CN105956189B - Information recommendation method and device based on artificial intelligence - Google Patents
Information recommendation method and device based on artificial intelligence Download PDFInfo
- Publication number
- CN105956189B CN105956189B CN201610405228.8A CN201610405228A CN105956189B CN 105956189 B CN105956189 B CN 105956189B CN 201610405228 A CN201610405228 A CN 201610405228A CN 105956189 B CN105956189 B CN 105956189B
- Authority
- CN
- China
- Prior art keywords
- recommendation
- association
- lexical analysis
- result
- environment
- 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)
Abstract
The invention provides an information recommendation method and device based on artificial intelligence, wherein the information recommendation method based on artificial intelligence comprises the following steps: receiving a search term; performing deep lexical analysis on the search words, and abstracting the search words according to the deep lexical analysis result; and acquiring a recommendation result according to the deep lexical analysis result and the search word abstract result. The method can improve the accuracy of the recommendation result.
Description
Technical Field
The invention relates to the technical field of internet, in particular to an artificial intelligence-based information recommendation method and device.
Background
Artificial Intelligence (AI) is a new technical science to study and develop theories, methods, techniques and application systems for simulating, extending and expanding human Intelligence. Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence, and research in this field includes intelligent ordering robots, language recognition, image recognition, natural language processing, and expert systems, among others.
Search engine recommendation systems are generally implemented by mining statistical associations between search terms (queries) and documents, and between queries and queries. However, the query-level statistical method is not satisfactory for processing long and cold queries, and the obtained result is not ideal.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, an object of the present invention is to provide an artificial intelligence-based information recommendation method, which can improve the accuracy of recommendation results.
Another objective of the present invention is to provide an artificial intelligence-based information recommendation apparatus.
In order to achieve the above object, an artificial intelligence based information recommendation method provided in an embodiment of a first aspect of the present invention includes: receiving a search term; performing deep lexical analysis on the search words, and abstracting the search words according to the deep lexical analysis result; and acquiring a recommendation result according to the deep lexical analysis result and the search word abstract result.
According to the information recommendation method based on artificial intelligence, disclosed by the embodiment of the first aspect of the invention, element-level recommendation can be realized by performing deep lexical analysis and search word abstraction on the search words, and the accuracy of a recommendation result can be improved compared with the search word-level recommendation.
In order to achieve the above object, an artificial intelligence-based information recommendation apparatus according to an embodiment of a second aspect of the present invention includes: the receiving module is used for receiving the search terms; the analysis module is used for carrying out deep lexical analysis on the search words and carrying out search word abstraction according to the deep lexical analysis result; and the recommending module is used for acquiring recommending results according to the deep lexical analysis results and the search term abstract results.
The information recommendation device based on artificial intelligence provided by the embodiment of the second aspect of the invention can realize element-level recommendation by performing deep lexical analysis and search term abstraction on the search term, and can improve the accuracy of a recommendation result compared with search term-level recommendation.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flowchart illustrating an artificial intelligence based information recommendation method according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating an artificial intelligence based information recommendation method according to another embodiment of the present invention;
FIG. 3 is a schematic flow chart of constructing a database of associated elements according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an online recommendation process in an embodiment of the invention;
FIG. 5 is a schematic structural diagram of an artificial intelligence-based information recommendation apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an artificial intelligence-based information recommendation apparatus according to another embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar modules or modules having the same or similar functionality throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention. On the contrary, the embodiments of the invention include all changes, modifications and equivalents coming within the spirit and terms of the claims appended hereto.
Fig. 1 is a flowchart illustrating an artificial intelligence based information recommendation method according to an embodiment of the present invention.
Referring to fig. 1, the method of the present embodiment includes:
s11: a search term is received.
The user can input the search terms into the client in the forms of text, voice and the like, and then the client can send the search terms to the server.
It is understood that when the user inputs the search word in a form of voice, voice recognition may be performed at the client or the server, and recognized as text.
The search term may be a hot search term, such as query, "iphone 4s money from the beginning", or a cold search term, such as query, "gallaxys 6edge + money from the beginning".
S12: and carrying out deep lexical analysis on the search words, and abstracting the search words according to the deep lexical analysis result.
When the search term is a cold search term, such as "gallaxys 6edge + first money" described above, the recommendation results obtained according to the statistical manner of the related art are: samsung s6edge, galaxy s6edge5.7, galaxy s6edge release, etc., but in reality, the user is most concerned about the prices of "galaxys 6edge +" original price "and" just coming to the market "of the mobile phone, but the recommendation result is not reflected; the "present price" related to the concept of "original price" is not shown either.
When the search term is a popular term, for example, the mobile phone model is changed into iphone4s, iphone5 and the like, the original price and the current price recommendation result can be obtained.
Therefore, a recommendation method using statistical information at the level of a general search term (query) cannot obtain a good recommendation result for a cold search term.
In order to obtain a good recommendation result for the cold search term, in this embodiment, deep lexical analysis and search term abstraction are performed on the search term to implement element-level recommendation.
The deep lexical analysis is to analyze important elements in the search term, for example, to perform word segmentation on the search term, and identify that the word segmentation is an entity, time, domain concept or common term, so as to determine the word segmentation of the entity, time, and domain concept as the important elements.
The term abstraction is to abstract each word segment in the deep lexical analysis result, for example, the category of the entity is used to replace the concrete entity, the time element is used to replace the concrete time, and the field concept element is used to replace the concrete field concept.
The above-mentioned domain concepts refer to specific concepts in a certain domain, and the domain concepts in different domains may be determined by means of statistics in advance and the like.
The flow of detailed deep lexical analysis and search term abstraction can be seen in the following description.
S13: and acquiring a recommendation result according to the deep lexical analysis result and the search word abstract result.
After deep lexical analysis and search word abstraction, the search words can be converted into element levels, and then element level recommendation can be achieved, so that the accuracy of recommendation results is improved.
The specific process of obtaining the recommendation result can be referred to the following description.
In addition, after obtaining the recommendation result, the server may send the recommendation result to the client, and the client displays the recommendation result to the user, for example, in a part of "related search".
In the embodiment, element-level recommendation can be realized by performing deep lexical analysis and search word abstraction on the search words, and the accuracy of a recommendation result can be improved compared with search word-level recommendation.
The previous embodiment describes a process of online recommendation, and in actual implementation, online recommendation may be performed according to an associated element database constructed offline, so some embodiments may further include a process of constructing an associated element database offline.
Fig. 2 is a flowchart illustrating an artificial intelligence based information recommendation method according to another embodiment of the present invention.
The method of the embodiment can be divided into two parts of off-line mining and on-line computing.
The offline calculation part is used for mining the association relation among query elements in the historical queries to form an association element database and provide a statistical basis for online calculation.
The online calculation is used for realizing element-based recommendation according to the associated element database determined by the offline calculation part when the user searches the query.
Specifically, referring to fig. 2, the process of this embodiment includes:
s21: a historical query is collected.
The historical query refers to search terms generated in a historical stage of a large number of users.
Specifically, the query may be extracted from a history session (session) and used as the history query. The historical conversation is the reflection of user behaviors, and embodies the process that a user clearly determines own requirements or arouses new requirements.
S22: and carrying out important element association mining on the historical search words based on deep lexical analysis to construct an association element database.
For details, reference may be made to the subsequent flow.
It is understood that S21-S22 may be done offline, thereby building a database of associated elements.
S23: and receiving the query.
Here, query at this time is a search word at the time of online search.
S24: the query is analyzed.
Wherein the analysis comprises: deep lexical analysis and search term abstraction according to the deep lexical analysis result.
S25: and recommending according to the analysis result and the associated element database to obtain a recommendation result.
The recommendation can be referred to as an accurate recommendation based on deep lexical analysis.
It is understood that S23-S25 are online processes.
Therefore, the above-mentioned accurate recommendation can be realized by combining the above-mentioned offline flow and online flow shown in fig. 2.
In the process shown in fig. 2, the process of constructing the association element database is involved.
In some embodiments, referring to fig. 3, performing important element association mining based on deep lexical analysis on the historical search terms, and the process of constructing the association element database may include:
s31: and performing deep lexical analysis on each historical search word to obtain a deep lexical analysis result corresponding to each historical search word.
It will be appreciated that the same principles can be applied to deep lexical analysis of historical queries either off-line or received on-line. Therefore, deep lexical analysis of a query will be described below without distinguishing between historical queries or online received queries, but collectively referred to as queries.
Wherein, deep lexical analysis is used for positioning important elements such as entities, time, field concepts and the like in the query. For example:
query ═ iphone6s jail crossing "
And (3) analysis results: iphone6sEntityPrison crossingConcept of domain
Query is a 2016 government work report in Beijing City "
And (3) analysis results: beijing market-Entity2016/Time of dayGovernment work reportConcept of domain
S32: and abstracting the search words according to the deep lexical analysis result to obtain an abstract result corresponding to each historical search word.
Similar to deep lexical analysis, the search term abstraction may also be collectively referred to as query without distinguishing historical query or online received query, and describes a search term abstraction flow of a deep lexical analysis result of query.
The problem of data sparseness on long-tail and cold-door queries can be solved through query abstraction.
The Query abstraction means that an entity in a deep lexical analysis result replaces a concrete entity by an entity category, and a time and field concept abstract element replaces a concrete time and field concept.
The Query abstraction may be a full abstraction or a partial abstraction. For example:
query ═ iphone6s jail crossing "
And (3) abstracting a result: 'ELEM-mobile phone jail-off', 'iphone 6s ELEM-field concept'
Query is a 2016 government work report in Beijing City "
And (3) abstracting a result: "ELEM _ Place name 2016 government work report", "ELEM _ Place name ELEM _ time government work report" … …
S33: and acquiring the associated elements according to the deep lexical analysis result and/or the abstract result corresponding to any two historical search words.
Wherein the general form of the associated element is: (rewrite | context), i.e., under the constraint of a certain "context", there is a specific "rewrite" relationship.
Taking the obtaining of the association element according to the deep lexical analysis result as an example, the process of obtaining the association element may specifically include:
obtaining deep lexical analysis results corresponding to any two historical search terms;
if the two deep lexical analysis results contain the same important elements, the same important elements are determined as the environment part of the related elements, and different parts are determined as the rewriting parts of the related elements.
Further, a flag for marking the rewriting position may be provided in the environment section.
For example, assume any two historical queries (denoted as q)1、q2) The results of deep lexical analysis of (1) are represented by L (q)1)、L(q2) And (4) showing.
Then first judge L (q)1)、L(q2) Whether or not there is at least one of the same important elements between, there beingAnd executing the subsequent flow in time, and ending the discovery flow if the subsequent flow is not executed.
When at least one same important element exists between L (q1) and L (q2), the association relationship is determined according to L (q1) and L (q 2). For example:
L(q1)＝iphone6s/entityPrison crossingConcept of domain
L(q2)＝iphone5s/EntityPrison crossingConcept of domain
And (3) association elements: (
|<SLOT>Cross prison)
< SLOT > defines the location where overwriting occurs.
Another example is:
l (q1) ═ Beijing-Entity2016/Time of dayGovernment work reportConcept of domain
L (q2) ═ starch zone-Entity2016/Time of dayGovernment work reportConcept of domain
And (3) association elements: (|<SLOT>2016 government work report)
While the above description has been given of the process of obtaining the association element according to the deep lexical analysis result, it is understood that the process of obtaining the association element according to the abstract result may also be executed by reference.
Suppose an arbitrary pair of queries (denoted as q)1、q2) The query abstract results of (1) are respectively represented by A (q)1)、A(q2) And (4) showing.
For example:
a (q1) ═ ELEM _ cell phone is at first-Time of dayHow much money is basedConcept of domain
A (q2) ═ ELEM _ cell phone original value-Concept of domain
And (3) association elements: (
I ELEM mobile phone<SLOT>)
Another example is:
a (q1) ═ Beijing-EntityELEM time government work reportConcept of domain
A (q2) is based on the area starch ═ sea-EntityELEM time government work reportConcept of domainFull text/common term
And (3) association elements: ( |<SLOT1>ELEM _ time<SLOT2>)
S34: and counting the frequency of the associated elements, setting the weight of the associated elements according to the frequency, and correspondingly storing the associated elements and the frequency into an associated element database.
After the associated elements are obtained, the frequencies of the associated elements may be counted, and when the frequency of one associated element is higher, a larger weight is set. The association elements and weight correspondences may then be stored in an association element database.
Further, the associated element may use the environment portion as an index when being stored, so that a corresponding rewritten portion may be queried according to the environment portion in a subsequent flow.
Therefore, the construction of the association element database is realized through the above-mentioned flow of fig. 3, and the constructed association element database can be used for online recommendation as shown in fig. 2.
In some embodiments, referring to fig. 4, the process of recommending according to the deep lexical analysis result and the search term abstraction result and the association element database may include:
s41: and determining the environment part of the search word according to the deep lexical analysis result and the abstract result of the search word, and calculating the environment score of the environment part of each search word.
After deep lexical analysis and search term abstraction are performed on the query received on line, the deep lexical analysis result and the search term abstraction result are obtained.
For example, the query received online is: when q is "gallaxys 6edge + initial money", the deep lexical analysis can obtain:
galaxys6edge+/entityMost initially-Time of dayHow much money is basedConcept of domain
When a search term is abstracted, important elements can be fully or partially abstracted, so that one abstract result is:
initial amount of money for ELEM _ handset
Of course, the search term may include a common term in addition to the above-mentioned important elements such as the entity, time, and domain concepts, and thus may include a common term in the abstract result.
After the deep lexical analysis results and the abstract results are obtained, one or more of the above components (including the important elements, abstract important elements, and common term therein) may be replaced with < SLOT > placeholders, forming an environmental portion of the search term.
For example, replacing the ELEM _ handset described above with < SLOT >, one can get how much the environmental part is e (q) < SLOT > the first money.
In addition, the score s (e (q)) of the corresponding environment section can be calculated according to the degree of importance of the < SLOT > replaced section in the search term.
S42: selecting the related elements with the same environmental parts as the environmental parts of the search words according to the related elements stored in the related element database, acquiring the corresponding weights of the selected related elements, and acquiring the recommendation results according to the selected related elements, the weights and the environmental scores.
The associated elements in the associated element database can be stored by using the environment part as an index, so that the corresponding associated elements can be queried according to the environment part.
For example, if the environment part of the search term is the first amount of money of < SLOT >, the associated element of how much the environment part is the first amount of money of < SLOT > can be obtained.
Further, obtaining a recommendation according to the selected association element, the weight, and the environment score may include:
generating a preliminary recommendation result according to the rewriting part of the selected related element;
corresponding to each preliminary recommendation result, calculating a corresponding recommendation score according to the corresponding weight and the environment score;
and selecting a preset number of preliminary recommendation results according to the recommendation scores, and determining the preliminary recommendation results as recommendation results to be displayed so as to display the recommendation results to be displayed.
Generating a preliminary recommendation result according to the rewriting part of the incidence relation, wherein the preliminary recommendation result can comprise: and combining the rewriting part with the environment part to obtain a preliminary recommendation result.
For example, one selected associated element (r | e (q)), (
|<SLOT>First amount of money), a preliminary recommendation may be generated based on combining the rewrite part "iphone 6s, iphone5 s" with the environment part "first amount of money", such that the preliminary recommendation includes: how much was the first of the iphone6s, and how much was the first of the iphone5 s.
After each preliminary recommendation result is obtained, a recommendation score f for each preliminary recommendation result may be calculated based on s (e (q)), the weight w (r | e (q)) of the corresponding associated element. Then, a preset number of preliminary recommendation results may be selected as recommendation results to be presented according to the recommendation score f from high to low, and then the recommendation results to be presented may be presented at the client.
Thus, online recommendation is achieved through the above-described flow of fig. 4.
Fig. 5 is a schematic structural diagram of an artificial intelligence-based information recommendation apparatus according to an embodiment of the present invention.
Referring to fig. 5, the apparatus 50 of the present embodiment includes: a receiving module 51, an analyzing module 52 and a recommending module 53.
A receiving module 51, configured to receive a search term;
the analysis module 52 is configured to perform deep lexical analysis on the search terms and perform search term abstraction according to a result of the deep lexical analysis;
and the recommending module 53 is configured to obtain a recommending result according to the deep lexical analysis result and the search term abstract result.
In some embodiments, referring to fig. 6, the apparatus 50 further comprises:
a collection module 54 for collecting historical search terms;
and the mining module 55 is configured to perform important element association mining based on deep lexical analysis on the historical search terms, and construct an associated element database.
In some embodiments, referring to fig. 6, the excavation module 55 includes:
the analysis submodule 551 is used for performing deep lexical analysis on each historical search word to obtain a deep lexical analysis result corresponding to each historical search word;
the abstraction submodule 552 is configured to abstract the search terms according to the deep lexical analysis result to obtain an abstraction result corresponding to each historical search term;
the association discovery sub-module 553, configured to obtain association elements according to the deep lexical analysis result and/or the abstract result corresponding to any two historical search terms;
the statistic submodule 554 is configured to count the frequency of the associated element, set the weight of the associated element according to the frequency, and correspondingly store the associated element and the frequency in an associated element database.
In some embodiments, the association element includes an environment part and a rewrite part, and referring to fig. 6, the recommending module 53 includes:
an environment determination submodule 531, configured to determine an environment part of the search term according to the deep lexical analysis result and the search term abstraction result, and calculate an environment score of the environment part of each search term;
the association query submodule 532 is configured to select, according to the association elements stored in the association element database, an association element having an environment part that is the same as the environment part of the search term, obtain a weight corresponding to the selected association element, and obtain a recommendation result according to the selected association element, the weight, and the environment score.
In some embodiments, the association query sub-module 532 is configured to obtain recommendation results according to the selected association elements, the weights and the environment scores, and includes:
generating a preliminary recommendation result according to the rewriting part of the selected related element;
corresponding to each preliminary recommendation result, calculating a corresponding recommendation score according to the corresponding weight and the environment score;
and selecting a preset number of preliminary recommendation results according to the recommendation scores, and determining the preliminary recommendation results as recommendation results to be displayed so as to display the recommendation results to be displayed.
It is understood that the apparatus of the present embodiment corresponds to the method embodiment described above, and specific contents may be referred to related descriptions in the method embodiment, and are not described in detail herein.
In the embodiment, element-level recommendation can be realized by performing deep lexical analysis and search word abstraction on the search words, and the accuracy of a recommendation result can be improved compared with search word-level recommendation.
It should be noted that the terms "first," "second," and the like in the description of the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Further, in the description of the present invention, the meaning of "a plurality" means at least two unless otherwise specified.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (6)
1. An information recommendation method based on artificial intelligence is characterized by comprising the following steps:
receiving a search term;
performing deep lexical analysis on the search words, and abstracting the search words according to the deep lexical analysis result;
obtaining a recommendation result according to the deep lexical analysis result and the search term abstraction result, wherein a pre-constructed association element database is obtained, wherein the association element database comprises association elements and weights, the association elements comprise an environment part and a rewriting part, the environment part comprises the same part in the deep lexical analysis result, the environment part of the search term is determined according to the deep lexical analysis result and the search term abstraction result, an environment score of the environment part of each search term is calculated, the association elements with the environment part being the same as the environment part of the search term are selected according to the association elements stored in the association element database, the corresponding weights of the selected association elements are obtained, and the recommendation result is obtained according to the selected association elements, the weights and the environment scores,
wherein the obtaining of recommendation results according to the selected association elements, the weights and the environment scores comprises: combining the rewriting part and the environment part of the selected associated elements to generate a preliminary recommendation result; corresponding to each preliminary recommendation result, calculating a corresponding recommendation score according to the corresponding weight and the environment score; and selecting a preset number of preliminary recommendation results according to the recommendation scores, and determining the preliminary recommendation results as recommendation results to be displayed so as to display the recommendation results to be displayed.
2. The method of claim 1, further comprising:
collecting historical search terms;
and carrying out important element association mining on the historical search words based on deep lexical analysis to construct an association element database.
3. The method of claim 2, wherein performing deep lexical analysis-based important element association mining on the historical search terms to construct an association element database comprises:
performing deep lexical analysis on each historical search word to obtain a deep lexical analysis result corresponding to each historical search word;
abstracting the search words according to the deep lexical analysis result to obtain an abstract result corresponding to each historical search word;
acquiring association elements according to the deep lexical analysis results and/or abstract results corresponding to any two historical search words;
and counting the frequency of the associated elements, setting the weight of the associated elements according to the frequency, and correspondingly storing the associated elements and the frequency into an associated element database.
4. An artificial intelligence-based information recommendation device, comprising:
the receiving module is used for receiving the search terms;
the analysis module is used for carrying out deep lexical analysis on the search words and carrying out search word abstraction according to the deep lexical analysis result;
a recommending module for obtaining recommending results according to the deep lexical analysis results and the abstract search word results,
the method comprises the steps of obtaining a pre-constructed associated element database, wherein the associated element database comprises associated elements and weights, the associated elements comprise an environment part and a rewriting part, the environment part comprises the same part in the deep lexical analysis result, and the recommending module comprises:
the environment determining submodule is used for determining the environment part of the search word according to the deep lexical analysis result and the abstract result of the search word and calculating the environment score of the environment part of each search word;
the association query submodule is used for selecting association elements with the same environment parts as the environment parts of the search terms according to the association elements stored in the association element database, acquiring corresponding weights of the selected association elements and acquiring recommendation results according to the selected association elements, the weights and the environment scores,
the association query submodule is configured to obtain a recommendation result according to the selected association element, the weight, and the environment score, and includes: combining the rewriting part and the environment part of the selected associated elements to generate a preliminary recommendation result; corresponding to each preliminary recommendation result, calculating a corresponding recommendation score according to the corresponding weight and the environment score; and selecting a preset number of preliminary recommendation results according to the recommendation scores, and determining the preliminary recommendation results as recommendation results to be displayed so as to display the recommendation results to be displayed.
5. The apparatus of claim 4, further comprising:
the collection module is used for collecting historical search terms;
and the mining module is used for performing important element association mining on the historical search words based on deep lexical analysis and constructing an associated element database.
6. The apparatus of claim 5, wherein the excavation module comprises:
the analysis submodule is used for carrying out deep lexical analysis on each historical search word to obtain a deep lexical analysis result corresponding to each historical search word;
the abstract submodule is used for abstracting the search words according to the deep lexical analysis result to obtain an abstract result corresponding to each historical search word;
the association discovery submodule is used for acquiring association elements according to the deep lexical analysis results and/or the abstract results corresponding to any two historical search words;
and the statistic submodule is used for counting the frequency of the associated elements, setting the weight of the associated elements according to the frequency and correspondingly storing the associated elements and the frequency into an associated element database.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610405228.8A CN105956189B (en) | 2016-06-08 | 2016-06-08 | Information recommendation method and device based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610405228.8A CN105956189B (en) | 2016-06-08 | 2016-06-08 | Information recommendation method and device based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105956189A CN105956189A (en) | 2016-09-21 |
| CN105956189B true CN105956189B (en) | 2020-01-07 |
Family
ID=56907987
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610405228.8A Active CN105956189B (en) | 2016-06-08 | 2016-06-08 | Information recommendation method and device based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105956189B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107145525B (en) * | 2017-04-14 | 2020-10-16 | 北京星选科技有限公司 | Data processing method for confirming search scene, search method and corresponding device |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103136257B (en) * | 2011-11-30 | 2016-09-28 | 阿里巴巴集团控股有限公司 | Information providing method and device thereof |
| CN102402619B (en) * | 2011-12-23 | 2013-10-30 | 广东威创视讯科技股份有限公司 | Search method and device |
| CN103353894A (en) * | 2013-07-19 | 2013-10-16 | 武汉睿数信息技术有限公司 | Data searching method and system based on semantic analysis |
-
2016
- 2016-06-08 CN CN201610405228.8A patent/CN105956189B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105956189A (en) | 2016-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11397772B2 (en) | Information search method, apparatus, and system | |
| EP3617952A1 (en) | Information search method, apparatus and system | |
| JP6515624B2 (en) | Method of identifying lecture video topics and non-transitory computer readable medium | |
| US20170010123A1 (en) | Hybrid road network and grid based spatial-temporal indexing under missing road links | |
| CN106126582A (en) | Recommend method and device | |
| CN106126524B (en) | Information pushing method and device | |
| US20180203857A1 (en) | Automated data exploration and validation | |
| CN109241525B (en) | Keyword extraction method, device and system | |
| US8380654B2 (en) | General market prediction using position specification language | |
| CN103778548A (en) | Goods information and keyword matching method, and goods information releasing method and device | |
| CN110532351B (en) | Recommendation word display method, device and equipment and computer readable storage medium | |
| CN109325179A (en) | A kind of method and device that content is promoted | |
| CN106471496A (en) | From inquiry is expressed again, identification is intended in the search | |
| CN116127020A (en) | Method for training generated large language model and searching method based on model | |
| CN116894711A (en) | Commodity recommendation reason generation method and device and electronic equipment | |
| CN110825889A (en) | Propaganda information interaction method and device, electronic equipment and storage medium | |
| CN111191133B (en) | Service search processing method, device and equipment | |
| CN104035964A (en) | Method and device for providing credit-related information | |
| CN111680213A (en) | Information recommendation method, data processing method and device | |
| CN105956189B (en) | Information recommendation method and device based on artificial intelligence | |
| CN105550282A (en) | User interest forecasting method by utilizing multidimensional data | |
| CN111125379B (en) | Knowledge base expansion method and device, electronic equipment and storage medium | |
| CN116662495A (en) | Question-answering processing method, and method and device for training question-answering processing model | |
| CN116304155A (en) | Three-dimensional member retrieval method, device, equipment and medium based on two-dimensional picture | |
| CN105446971B (en) | A kind of information display method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |