry model, and outputting to obtain a corresponding theme label, a type label and an industry label;
the matching module is specifically configured to:
comparing the subject label of the extended phrase of the current search word with the subject label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the first preset score is added into the theme score of the current advertisement to be recommended aiming at the current search word;
comparing the type label of the extended phrase of the current search word with the type label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the second preset score is added to the type score of the current advertisement to be recommended aiming at the current search word;
comparing the industry label of the extended phrase of the current search word with the industry label of the extended phrase of each advertisement to be recommended;
if the industry label of the expanded phrase of the current search word is the same as the industry label of the expanded phrase of the current advertisement to be recommended, the third preset score is added to the industry score of the current advertisement to be recommended for the current search word;
and adding or weighting and adding the theme score, the type score and the industry score of the current advertisement to be recommended aiming at the current search word to obtain the attribute similarity score of the current advertisement to be recommended aiming at the current search word.
In the above technical solution, the recommending unit 14 is specifically configured to:
sequencing each advertisement to be recommended from high to low according to the comprehensive score of the advertisement to be recommended for the current search term;
taking the first n (n is more than 0) advertisements to be recommended as recommended advertisements;
and displaying the recommended advertisements on the corresponding pages of the current search terms according to the sequence.
The above technical solutions of the embodiments of the present invention are described in detail below with reference to application examples:
take microblog as an example;
as shown in fig. 3, the advertisement recommendation method based on search terms includes:
step 1, judging that a current search word input by a user is a hot search word; the hot search word is a search word of which the word frequency exceeds a preset word frequency;
the amount of searches for users to enter the microblog search box every day is huge, and if each search word is subjected to expansion matching advertisements, the burden of a server is very heavy. Statistics show that search words related to hot time occurring in a short period of time in a microblog search box account for most of the search words, so that the occurrence frequency of some search words is very high, and the words are called hot search words. Therefore, in the embodiment, only the hot search words are subjected to subsequent expansion and advertisement matching, so that the pressure of the microblog server can be greatly reduced, and the advertisement putting efficiency can be improved.
in actual working implementation, a composite score needs to be calculated for each search term input by all users and all advertisements to be recommended. The calculation workload is huge, and since different users input the same search term for a certain hot event in a certain time period, in fact, the search term is also called a hot search term, so that a large amount of repeated calculation is performed in the calculation workload.
Therefore, the comprehensive scores between the same search word and the same advertisement to be recommended are stored in the comprehensive score table, and whether the comprehensive scores of the current search word and the current advertisement to be recommended exist in the comprehensive score table is inquired before the comprehensive scores of the current search word and the current advertisement to be recommended are calculated.
If the current search word and the current advertisement to be recommended store the comprehensive score, directly returning the comprehensive score;
and if the current search word and the current advertisement to be recommended do not store the comprehensive score, namely the comprehensive score is not calculated, entering the step 3.
It should be noted that, in the implementation of the operation, the technical effect of the present invention can be achieved by omitting step 2.
Step 3, judging whether the current search word input by the user stores an expansion phrase or not;
in the same way, step 2, in the actual working implementation, the expansion processing is required to be performed for each search term input by each user. This causes a large amount of work, and the search term (hot search term) after the expansion processing is stored in the expansion phrase storage table.
If the extended phrase of the current search word is stored in the extended phrase storage table, directly returning to the extended phrase;
and if the extended phrase storage table does not have the extended phrase of the current search word, entering the step 4.
It should be noted that, in the implementation of the operation, the technical effect of the present invention can be achieved by omitting step 3.
Step 4, carrying out expansion processing on the current search word input by the user to obtain an expansion phrase of the current search word; specifically, the method comprises the following steps:
obtaining an expansion word of a current search word, wherein the expansion word comprises but is not limited to: related words, impression words, synonyms, co-occurrence words, associated words and log expanded words;
and combining the current search word and the corresponding expansion word into an expansion phrase of the current search word.
The related words are words input by the same user within a preset time period after the current search word is input;
the impression words are words having public sentiment relation with the current search words;
the co-occurrence word is a word input in the search box simultaneously with the current search word;
the associated words are words which have an association relation with the current search word in the word sequence;
the log expansion words are words with higher word frequency in the context, which are used for checking the context of the current search word through the search log.
In an actual working implementation, the expansion word of the current search word may also not be limited to the above-mentioned word.
Step 5, judging whether the current advertisement to be recommended stores an expansion phrase or not;
similarly, in step 3, in the actual work implementation, the extension processing is required for each advertisement to be recommended. Thus, a large amount of workload is generated, and the advertisements to be recommended after the expansion processing are all stored in the expansion phrase storage table.
If the extended phrase of the current advertisement to be recommended is stored in the extended phrase storage table, directly returning to the extended phrase;
and if the extended phrase of the current advertisement to be recommended does not exist in the extended phrase storage table, entering the step 6.
It should be noted that, in the implementation of the operation, the technical effect of the present invention can be achieved by omitting step 5.
Step 6, carrying out expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
before the advertisement to be recommended is expanded, a keyword in the current advertisement to be recommended needs to be found, and then the keyword is expanded to obtain an expanded phrase of the current advertisement to be recommended.
Step 7, sequentially performing word segmentation, single word removal and stop word removal on the extended word group of the current search word;
step 8, performing word segmentation, word removal and stop word removal processing on the extended word group corresponding to each text keyword of the advertisement to be recommended in sequence;
when the expanded phrases and the expanded phrases are segmented, the specialized segmentation can be adopted. The word removing means removing words formed by a single word appearing in the extended phrase and the extended phrase. Stop words, which are generally determined by developers based on actual operating conditions, typically include: the words of (1), (d), (a), (b), (c), etc. do not have a specific meaning.
Step 9, calculating to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the expansion word group of the current search word and the expansion word group of each advertisement to be recommended; specifically, the method comprises the following steps:
step 9.1, respectively coding the extended phrases of the current search word and the extended phrases of each advertisement to be recommended to obtain the vector of the current search word and the vector of each advertisement to be recommended;
for example, the expanded phrase of the current search term is: "some star, son, birth, daddy";
the current word group of the advertisement to be recommended is as follows: "Germany, milk powder, daddy";
coding the current extended phrase and the extended phrase;
firstly, combining the current extended phrase and the extended phrase to obtain a combined phrase: "some star, son, birth, daddy, germany, milk powder";
then, comparing the current extended phrase with the combined phrase, wherein the same word code is 1, and different word codes are 0; obtaining a vector A of the current search word as [ 1, 1, 1, 1, 0, 0 ];
comparing the current extended phrase with the combined phrase, wherein the same word code is 1, and the different word code is 0; obtaining a vector B of the current search word as [ 0, 0, 0, 1, 1, 1 ];
9.2, respectively calculating the cosine similarity of the vector of each advertisement to be recommended to the vector of the current search word, so as to obtain a word similarity score of each advertisement to be recommended to the current search word;
calculating cosine similarity of A and B by using the Euclidean distance; the following formula:
Similarity＝cos(theta)＝A·B/||A||||B||(1)
9.3, acquiring each attribute label of the extended phrase of the current search word;
respectively establishing a main body model, a type model and an industry model according to historical data;
and respectively inputting the extended phrases of the current search word into the theme model, the type model and the industry model, and outputting to obtain a corresponding theme label, a corresponding type label and a corresponding industry label.
On the topic level, combining historical hot search data and text inner cylinders of advertisements, and generating a training topic model such as a model LDA by using a document topic. And taking the extended phrase of the current search word as the input of the topic model, and outputting to obtain at least one topic.
Predefining various types, building classification models from historical data, typically using support vector machines svm or Bayes
Bayes for classification.
9.4, judging each attribute label of the extended phrase of each advertisement to be recommended;
9.5, matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended respectively, and calculating to obtain an attribute similarity score of each advertisement to be recommended for the current search word; specifically, the method comprises the following steps:
comparing the subject label of the extended phrase of the current search word with the subject label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the first preset score is added into the theme score of the current advertisement to be recommended aiming at the current search word;
comparing the type label of the extended phrase of the current search word with the type label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the second preset score is added to the type score of the current advertisement to be recommended aiming at the current search word;
comparing the industry label of the extended phrase of the current search word with the industry label of the extended phrase of each advertisement to be recommended;
if the industry label of the expanded phrase of the current search word is the same as the industry label of the expanded phrase of the current advertisement to be recommended, the third preset score is added to the industry score of the current advertisement to be recommended for the current search word;
and adding or weighting and adding the theme score, the type score and the industry score of the current advertisement to be recommended aiming at the current search word to obtain the attribute similarity score of the current advertisement to be recommended aiming at the current search word.
And comparing the obtained at least one theme with the theme of the current advertisement to be recommended after the theme model is output. If one topic is matched, the first preset score is set to be 10 points, and the topic score of the current advertisement to be recommended aiming at the current search word is 10 points. If the two topics are matched, the topic of the current advertisement to be recommended aiming at the current search word is divided into 20 points; and so on.
The same reasoning holds for the calculation of the type score and the calculation of the topic score. The types are predefined, the hot search expansion words are directly classified by the classification model and then are compared with the classification of the advertisement postscript, and if the hot search expansion words are matched with the advertisement postscript, a second preset score is added. If a hierarchical classification system is used, the second predetermined score may be obtained by taking a weighted average of the correlation scores between the different layers.
And 9.6, adding or weighting and adding the word similarity scores and the attribute similarity scores to obtain a comprehensive score.
Step 10, selecting at least one recommended advertisement from the advertisements to be recommended to be displayed on a corresponding page of the current search term according to the comprehensive score of each advertisement to be recommended for the current search term; specifically, the method comprises the following steps:
step 10.1, sorting each advertisement to be recommended according to the comprehensive score of the advertisement to be recommended for the current search term from high to low;
step 10.2, taking the first n (n is more than 0) advertisements to be recommended as recommended advertisements;
the recommended advertisement can also be selected from the advertisements to be recommended, the comprehensive scores of which exceed a certain preset score.
And step 10.3, displaying the recommended advertisements on the corresponding pages of the current search terms according to the sequence.
In the prior art, when the search words are matched with the advertisements to be recommended, only the keywords are limited, and a plurality of keywords are not in sequence, so that the n-gram of the Chinese language model or the semantic expression of the sentence level is lacked. In the present invention. After information such as subject, category, industry and the like is added, more semantic matches can be generated, so that more matched advertisements can be found through the current search words, the advertisements are accurately delivered, and the advertisement recalls are increased.
It should be understood that the specific order or hierarchy of steps in the processes disclosed is an example of exemplary approaches. Based upon design preferences, it is understood that the specific order or hierarchy of steps in the processes may be rearranged without departing from the scope of the present disclosure. The accompanying method claims present elements of the various steps in a sample order, and are not intended to be limited to the specific order or hierarchy presented.
In the foregoing detailed description, various features are grouped together in a single embodiment for the purpose of streamlining the disclosure. This method of disclosure is not to be interpreted as reflecting an intention that the claimed embodiments of the subject matter require more features than are expressly recited in each claim. Rather, as the following claims reflect, invention lies in less than all features of a single disclosed embodiment. Thus, the following claims are hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate preferred embodiment of the invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. To those skilled in the art; various modifications to these embodiments will be readily apparent, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
What has been described above includes examples of one or more embodiments. It is, of course, not possible to describe every conceivable combination of components or methodologies for purposes of describing the aforementioned embodiments, but one of ordinary skill in the art may recognize that many further combinations and permutations of various embodiments are possible. Accordingly, the embodiments described herein are intended to embrace all such alterations, modifications and variations that fall within the scope of the appended claims. Furthermore, to the extent that the term "includes" is used in either the detailed description or the claims, such term is intended to be inclusive in a manner similar to the term "comprising" as "comprising" is interpreted when employed as a transitional word in a claim. Furthermore, any use of the term "or" in the specification of the claims is intended to mean a "non-exclusive or".
Those of skill in the art will further appreciate that the various illustrative logical blocks, units, and steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate the interchangeability of hardware and software, various illustrative components, elements, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design requirements of the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present embodiments.
The various illustrative logical blocks, or elements, described in connection with the embodiments disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor, an Application Specific Integrated Circuit (ASIC), a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other similar configuration.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may be stored in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. For example, a storage medium may be coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC, which may be located in a user terminal. In the alternative, the processor and the storage medium may reside in different components in a user terminal.
In one or more exemplary designs, the functions described above in connection with the embodiments of the invention may be implemented in hardware, software, firmware, or any combination of the three. If implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media that facilitate transfer of a computer program from one place to another. Storage media may be any available media that can be accessed by a general purpose or special purpose computer. For example, such computer-readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to carry or store program code in the form of instructions or data structures and which can be read by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor. Additionally, any connection is properly termed a computer-readable medium, and, thus, is included if the software is transmitted from a website, server, or other remote source via a coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL), or wirelessly, e.g., infrared, radio, and microwave. Such discs (disk) and disks (disc) include compact disks, laser disks, optical disks, DVDs, floppy disks and blu-ray disks where disks usually reproduce data magnetically, while disks usually reproduce data optically with lasers. Combinations of the above may also be included in the computer-readable medium.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A method for recommending advertisements based on search terms, the method comprising:
carrying out expansion processing on a current search word input by a user to obtain an expansion phrase of the current search word;
performing expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
calculating to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the expanded word group of the current search word and the expanded word group of each advertisement to be recommended;
selecting at least one recommended advertisement from the advertisements to be recommended to be displayed on a corresponding page of the current search term according to the comprehensive score of each advertisement to be recommended for the current search term;
the method includes the steps of calculating a comprehensive score of each advertisement to be recommended for a current search word according to an extended phrase of the current search word and an extended phrase of each advertisement to be recommended, and specifically includes:
respectively coding the extended phrases of the current search word and the extended phrases of each advertisement to be recommended to obtain the vector of the current search word and the vector of each advertisement to be recommended;
respectively calculating the cosine similarity of the vector of each advertisement to be recommended to the vector of the current search word, so as to obtain a word similarity score of each advertisement to be recommended to the current search word;
acquiring each attribute label of an extended phrase of a current search word;
judging each attribute label of the extended phrase of each advertisement to be recommended;
matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended respectively, and calculating to obtain an attribute similarity score of each advertisement to be recommended for the current search word;
adding or weighted adding is carried out on the word similarity score and the attribute similarity score of each advertisement to be recommended aiming at the current search word, so as to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word;
wherein the attribute tag comprises: a subject label, a type label, and an industry label;
the obtaining of each attribute label of the extended phrase of the current search word specifically includes:
respectively establishing a main body model, a type model and an industry model according to historical data;
respectively inputting the extended phrases of the current search word into a theme model, a type model and an industry model, and outputting to obtain a corresponding theme label, a type label and an industry label;
the method for calculating the attribute similarity score of each advertisement to be recommended aiming at the current search word comprises the following steps of matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended respectively, and calculating the attribute similarity score of each advertisement to be recommended aiming at the current search word, wherein the method specifically comprises the following steps:
comparing the subject label of the extended phrase of the current search word with the subject label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the first preset score is added into the theme score of the current advertisement to be recommended aiming at the current search word;
comparing the type label of the extended phrase of the current search word with the type label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the second preset score is added to the type score of the current advertisement to be recommended aiming at the current search word;
comparing the industry label of the extended phrase of the current search word with the industry label of the extended phrase of each advertisement to be recommended;
if the industry label of the expanded phrase of the current search word is the same as the industry label of the expanded phrase of the current advertisement to be recommended, the third preset score is added to the industry score of the current advertisement to be recommended for the current search word;
and adding or weighting and adding the theme score, the type score and the industry score of the current advertisement to be recommended aiming at the current search word to obtain the attribute similarity score of the current advertisement to be recommended aiming at the current search word.
2. The method according to claim 1, wherein the expanding process is performed on the current search term input by the user to obtain an expanded phrase of the current search term, and the method further comprises:
judging that the current search word input by the user is a hot search word; the hot search word is a search word with the word frequency exceeding the preset word frequency.
3. The search word-based advertisement recommendation method according to claim 1 or 2, wherein the expanding process is performed on the current search word input by the user to obtain an expanded phrase of the current search word, and specifically includes:
obtaining an expansion word of a current search word, wherein the expansion word comprises but is not limited to: related words, impression words, synonyms, co-occurrence words, associated words and log expanded words;
and combining the current search word and the corresponding expansion word into an expansion phrase of the current search word.
4. The method according to claim 1 or 2, wherein the current search term inputted by the user is expanded to obtain an expanded phrase of the current search term, and then the method further comprises:
sequentially performing word segmentation, single word removal and stop word removal on the extended word group of the current search word;
the method comprises the following steps of carrying out expansion processing on text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended, and then:
and sequentially performing word segmentation, word removal and stop word removal on the extended word group corresponding to each text keyword of the advertisement to be recommended.
5. The method for recommending advertisements based on search terms according to claim 1 or 2, wherein at least one recommended advertisement is selected from the advertisements to be recommended and displayed on a corresponding page of the current search term according to the comprehensive score of each advertisement to be recommended for the current search term, and specifically comprises:
sequencing each advertisement to be recommended from high to low according to the comprehensive score of the advertisement to be recommended for the current search term;
taking the first n (n is more than 0) advertisements to be recommended as recommended advertisements;
and displaying the recommended advertisements on the corresponding pages of the current search terms according to the sequence.
6. A search term based advertisement recommendation system, the system comprising:
the expansion unit is used for carrying out expansion processing on the current search word input by the user to obtain an expansion phrase of the current search word;
the expansion unit is used for performing expansion processing on the text keywords of each advertisement to be recommended to obtain an expansion phrase of each advertisement to be recommended;
the computing unit is used for computing to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word according to the extended phrase of the current search word and the extended phrase of each advertisement to be recommended;
the recommending unit is used for selecting at least one recommended advertisement from the advertisements to be recommended according to the comprehensive score of each advertisement to be recommended for the current search term and displaying the recommended advertisement on a corresponding page of the current search term;
wherein the computing unit comprises:
the encoding module is used for respectively encoding the extended phrases of the current search word and the extended phrases of each advertisement to be recommended to obtain the vector of the current search word and the vector of each advertisement to be recommended;
the calculation module is used for calculating the cosine similarity of the vector of each advertisement to be recommended to the vector of the current search word respectively so as to obtain a word similarity score of each advertisement to be recommended to the current search word;
the acquisition module is used for acquiring each attribute label of the extended phrase of the current search word;
the judging module is used for judging each attribute label of the extended phrase of each advertisement to be recommended;
the matching module is used for respectively matching each attribute label of the extended phrase of the current search word with each attribute label of the extended phrase of each advertisement to be recommended and calculating to obtain an attribute similarity score of each advertisement to be recommended for the current search word;
the adding module is used for adding or weighting and adding the word similarity score and the attribute similarity score of each advertisement to be recommended aiming at the current search word to obtain the comprehensive score of each advertisement to be recommended aiming at the current search word;
the attribute tag includes: a subject label, a type label, and an industry label;
the acquisition module is specifically configured to:
respectively establishing a main body model, a type model and an industry model according to historical data;
respectively inputting the extended phrases of the current search word into a theme model, a type model and an industry model, and outputting to obtain a corresponding theme label, a type label and an industry label;
the matching module is specifically configured to:
comparing the subject label of the extended phrase of the current search word with the subject label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the first preset score is added into the theme score of the current advertisement to be recommended aiming at the current search word;
comparing the type label of the extended phrase of the current search word with the type label of the extended phrase of each advertisement to be recommended;
if the theme label of the expanded phrase of the current search word is the same as the theme label of the expanded phrase of the current advertisement to be recommended, the second preset score is added to the type score of the current advertisement to be recommended aiming at the current search word;
comparing the industry label of the extended phrase of the current search word with the industry label of the extended phrase of each advertisement to be recommended;
if the industry label of the expanded phrase of the current search word is the same as the industry label of the expanded phrase of the current advertisement to be recommended, the third preset score is added to the industry score of the current advertisement to be recommended for the current search word;
and adding or weighting and adding the theme score, the type score and the industry score of the current advertisement to be recommended aiming at the current search word to obtain the attribute similarity score of the current advertisement to be recommended aiming at the current search word.
7. The search term based advertisement recommendation system according to claim 6, further comprising:
the judging unit is used for triggering the expansion unit when judging that the current search word input by the user is a hot search word; the hot search word is a search word with the word frequency exceeding the preset word frequency.
8. The search term-based advertisement recommendation system according to claim 6 or 7, wherein the expansion unit is specifically configured to:
obtaining an expansion word of a current search word, wherein the expansion word comprises but is not limited to: related words, impression words, synonyms, co-occurrence words, associated words and log expanded words;
and combining the current search word and the corresponding expansion word into an expansion phrase of the current search word.
9. The search term based advertisement recommendation system according to claim 6 or 7, further comprising:
the first filtering unit is used for sequentially carrying out word segmentation, single word removal and stop word removal on the extended word group of the current search word;
and the second filtering unit is used for sequentially performing word segmentation, word removal and stop word removal on the extended word group corresponding to each text keyword of the advertisement to be recommended.
10. The search term-based advertisement recommendation system according to claim 6 or 7, wherein the recommendation unit is specifically configured to:
sequencing each advertisement to be recommended from high to low according to the comprehensive score of the advertisement to be recommended for the current search term;
taking the first n (n is more than 0) advertisements to be recommended as recommended advertisements;
and displaying the recommended advertisements on the corresponding pages of the current search terms according to the sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810392779.4A CN108776901B (en) | 2018-04-27 | 2018-04-27 | Advertisement recommendation method and system based on search terms |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810392779.4A CN108776901B (en) | 2018-04-27 | 2018-04-27 | Advertisement recommendation method and system based on search terms |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108776901A CN108776901A (en) | 2018-11-09 |
| CN108776901B true CN108776901B (en) | 2021-01-15 |
Family
ID=64026649
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810392779.4A Active CN108776901B (en) | 2018-04-27 | 2018-04-27 | Advertisement recommendation method and system based on search terms |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108776901B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111383069A (en) * | 2018-12-29 | 2020-07-07 | 美宅科技（北京）有限公司 | Furniture assembly recommendation method and device |
| CN109885753A (en) * | 2019-01-16 | 2019-06-14 | 苏宁易购集团股份有限公司 | A kind of method and device for expanding commercial articles searching and recalling |
| CN110222147A (en) * | 2019-05-15 | 2019-09-10 | 北京百度网讯科技有限公司 | Label extending method, device, computer equipment and storage medium |
| CN110781204B (en) * | 2019-09-09 | 2024-02-20 | 腾讯大地通途（北京）科技有限公司 | Identification information determining method, device, equipment and storage medium of target object |
| CN110706021A (en) * | 2019-09-12 | 2020-01-17 | 微梦创科网络科技（中国）有限公司 | Advertisement putting method and system |
| CN113111216B (en) * | 2020-01-13 | 2023-11-03 | 百度在线网络技术（北京）有限公司 | Advertisement recommendation method, device, equipment and storage medium |
| CN112256822A (en) * | 2020-10-21 | 2021-01-22 | 平安科技（深圳）有限公司 | Text search method and device, computer equipment and storage medium |
| CN113297489A (en) * | 2021-05-28 | 2021-08-24 | 平安科技（深圳）有限公司 | Rehabilitation aid recommendation method and device, computer equipment and storage medium |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063468A (en) * | 2010-12-03 | 2011-05-18 | 百度在线网络技术（北京）有限公司 | Device and method for confirming query sort of query sequence |
| CN102254039A (en) * | 2011-08-11 | 2011-11-23 | 武汉安问科技发展有限责任公司 | Searching engine-based network searching method |
| CN102855252A (en) * | 2011-06-30 | 2013-01-02 | 北京百度网讯科技有限公司 | Method and device for data retrieval based on demands |
| CN102999625A (en) * | 2012-12-05 | 2013-03-27 | 北京海量融通软件技术有限公司 | Method for realizing semantic extension on retrieval request |
| CN103853722A (en) * | 2012-11-29 | 2014-06-11 | 腾讯科技（深圳）有限公司 | Query based keyword extension method, device and system |
| CN103914543A (en) * | 2014-04-03 | 2014-07-09 | 北京百度网讯科技有限公司 | Search result displaying method and device |
| CN104199822A (en) * | 2014-07-11 | 2014-12-10 | 五八同城信息技术有限公司 | Method and system for identifying demand classification corresponding to searching |
| CN105912630A (en) * | 2016-04-07 | 2016-08-31 | 北京搜狗科技发展有限公司 | Information expansion method and device |
| CN105975596A (en) * | 2016-05-10 | 2016-09-28 | 上海珍岛信息技术有限公司 | Query expansion method and system of search engine |
| CN106528588A (en) * | 2016-09-14 | 2017-03-22 | 厦门幻世网络科技有限公司 | Method and apparatus for matching resources for text information |
| CN106557476A (en) * | 2015-09-24 | 2017-04-05 | 北京奇虎科技有限公司 | The acquisition methods and device of relevant information |
| CN106610972A (en) * | 2015-10-21 | 2017-05-03 | 阿里巴巴集团控股有限公司 | Query rewriting method and apparatus |
| CN107544982A (en) * | 2016-06-24 | 2018-01-05 | 中兴通讯股份有限公司 | Text message processing method, device and terminal |
| CN107609152A (en) * | 2017-09-22 | 2018-01-19 | 百度在线网络技术（北京）有限公司 | Method and apparatus for expanding query formula |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567408B (en) * | 2010-12-31 | 2014-06-04 | 阿里巴巴集团控股有限公司 | Method and device for recommending search keyword |
-
2018
- 2018-04-27 CN CN201810392779.4A patent/CN108776901B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063468A (en) * | 2010-12-03 | 2011-05-18 | 百度在线网络技术（北京）有限公司 | Device and method for confirming query sort of query sequence |
| CN102855252A (en) * | 2011-06-30 | 2013-01-02 | 北京百度网讯科技有限公司 | Method and device for data retrieval based on demands |
| CN102254039A (en) * | 2011-08-11 | 2011-11-23 | 武汉安问科技发展有限责任公司 | Searching engine-based network searching method |
| CN103853722A (en) * | 2012-11-29 | 2014-06-11 | 腾讯科技（深圳）有限公司 | Query based keyword extension method, device and system |
| CN102999625A (en) * | 2012-12-05 | 2013-03-27 | 北京海量融通软件技术有限公司 | Method for realizing semantic extension on retrieval request |
| CN103914543A (en) * | 2014-04-03 | 2014-07-09 | 北京百度网讯科技有限公司 | Search result displaying method and device |
| CN104199822A (en) * | 2014-07-11 | 2014-12-10 | 五八同城信息技术有限公司 | Method and system for identifying demand classification corresponding to searching |
| CN106557476A (en) * | 2015-09-24 | 2017-04-05 | 北京奇虎科技有限公司 | The acquisition methods and device of relevant information |
| CN106610972A (en) * | 2015-10-21 | 2017-05-03 | 阿里巴巴集团控股有限公司 | Query rewriting method and apparatus |
| CN105912630A (en) * | 2016-04-07 | 2016-08-31 | 北京搜狗科技发展有限公司 | Information expansion method and device |
| CN105975596A (en) * | 2016-05-10 | 2016-09-28 | 上海珍岛信息技术有限公司 | Query expansion method and system of search engine |
| CN107544982A (en) * | 2016-06-24 | 2018-01-05 | 中兴通讯股份有限公司 | Text message processing method, device and terminal |
| CN106528588A (en) * | 2016-09-14 | 2017-03-22 | 厦门幻世网络科技有限公司 | Method and apparatus for matching resources for text information |
| CN107609152A (en) * | 2017-09-22 | 2018-01-19 | 百度在线网络技术（北京）有限公司 | Method and apparatus for expanding query formula |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108776901A (en) | 2018-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108776901B (en) | Advertisement recommendation method and system based on search terms | |
| US11403680B2 (en) | Method, apparatus for evaluating review, device and storage medium | |
| CN109145153B (en) | Intention category identification method and device | |
| CN107704512B (en) | Financial product recommendation method based on social data, electronic device and medium | |
| US10515125B1 (en) | Structured text segment indexing techniques | |
| CN106462604B (en) | Identifying query intent | |
| CN109637537B (en) | Method for automatically acquiring annotated data to optimize user-defined awakening model | |
| KR102170206B1 (en) | Information Search System and Method using keyword and relation information | |
| CN111241267A (en) | Abstract extraction and abstract extraction model training method, related device and storage medium | |
| CN111159363A (en) | Knowledge base-based question answer determination method and device | |
| CN105095204A (en) | Method and device for obtaining synonym | |
| CN109858626B (en) | Knowledge base construction method and device | |
| CN111414757B (en) | Text recognition method and device | |
| CN106776566B (en) | Method and device for recognizing emotion vocabulary | |
| CN115438166A (en) | Keyword and semantic-based searching method, device, equipment and storage medium | |
| CN106777236B (en) | Method and device for displaying query result based on deep question answering | |
| CN111310440A (en) | Text error correction method, device and system | |
| CN111160007B (en) | Search method and device based on BERT language model, computer equipment and storage medium | |
| CN110096572B (en) | Sample generation method, device and computer readable medium | |
| CN109255022B (en) | Automatic abstract extraction method for network articles | |
| CA3233457A1 (en) | Machine learning-implemented chat bot database query system for multi-format database queries | |
| CN106021532B (en) | Keyword display method and device | |
| CN107239455B (en) | Core word recognition method and device | |
| CN111241290A (en) | Comment tag generation method and device and computing equipment | |
| CN111985212A (en) | Text keyword recognition method and device, computer equipment and readable storage medium |
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 |