CN111666448A - Search method, search device, electronic equipment and computer-readable storage medium - Google Patents
Search method, search device, electronic equipment and computer-readable storage medium Download PDFInfo
- Publication number
- CN111666448A CN111666448A CN202010318720.8A CN202010318720A CN111666448A CN 111666448 A CN111666448 A CN 111666448A CN 202010318720 A CN202010318720 A CN 202010318720A CN 111666448 A CN111666448 A CN 111666448A
- Authority
- CN
- China
- Prior art keywords
- search
- word
- words
- word segmentation
- result
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 230000011218 segmentation Effects 0.000 claims abstract description 277
- 238000010801 machine learning Methods 0.000 claims abstract description 59
- 238000012545 processing Methods 0.000 claims abstract description 39
- 238000004422 calculation algorithm Methods 0.000 claims description 57
- 238000013507 mapping Methods 0.000 claims description 31
- 239000013598 vector Substances 0.000 claims description 20
- 238000004891 communication Methods 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 9
- 238000004140 cleaning Methods 0.000 claims description 8
- 230000006399 behavior Effects 0.000 description 14
- 238000010586 diagram Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
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
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/73—Querying
- G06F16/732—Query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/23—Updating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/71—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The embodiment of the invention provides a searching method, a searching device, electronic equipment and a computer readable storage medium, wherein the method comprises the following steps: acquiring a target search word; performing word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result; searching data matched with the first segmentation result in a pre-established database; the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words. Therefore, the scheme of the invention can solve the problems of long search time and low search result accuracy during searching based on the existing word segmentation method to a certain extent.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a search method, an apparatus, an electronic device, and a computer-readable storage medium.
Background
An information flow background that provides a full library short video search based on the Elasticsearch (i.e., a Lucene (search engine) based search server) where the user can search by video ID, video title, etc. However, the default chinese word segmentation algorithm based on the Elasticsearch may not only be very accurate in the matching result, but also many of the matching results may not be really needed by the user, and returning too many results may also cause the time consumption to be long, and the user experience to be very bad. For example, a "machine learning algorithm video tutorial" searched by a user should not return results that contain only "machine" or only "learning" or only "algorithm", etc., but must contain results for the entire word of the "machine learning algorithm". However, the current default chinese word segmentation algorithm (e.g. ik segmenter) based on the Elasticsearch, the "machine learning algorithm" is divided into multiple words.
Therefore, based on the existing word segmentation method, the time consumption of searching is long and the accuracy of the searching result is low when searching is carried out.
Disclosure of Invention
Embodiments of the present invention provide a searching method, an apparatus, an electronic device, and a computer-readable storage medium, so as to solve the problems that, based on the existing word segmentation method, the time consumed by searching is long and the accuracy of the search result is low.
The specific technical scheme is as follows:
in a first aspect of the present invention, there is provided a search method, including:
acquiring a target search word;
performing word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result;
searching data matched with the first segmentation result in a pre-established database;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
In yet another aspect of the present invention, there is also provided a search apparatus, including:
the search word acquisition module is used for acquiring a target search word;
the first word segmentation module is used for carrying out word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result;
the searching module is used for searching data matched with the first segmentation result in a pre-established database;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
In another aspect of the present invention, there is also provided an electronic device, including a processor, a communication interface, a memory and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing the searching method when executing the program stored in the memory.
In yet another aspect of the present invention, there is also provided a computer-readable storage medium on which a computer program is stored, the program implementing the above-mentioned search method when executed by a processor.
In yet another aspect of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform any of the above-described search methods.
According to the searching method provided by the embodiment of the invention, through collecting the historical searching records of the user and based on a machine learning mode, the words for describing the searching target of the historical searching words are extracted from the historical searching records of the user, and the word segmentation library is constructed by the words, so that when searching is needed based on the searching words in the following, the word segmentation can be carried out on the target searching words according to the word segmentation library, and then the required data is searched in the pre-established database according to the word segmentation result. The words extracted from the historical search records and used for describing the search targets of the historical search words represent the search requirements of the user, so that when the word segmentation processing is performed on the subsequent target search words according to the word segmentation library, the words can be segmented according to the requirements of the user, unreasonable word segmentation is avoided to a certain extent, excessive search results cannot be returned, the search time is shortened to a certain extent, and the accuracy of the search results is improved.
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.
Fig. 1 is a flowchart illustrating steps of a searching method according to an embodiment of the present invention;
FIG. 2 is a flow chart illustrating steps of another searching method according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating steps in a process for creating a thesaurus according to an embodiment of the present invention;
FIG. 4 is a flowchart illustrating a specific implementation of a search method according to an embodiment of the present invention;
fig. 5 is a block diagram of a search apparatus according to an embodiment of the present invention;
FIG. 6 is a block diagram of another searching apparatus according to an embodiment of the present invention;
fig. 7 is a block diagram of an electronic device provided in an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention.
Fig. 1 is a search method according to an embodiment of the present invention. As shown in fig. 1, the search method may include the steps of:
step 101: and acquiring a target search word.
The embodiment of the invention relates to a searching method, so that when searching is carried out, a target searching word for searching needs to be acquired. The target search word is text information input by a user, such as a "machine learning algorithm video tutorial", a "word application video tutorial", and the like.
Step 102: and performing word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result.
The words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
Wherein, the history search record records that the search target of the user based on a certain search word is a certain target object, for example, after the user a inputs a search word "machine learning algorithm video course", the search result includes a video with "machine" two words in the video name, a video with "learning" two words in the video name, a video with "algorithm" two words in the video name, and a video with "machine learning algorithm" six words in the video name, and the user a only clicks the video with "machine learning algorithm" six words in the video name for watching, then the history search record: the search word is a machine learning algorithm video course, and if the user clicks and watches the machine learning algorithm video, the purpose of the search word of the machine learning algorithm video course is to search the video named as the machine learning algorithm.
Therefore, the target object to be searched by the search word can be analyzed and obtained from the historical search records, and further, the word describing the target object can be extracted from the historical search records, for example, if the search word is a "machine learning algorithm video course", and the user clicks and watches the historical search record of the "machine learning algorithm" video, the name of the target object to be searched by the search word, namely the "machine learning algorithm", can be extracted.
Therefore, the words extracted from the historical search records and used for describing the search targets of the historical search words represent the search requirements of the user. Therefore, the embodiment of the invention can reasonably divide the target search word input by the user according to the word division library, namely, divide the word according to the search requirement of the user, further can avoid returning too many search results to a certain extent, shortens the search time to a certain extent and improves the accuracy of the search results.
Step 103: and searching data matched with the first segmentation result in a pre-established database.
The first word segmentation result is obtained after the word segmentation processing is carried out on the target search word according to the word segmentation library, and the word in the word segmentation library represents the search requirement of the user, so that the first word segmentation result is reasonable word segmentation according to the search requirement of the user, and further, when the search is carried out in a pre-established database according to the first word segmentation result, data which are not needed by the user can be avoided to a certain extent, so that the search time is shortened, and the accuracy of the search result is improved.
For example, when the target search word is a "machine learning algorithm video tutorial", the first word segmentation result does not separate the three words of "machine", "learning", and "algorithm", but takes the "machine learning algorithm" as a complete word, and then the search can be performed in the database according to the complete word of the "machine learning algorithm", so that a video including only the word of "machine" is not returned, a video including only the word of "learning" is not returned, a video including only the word of "algorithm" is not returned, and a video including only the complete word of the "machine learning algorithm" is returned, so that the search method of the embodiment of the present invention shortens the search time to a certain extent, and improves the accuracy of the search result.
In summary, in the search method according to the embodiment of the present invention, by collecting the historical search records of the user, and based on the machine learning manner, the words describing the search targets of the historical search words are extracted from the historical search records of the user, and a word segmentation library is constructed from the words, so that when a search needs to be performed subsequently based on the search words, the target search words can be segmented according to the word segmentation library, and then the required data is searched in the pre-established database according to the word segmentation result. Therefore, the embodiment of the invention analyzes the historical search records of the user in advance to obtain the target object which the user wants to search based on the historical search words, so that the word segmentation library is constructed by the words describing the target object, and when the word segmentation processing is carried out on the subsequent target search words according to the word segmentation library, the word segmentation can be carried out according to the requirements of the user, the unreasonable word segmentation is avoided to a certain extent, so that excessive search results cannot be returned, the search time is shortened to a certain extent, and the accuracy of the search results is improved.
Optionally, the object to be stored in the database is a video, the parameter information is a video title, the word segmentation library includes words for describing the video title, and the target search word is a word for describing the video title; searching data matched with the first segmentation result in a pre-established database, wherein the searching comprises the following steps:
searching a word matched with the first word segmentation result in a pre-established database;
and determining the video with the mapping relation of the words matched with the first word segmentation result as the data matched with the first word segmentation result.
That is, for a search based on a video title, the search method of the embodiment of the present invention may be employed.
Specifically, for the search based on the video titles, the history search records of the user, that is, the history search words, and the titles of the videos selected to be watched by the user from the videos searched according to the history search words may be obtained in advance, and then the words describing the videos to be searched by the history search words are extracted from the history search records (that is, the history search words and the titles of the videos selected to be watched by the user), so as to construct a word segmentation library from the words. Therefore, when searching is needed to be carried out in the database based on the video titles subsequently, the target search word is obtained, then word segmentation processing is carried out on the target search word according to the word segmentation library, a first word segmentation result is obtained, and then the video title matched with the first word segmentation result is searched in the video titles stored in the database.
Therefore, when the searching method of the embodiment of the invention is adopted for searching based on the video titles, by collecting the historical search records of the user and analyzing the historical search records of the user, the words for describing the videos to be searched by the historical search words are extracted from the historical search records based on a machine learning mode, namely, the search target of the user is obtained through analyzing the historical search records, so that the video to be searched based on a certain historical search word is determined, then constructing a word segmentation library by the extracted words for describing the videos to be searched by the historical search words, and further searching by the search words based on the video titles according to the word segmentation library, reasonably segmenting the search words so as to return the videos really needed by the user and related videos, rather than returning much video information that is completely irrelevant to the user's search behavior as in the prior art.
Fig. 2 is another searching method provided in the embodiment of the present invention. As shown in fig. 2, the search method may include the steps of:
step 201: and according to the pre-established word segmentation library, performing word segmentation processing on the parameter information of the object to be stored in the pre-established database to obtain a fourth word segmentation result.
The parameter information of the object to be stored in the pre-established database comprises at least one of the name and the characteristics of the object.
In addition, the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words. The term extracted from the history search record and used for describing the search target of the history search term represents the search requirement of the user, when the term segmentation processing is carried out on the parameter information of the object to be stored in the database according to the term segmentation library, the term segmentation can be carried out according to the requirement of the user, for example, when the term segmentation processing is carried out on the video name of the video of the machine learning algorithm to be stored in the database, namely the machine learning algorithm, the three terms of the machine learning algorithm and the algorithm are not separated, but the machine learning algorithm is stored as a whole so as to facilitate the later search.
In addition, for the video search, the video is searched from a search library (the search library can be regarded as a database), and an Elasticsearch, namely a search server based on a search engine (Lucene), is commonly used at present. The information to be retrieved by the video is stored in the elastic search, such as the title, duration, director, and actors of the video. When the data is added into a search library, the fields, such as titles, which need to be participled and searched are well participled, then the mapping relation between each word obtained by the participle and the video is established (namely, the fact that each word is respectively shown in the titles of the videos is determined), and the mapping relation is stored in the inverted index, so that later-stage search is facilitated.
However, according to the embodiment of the present invention, when the video is stored in the search library, the titles of the video may be segmented according to the segmentation library, for example, for the video with the title of "machine learning algorithm", the "machine learning algorithm" is stored as a word in the inverted index.
Step 202: and establishing a first mapping relation between the fourth word segmentation result and the object to be stored in the database, and storing the object to be stored in the database, the fourth word segmentation result and the first mapping relation in the database.
In step 201, the parameter information of the object to be stored in the database is segmented according to the segmentation library, so that the parameter information of the object to be stored in the database is reasonably segmented according to the requirement of the user, and therefore when the subsequent search is performed in the database based on the search term, the accuracy of the information returned to the user is higher, irrelevant content is less, time consumption is better, and user experience is better.
Step 203: and acquiring a target search word.
Step 204: and performing word segmentation processing on the target search word according to the word segmentation library to obtain a first word segmentation result.
The words in the word segmentation library are target words extracted from the historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words, so that the words extracted from the historical search records and used for describing the search targets of the historical search words represent search requirements of the user. Therefore, the embodiment of the invention can reasonably divide the target search word input by the user according to the word division library, namely, divide the word according to the search requirement of the user, and further can avoid returning excessive search results to a certain extent, thereby shortening the search time to a certain extent and improving the accuracy of the search results.
For example, if the "machine learning algorithm" is added as an integral word to the segmentation library, the "machine learning algorithm" will participate in segmentation as an integral word and search when the user searches based on the search word "machine learning algorithm video tutorial".
Step 205: and searching the database for data matched with the first segmentation result.
The first word segmentation result is obtained after the word segmentation processing is carried out on the target search word according to the word segmentation library, and the word in the word segmentation library represents the search requirement of the user, so that the first word segmentation result is reasonable word segmentation according to the search requirement of the user, and further, when the search is carried out in a pre-established database according to the first word segmentation result, data which are not needed by the user can be avoided to a certain extent, so that the search time is shortened, and the accuracy of the search result is improved.
Optionally, after the object to be stored in the database, the fourth word segmentation result, and the first mapping relationship are stored in the database, the method further includes:
updating the word segmentation library at intervals of preset time;
performing word segmentation processing on the parameter information of the object in the database according to the updated word segmentation library to obtain a fifth word segmentation result;
and establishing a second mapping relation between the fifth word segmentation result and the object in the database, and updating the first mapping relation in the database into the second mapping relation.
Therefore, in the embodiment of the present invention, the segmentation library may also be updated once at preset time intervals, where the segmentation library is updated once, that is, reasonable segmentation is obtained according to the same method for creating the segmentation library (that is, words whose occurrence times are greater than a first preset threshold value in candidate words formed by the historical search words in the historical search record and the same words in the browsing result are obtained), and then the segmentation is supplemented to the segmentation library, so that the updated segmentation library is used to perform segmentation processing on parameter information of objects stored in the database again, and a mapping relationship between the segmentation result after being re-segmented and the objects is established, that is, an index of the objects stored in the database is re-established.
The word segmentation database is updated regularly and the index of the database is reestablished according to the updated word segmentation database, so that more reasonable word segmentation can be performed according to the word segmentation database when the database is searched based on the keywords subsequently, and data required by a user can be searched more quickly.
Therefore, the searching method of the embodiment of the invention analyzes the historical search records of the user in advance to obtain the target object which the user wants to search based on the historical search words, so that the word segmentation library is constructed by the words describing the target object, and then when the word segmentation processing is carried out on the subsequent target search words according to the word segmentation library, the word segmentation can be carried out according to the requirements of the user, the unreasonable word segmentation is avoided to a certain extent, so that excessive search results cannot be returned, the searching time is shortened to a certain extent, and the accuracy of the search results is improved. And the parameter information of the object to be stored in the database is subjected to word segmentation according to the word segmentation library, so that the parameter information of the object to be stored in the database is subjected to reasonable word segmentation according to the requirements of the user, and the accuracy of the information returned to the user is higher, irrelevant content is less, time consumption is better, and user experience is better when searching is performed in the database based on the search words.
Fig. 3 is a flowchart of a process of creating a thesaurus in an embodiment of the present invention. As shown in fig. 3, the process of establishing the thesaurus may include the following steps:
step 301: a plurality of historical search records of a user are obtained.
The historical search records comprise historical search words and browsing results, and the browsing results are results of browsing selected by a user from the search results of the historical search words;
step 302: and obtaining the same words in the history search words and the browsing results in each history search record, and determining the same words as the candidate words.
The historical search words in one historical search record and all the same words of the browsing result form a candidate word.
Step 303: and acquiring the occurrence times of each candidate word in the candidate words.
Step 304: and storing the candidate words with the occurrence times larger than a first preset threshold value to form the word segmentation library.
It should be noted that, in the embodiment of the present invention, one history search record includes a group of history search terms and one browsing result, even if a user performs a search based on the same group of search terms in an actual search process, but browses a plurality of search results in sequence, one browsing result still corresponds to one history search record, for example, if the user a clicks and browses the search result X and the search result Y based on the search term "ABC", two history search records are generated here, that is, the first history search record is: the search word is ABC, and the browsing result is search result X; the second historical search record is: the search word is "ABC" and the browsing result is "search result Y".
In addition, for example, if there are one thousand history search records, in the embodiment of the present invention, when a part word library is established, the same word in the history search word and the browsing result needs to be obtained for each history search record, and a candidate word is composed of all the same words in the record, for example, a certain history search word is a "machine learning algorithm video tutorial", the browsing result is a video titled "machine learning algorithm", and the same word in the history search record and the browsing result is "machine learning" and "algorithm", and then the two words constitute a candidate word, that is, "machine learning algorithm"; then, counting each candidate word; and finally, storing the candidate words with the occurrence frequency exceeding a first preset threshold value into a word segmentation library.
For example, the occurrence frequency of the word "machine" is smaller than a first preset threshold value through statistics, and the occurrence frequency of the word "machine learning algorithm" is larger than the first preset threshold value, the word "machine" is not stored in the word segmentation library, and the word "machine learning algorithm" is stored in the word segmentation library.
Therefore, in the embodiment of the invention, the words in the pre-established word segmentation library are the words with the occurrence frequency larger than the first preset threshold value in the candidate words consisting of the historical search words in the historical search records and the same words in the browsing results, so that the words in the word segmentation library are the words meeting the search requirements of most users. Therefore, the search words are segmented according to the segmentation library, the search requirements of most users can be met, and the search experience of the users is further improved.
Optionally, before obtaining a same word in the history search record, the history search word and the browsing result, and determining the same word as the candidate word, the process of establishing the word segmentation library further includes:
and eliminating the historical search records of which the similarity between the historical search words and the browsing results is less than a second preset threshold value from the plurality of historical search records.
Among many historical search records, a user who may be less may click the search result to browse after searching based on some historical search terms, and the search result is not consistent with the previous search expectation, so that the relevance between the historical search terms and the browsing result in such historical search records is small, and the creation of a word segmentation library is not facilitated. Therefore, the historical search records can be eliminated, so that the subsequent process of creating the word segmentation library is simplified, and the influence on the accuracy of the words in the word segmentation caused by the historical search records with small relevance between the historical search words and the browsing results is avoided.
Optionally, the removing, from the plurality of historical search records, the historical search record whose similarity between the historical search terms and the browsing result is smaller than a second preset threshold includes:
acquiring a first characteristic vector of a history search word in each history search record and a second characteristic vector of a browsing result;
calculating the similarity between the first feature vector and a second feature vector belonging to the same historical search record;
and eliminating the historical search records corresponding to the similarity smaller than the second preset threshold from the plurality of historical search records.
Wherein, the similarity between the feature vectors can be represented by cosine similarity. It will be appreciated that the similarity between feature vectors may be calculated using other algorithms.
Optionally, the obtaining the same word in the history search word and the browsing result in each history search record, and determining the same word as the candidate word includes:
performing word segmentation processing on the historical search words in each historical search record by adopting a preset word segmentation algorithm to obtain a second word segmentation result;
performing word segmentation processing on the browsing result in each historical search record by adopting the preset word segmentation algorithm to obtain a third word segmentation result;
and forming a candidate word by using the second word segmentation result and the same word in the third word segmentation result belonging to the same historical search record.
Therefore, in the embodiment of the invention, when segmenting the historical search words and the browsing results, the same segmentation algorithm is adopted, and then the words which are commonly appeared in the segmentation of the historical search words and the segmentation of the browsing results in the same historical search record form a candidate word. The historical search words and the browsing results are subjected to word segmentation processing by adopting the same word segmentation algorithm, so that the same words in the historical search words and the browsing results can be more accurately acquired.
Optionally, before obtaining the same word in the history search word and the browsing result in each history search record and determining the same word as the candidate word, the method further includes:
and performing data cleaning on the historical search words.
Alternatively, the data cleansing process may employ the python language. Where Python is a high level programming language for interpreted, object oriented, dynamic data types.
The data cleaning is a procedure for finding and correcting recognizable errors in the data file, and comprises the steps of checking data consistency, processing invalid values and missing values and the like. In the embodiment of the invention, the historical search words can be regularized by cleaning the data of the historical search words, for example, a certain historical search word is a machine learning algorithm-video course which is a machine learning algorithm video course after being cleaned, so that the accuracy of the words in the word segmentation library created subsequently is further improved.
To sum up, for the video search, a flowchart of a specific implementation manner of the embodiment of the present invention can be shown in fig. 4:
firstly, when a user searches a video title through a video background, the searching behaviors of the user can be collected through real-time streaming, and the videos are clicked after searching, and then the searching behaviors of the user are stored in a searching behavior database after data cleaning.
Secondly, analyzing the cleaned user behavior data in a machine learning mode, so that the fact that some words searched by the user are for finding some videos can be known, the same words in the titles of the searched words and the videos clicked and watched by the user are obtained, and how to divide the words of the searched words is more reasonable.
That is, by analyzing a large number of user search behaviors and data of videos viewed by clicking in the corresponding search behaviors, it can be roughly understood that a user searches for some words for some videos. Such as: many users search for a "machine learning algorithm video tutorial", which is a word segmentation in the prior art: the method comprises the steps of machine learning, algorithm, video and course segmentation, wherein the word segmentation mode can search out a lot of irrelevant contents, and the video related to the course can be unrelated to the machine learning algorithm and can be searched out, and a user does not click. By analyzing the click behavior of the user search, it can be known that: the video related to the machine learning algorithm is actually searched by the user. Therefore, the correct segmentation of the user's search results should be: the machine learning algorithm, the video and the course are reasonable. Therefore, the search behavior of the user is analyzed in a machine learning mode, and the obtained search word is the same as the word in the video title clicked and watched by the user, namely, the word is a reasonable word segmentation.
And thirdly, regularly maintaining the words obtained through machine learning to a Chinese word stock of the ik word segmentation device, and reconstructing an Elasticissearch index of the video stock.
Therefore, when a new video is put in a warehouse and a user searches, word segmentation can be carried out according to a new word segmentation rule. In the search result, the accuracy of the information returned to the user is higher, irrelevant content is less, time consumption is better, and user experience is better.
Therefore, by adopting the searching method provided by the embodiment of the invention, a new word segmentation library can be constructed by collecting the searching behaviors of the user, analyzing the searching behaviors of the user and based on a machine learning mode. And using the word segmentation library in the elastic search to perform word segmentation and storage and search behaviors of the video title. In this way, the user's needs and search behavior can be better understood, and thus videos really needed by the user and related videos are returned, rather than much video information completely irrelevant to the user's search behavior as in the prior art.
The video searching is carried out according to the word segmentation mode in the prior art, millions or even tens of millions of video searching results can be returned once, and the time consumption is 2-3 seconds at most. However, if the search method of the embodiment of the present invention is used to search videos, the number of search results returned to videos is generally dozens, hundreds, or thousands, and the time consumption is at most 1 second.
Fig. 5 is a block diagram of a search apparatus according to an embodiment of the present invention, where the search apparatus 40 includes:
a search word obtaining module 401, configured to obtain a target search word;
a first word segmentation module 402, configured to perform word segmentation processing on the target search word according to a pre-established word segmentation library, so as to obtain a first word segmentation result;
a searching module 403, configured to search a pre-established database for data matching the first segmentation result;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
Therefore, the search device of the embodiment of the invention extracts the words for describing the search targets of the historical search words from the historical search records of the user by collecting the historical search records of the user and based on the machine learning mode, and constructs the word segmentation library by the words, so that when searching is needed subsequently based on the search words, the word segmentation can be carried out on the target search words according to the word segmentation library, and then the required data is searched in the pre-established database according to the word segmentation result. Therefore, the embodiment of the invention analyzes the historical search records of the user in advance to obtain the target object which the user wants to search based on the historical search words, so that the word segmentation library is constructed by the words describing the target object, and when the word segmentation processing is carried out on the subsequent target search words according to the word segmentation library, the word segmentation can be carried out according to the requirements of the user, the unreasonable word segmentation is avoided to a certain extent, so that excessive search results cannot be returned, the search time is shortened to a certain extent, and the accuracy of the search results is improved.
Fig. 6 is a block diagram of another search apparatus according to an embodiment of the present invention, where the search apparatus 50 includes:
a search term obtaining module 501, configured to obtain a target search term;
a first word segmentation module 502, configured to perform word segmentation processing on the target search word according to a pre-established word segmentation library, so as to obtain a first word segmentation result;
a searching module 503, configured to search a pre-established database for data matching the first segmentation result;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
Optionally, the apparatus further comprises: a segmentation library establishing module 504;
the word segmentation library establishing module 504 includes:
the history record obtaining sub-module 5041 is configured to obtain multiple history search records of a user, where the history search records include history search terms and browsing results, and the browsing results are results obtained by selecting from search results of the history search terms and browsing by the user;
the candidate word determining submodule 5044 is configured to obtain a word in each historical search record, which is the same as a word in the browsing result, and determine the word as a candidate word; the historical search words in one historical search record and all the same words of the browsing result form a candidate word;
the frequency counting submodule 5045 is used for acquiring the occurrence frequency of each candidate word in the candidate words;
the word extraction sub-module 5046 is configured to store the candidate words whose occurrence times are greater than a first preset threshold, so as to form the word segmentation library.
Optionally, the word segmentation library creating module 504 further includes:
and the removing sub-module 5043 is configured to remove, from the plurality of historical search records, a historical search record whose similarity between the historical search term and the browsing result is less than a second preset threshold.
Optionally, the culling sub-module 5043 includes:
a vector obtaining unit 50431, configured to obtain a first feature vector of a history search term in each history search record and a second feature vector of a browsing result;
a similarity calculation unit 50432, configured to calculate a similarity between the first feature vector and a second feature vector that belongs to the same history search record as the first feature vector;
a removing unit 50433, configured to remove history search records corresponding to the similarity smaller than the second preset threshold from the multiple history search records.
Optionally, the candidate word determination sub-module 5044 includes:
the first segmentation unit 50441 is configured to perform a segmentation process on each historical search term in each historical search record by using a preset segmentation algorithm, and obtain a second segmentation result;
a second word segmentation unit 50442, configured to perform word segmentation processing on the browsing result in each historical search record by using the preset word segmentation algorithm, so as to obtain a third word segmentation result;
and the candidate word determining unit 50443 is configured to combine the second segmentation result and the same word in the third segmentation result belonging to the same historical search record into a candidate word.
Optionally, the word segmentation library creating module 504 further includes:
and the data cleaning sub-module 5042 is used for performing data cleaning on the historical search terms.
Optionally, the apparatus further comprises:
a second word segmentation module 505, configured to perform word segmentation processing on the parameter information of the object to be stored in the database according to the word segmentation library, so as to obtain a fourth word segmentation result;
a storage module 506, configured to establish a first mapping relationship between the fourth word segmentation result and the object to be stored in the database, and store the object to be stored in the database, the fourth word segmentation result, and the first mapping relationship in the database.
Optionally, the apparatus further comprises:
a word segmentation bank updating module 507, configured to update the word segmentation bank at preset time intervals;
a third word segmentation module 508, configured to perform word segmentation processing on the parameter information of the object in the database according to the updated word segmentation library, so as to obtain a fifth word segmentation result;
a mapping relationship updating module 509, configured to establish a second mapping relationship between the fifth word segmentation result and the object in the database, and update the first mapping relationship in the database to the second mapping relationship.
Optionally, the object to be stored in the database is a video, the parameter information is a video title, the word segmentation library includes words for describing the video title, and the target search word is a word for describing the video title; the search module 503 is specifically configured to:
searching a word matched with the first word segmentation result in a pre-established database;
and determining the video with the mapping relation of the words matched with the first word segmentation result as the data matched with the first word segmentation result.
As can be seen from the above description, the search apparatus according to the embodiment of the present invention analyzes the historical search records of the user in advance, obtains the target object that the user desires to search based on the historical search word, and thus constructs the word segmentation library from the words describing the target object, and when performing word segmentation processing on the subsequent target search word according to the word segmentation library, word segmentation can be performed according to the needs of the user, so that unreasonable word segmentation is avoided to a certain extent, and thus excessive search results are not returned, so that the search time is shortened to a certain extent, and the accuracy of the search results is improved. And the parameter information of the object to be stored in the database is subjected to word segmentation according to the word segmentation library, so that the parameter information of the object to be stored in the database is subjected to reasonable word segmentation according to the requirements of the user, and the accuracy of the information returned to the user is higher, irrelevant content is less, time consumption is better, and user experience is better when searching is performed in the database based on the search words.
An embodiment of the present invention further provides an electronic device, as shown in fig. 7, including a processor 61, a communication interface 62, a memory 63, and a communication bus 64, where the processor 61, the communication interface 62, and the memory 63 complete mutual communication through the communication bus 64;
a memory 63 for storing a computer program;
the processor 61 is configured to implement the following steps when executing the program stored in the memory 63:
acquiring a target search word;
performing word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result;
searching data matched with the first segmentation result in a pre-established database;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
Optionally, the process of establishing the thesaurus includes:
acquiring a plurality of historical search records of a user, wherein the historical search records comprise historical search words and browsing results, and the browsing results are results selected by the user from the search results of the historical search words for browsing;
obtaining the same words in the history search words and the browsing results in each history search record, and determining the same words as candidate words; the historical search words in one historical search record and all the same words of the browsing result form a candidate word;
acquiring the occurrence frequency of each candidate word in the candidate words;
and storing the candidate words with the occurrence times larger than a first preset threshold value to form the word segmentation library.
Optionally, before obtaining the same word in the history search record as the word in the browsing result in each history search record and determining the word as a candidate word, the processor 61 is further configured to:
and eliminating the historical search records of which the similarity between the historical search words and the browsing results is less than a second preset threshold value from the plurality of historical search records.
Optionally, when history search records with a similarity between a history search word and a browsing result smaller than a second preset threshold are removed from the plurality of history search records, the processor 61 is specifically configured to:
acquiring a first characteristic vector of a history search word in each history search record and a second characteristic vector of a browsing result;
calculating the similarity between the first feature vector and a second feature vector belonging to the same historical search record;
and eliminating the historical search records corresponding to the similarity smaller than the second preset threshold from the plurality of historical search records.
Optionally, when obtaining the same word in the history search word and the browsing result in each history search record and determining the word as a candidate word, the processor 61 is specifically configured to:
performing word segmentation processing on the historical search words in each historical search record by adopting a preset word segmentation algorithm to obtain a second word segmentation result;
performing word segmentation processing on the browsing result in each historical search record by adopting the preset word segmentation algorithm to obtain a third word segmentation result;
and forming a candidate word by using the second word segmentation result and the same word in the third word segmentation result belonging to the same historical search record.
Optionally, before obtaining the same word in the history search record as the word in the browsing result in each history search record and determining the word as a candidate word, the processor 61 is further configured to:
and performing data cleaning on the historical search words.
Optionally, the processor 61 is further configured to:
according to the word segmentation library, carrying out word segmentation processing on the parameter information of the object to be stored in the database to obtain a fourth word segmentation result;
and establishing a first mapping relation between the fourth word segmentation result and the object to be stored in the database, and storing the object to be stored in the database, the fourth word segmentation result and the first mapping relation in the database.
Optionally, after the object to be stored in the database, the fourth word segmentation result, and the first mapping relationship are stored in the database, the processor 61 is further configured to: :
updating the word segmentation library at intervals of preset time;
performing word segmentation processing on the parameter information of the object in the database according to the updated word segmentation library to obtain a fifth word segmentation result;
and establishing a second mapping relation between the fifth word segmentation result and the object in the database, and updating the first mapping relation in the database into the second mapping relation.
Optionally, the object to be stored in the database is a video, the parameter information is a video title, the word segmentation library includes words for describing the video title, and the target search word is a word for describing the video title; when searching for data matching the first segmentation result in a pre-established database, the processor 61 is specifically configured to:
searching a word matched with the first word segmentation result in a pre-established database;
and determining the video with the mapping relation of the words matched with the first word segmentation result as the data matched with the first word segmentation result.
The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
In yet another embodiment of the present invention, a computer-readable storage medium is further provided, which has instructions stored therein, which when run on a computer, cause the computer to perform the search method described in any of the above embodiments.
In yet another embodiment, the present invention further provides a computer program product containing instructions which, when run on a computer, cause the computer to perform the search method described in any of the above embodiments.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (12)
1. A method of searching, the method comprising:
acquiring a target search word;
performing word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result;
searching data matched with the first segmentation result in a pre-established database;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
2. The search method according to claim 1, wherein the process of establishing the thesaurus comprises:
acquiring a plurality of historical search records of a user, wherein the historical search records comprise historical search words and browsing results, and the browsing results are results selected by the user from the search results of the historical search words for browsing;
obtaining the same words in the history search words and the browsing results in each history search record, and determining the same words as candidate words; the historical search words in one historical search record and all the same words of the browsing result form a candidate word;
acquiring the occurrence frequency of each candidate word in the candidate words;
and storing the candidate words with the occurrence times larger than a first preset threshold value to form the word segmentation library.
3. The searching method according to claim 2, wherein before obtaining the same word in each historical search record as the word in the browsing result and determining the word as a candidate word, further comprising:
and eliminating the historical search records of which the similarity between the historical search words and the browsing results is less than a second preset threshold value from the plurality of historical search records.
4. The searching method according to claim 3, wherein the removing, from the plurality of historical search records, the historical search record whose similarity between the historical search word and the browsing result is less than a second preset threshold value comprises:
acquiring a first characteristic vector of a history search word in each history search record and a second characteristic vector of a browsing result;
calculating the similarity between the first feature vector and a second feature vector belonging to the same historical search record;
and eliminating the historical search records corresponding to the similarity smaller than the second preset threshold from the plurality of historical search records.
5. The searching method according to claim 2, wherein the obtaining of the same word in each historical search record as the word in the browsing result and determining the same word as the word in the browsing result as a candidate word comprises:
performing word segmentation processing on the historical search words in each historical search record by adopting a preset word segmentation algorithm to obtain a second word segmentation result;
performing word segmentation processing on the browsing result in each historical search record by adopting the preset word segmentation algorithm to obtain a third word segmentation result;
and forming a candidate word by using the second word segmentation result and the same word in the third word segmentation result belonging to the same historical search record.
6. The searching method according to claim 2, wherein before obtaining the same word in each of the historical search records as the word in the browsing result and determining the word as the candidate word, further comprising:
and performing data cleaning on the historical search words.
7. The search method of claim 1, further comprising:
according to the word segmentation library, carrying out word segmentation processing on the parameter information of the object to be stored in the database to obtain a fourth word segmentation result;
and establishing a first mapping relation between the fourth word segmentation result and the object to be stored in the database, and storing the object to be stored in the database, the fourth word segmentation result and the first mapping relation in the database.
8. The searching method according to claim 7, wherein after storing the object to be stored in the database, the fourth segmentation result and the first mapping relationship in the database, further comprising:
updating the word segmentation library at intervals of preset time;
performing word segmentation processing on the parameter information of the object in the database according to the updated word segmentation library to obtain a fifth word segmentation result;
and establishing a second mapping relation between the fifth word segmentation result and the object in the database, and updating the first mapping relation in the database into the second mapping relation.
9. The searching method according to claim 7, wherein the object to be stored in the database is a video, the parameter information is a video title, the thesaurus includes words for describing the video title, and the target search word is a word for describing the video title;
searching data matched with the first segmentation result in a pre-established database, wherein the searching comprises the following steps:
searching a word matched with the first word segmentation result in a pre-established database;
and determining the video with the mapping relation of the words matched with the first word segmentation result as the data matched with the first word segmentation result.
10. A search apparatus, characterized in that the apparatus comprises:
the search word acquisition module is used for acquiring a target search word;
the first word segmentation module is used for carrying out word segmentation processing on the target search word according to a pre-established word segmentation library to obtain a first word segmentation result;
the searching module is used for searching data matched with the first segmentation result in a pre-established database;
the words in the word segmentation library are target words extracted from historical search records of the user in a machine learning mode, and the target words are used for describing search targets of the historical search words.
11. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the search method of any one of claims 1 to 9 when executing a program stored in the memory.
12. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, carries out the search method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010318720.8A CN111666448B (en) | 2020-04-21 | 2020-04-21 | Search method, search device, electronic equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010318720.8A CN111666448B (en) | 2020-04-21 | 2020-04-21 | Search method, search device, electronic equipment and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111666448A true CN111666448A (en) | 2020-09-15 |
| CN111666448B CN111666448B (en) | 2024-01-26 |
Family
ID=72382679
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010318720.8A Active CN111666448B (en) | 2020-04-21 | 2020-04-21 | Search method, search device, electronic equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111666448B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158091A (en) * | 2021-03-24 | 2021-07-23 | 北京奇艺世纪科技有限公司 | Recall method, apparatus, electronic device and storage medium |
| CN113177061A (en) * | 2021-05-25 | 2021-07-27 | 马上消费金融股份有限公司 | Searching method and device and electronic equipment |
| CN113342926A (en) * | 2021-02-22 | 2021-09-03 | 武汉烽火众智数字技术有限责任公司 | Case serial-parallel method and system based on natural language processing technology |
| CN113468886A (en) * | 2021-07-28 | 2021-10-01 | 拉扎斯网络科技（上海）有限公司 | Work order processing method and device and computer equipment |
| CN113656679A (en) * | 2021-08-27 | 2021-11-16 | 支付宝(杭州)信息技术有限公司 | User searching method and device |
| WO2023125315A1 (en) * | 2021-12-31 | 2023-07-06 | 北京百度网讯科技有限公司 | Information search method and apparatus, electronic device and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095187A (en) * | 2015-08-07 | 2015-11-25 | 广州神马移动信息科技有限公司 | Search intention identification method and device |
| 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 |
| WO2017088126A1 (en) * | 2015-11-25 | 2017-06-01 | 华为技术有限公司 | Method and device for obtaining out-of-vocabulary word |
| US20190179965A1 (en) * | 2017-12-13 | 2019-06-13 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for generating information |
-
2020
- 2020-04-21 CN CN202010318720.8A patent/CN111666448B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095187A (en) * | 2015-08-07 | 2015-11-25 | 广州神马移动信息科技有限公司 | Search intention identification method and device |
| WO2017024884A1 (en) * | 2015-08-07 | 2017-02-16 | 广州神马移动信息科技有限公司 | Search intention identification method and device |
| WO2017088126A1 (en) * | 2015-11-25 | 2017-06-01 | 华为技术有限公司 | Method and device for obtaining out-of-vocabulary word |
| 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 |
| US20190179965A1 (en) * | 2017-12-13 | 2019-06-13 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for generating information |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113342926A (en) * | 2021-02-22 | 2021-09-03 | 武汉烽火众智数字技术有限责任公司 | Case serial-parallel method and system based on natural language processing technology |
| CN113158091A (en) * | 2021-03-24 | 2021-07-23 | 北京奇艺世纪科技有限公司 | Recall method, apparatus, electronic device and storage medium |
| CN113177061A (en) * | 2021-05-25 | 2021-07-27 | 马上消费金融股份有限公司 | Searching method and device and electronic equipment |
| CN113177061B (en) * | 2021-05-25 | 2023-05-16 | 马上消费金融股份有限公司 | Searching method and device and electronic equipment |
| CN113468886A (en) * | 2021-07-28 | 2021-10-01 | 拉扎斯网络科技（上海）有限公司 | Work order processing method and device and computer equipment |
| CN113656679A (en) * | 2021-08-27 | 2021-11-16 | 支付宝(杭州)信息技术有限公司 | User searching method and device |
| CN113656679B (en) * | 2021-08-27 | 2024-03-12 | 支付宝(杭州)信息技术有限公司 | User searching method and device |
| WO2023125315A1 (en) * | 2021-12-31 | 2023-07-06 | 北京百度网讯科技有限公司 | Information search method and apparatus, electronic device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111666448B (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111666448B (en) | Search method, search device, electronic equipment and computer readable storage medium | |
| CN110162695B (en) | Information pushing method and equipment | |
| CN107220365B (en) | Accurate recommendation system and method based on collaborative filtering and association rule parallel processing | |
| RU2501078C2 (en) | Ranking search results using edit distance and document information | |
| US7502780B2 (en) | Information storage and retrieval | |
| EP2438539B1 (en) | Co-selected image classification | |
| CN110795657B (en) | Article pushing and model training method and device, storage medium and computer equipment | |
| US20060095852A1 (en) | Information storage and retrieval | |
| CN103279513A (en) | Method for generating content label and method and device for providing multi-media content information | |
| WO2008106668A1 (en) | User query mining for advertising matching | |
| CN111666450A (en) | Video recall method and device, electronic equipment and computer-readable storage medium | |
| GB2395807A (en) | Information retrieval | |
| Sisodia et al. | Fast prediction of web user browsing behaviours using most interesting patterns | |
| CN111400546A (en) | Video recall method and video recommendation method and device | |
| CN113065018A (en) | Audio and video index library creating and retrieving method and device and electronic equipment | |
| CN110795613A (en) | Commodity searching method, device and system and electronic equipment | |
| CN113220974A (en) | Click rate prediction model training and search recall method, device, equipment and medium | |
| CN110765348B (en) | Hot word recommendation method and device, electronic equipment and storage medium | |
| CN110851708B (en) | Negative sample extraction method, device, computer equipment and storage medium | |
| CN111212304A (en) | Video recommendation method and server | |
| CN112163415A (en) | User intention identification method and device for feedback content and electronic equipment | |
| CN115794898B (en) | Financial information recommendation method and device, electronic equipment and storage medium | |
| CN111314790A (en) | Video playing record sequencing method and device and electronic equipment | |
| CN110727867A (en) | Semantic entity recommendation method based on fuzzy mechanism | |
| CN116225848A (en) | Log monitoring method, device, equipment and 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 |