CN112364126A

CN112364126A - Keyword prompting method and device, computer equipment and storage medium

Info

Publication number: CN112364126A
Application number: CN202011134797.6A
Authority: CN
Inventors: 张旭东
Original assignee: Guangzhou Baiguoyuan Information Technology Co Ltd
Current assignee: Guangzhou Baiguoyuan Information Technology Co Ltd
Priority date: 2020-10-21
Filing date: 2020-10-21
Publication date: 2021-02-12

Abstract

The embodiment of the invention provides a keyword prompting method, a keyword prompting device, computer equipment and a storage medium, wherein the method comprises the following steps: determining an ordered tree, wherein nodes in the ordered tree represent characters and are associated with keywords for searching, the keywords take characters from root nodes to nodes as prefixes, receiving characters which are sent by a client and are currently used for searching as original character strings, searching characters which are the same as and/or similar to the original character strings in the ordered tree to form target character strings, inquiring keywords which take the target character strings as prefixes in the ordered tree, taking the keywords as candidate keywords, extracting part of the candidate keywords as target keywords, and sending the target keywords to the client for displaying.

Description

Keyword prompting method and device, computer equipment and storage medium

Technical Field

The embodiment of the invention relates to the technical field of search, in particular to a keyword prompting method and device, computer equipment and a storage medium.

Background

A user usually inputs keywords to search for related information in scenes such as short videos, news, shopping and the like, and at this time, a search engine usually provides a pull-down prompt (Query suggestion) of the keywords, that is, a candidate list of the keywords is automatically provided for the user to select according to the keywords currently input by the user, so that the user is helped to clarify the search intention, the input operation of the user is simplified, and the search time is saved.

In order to meet the requirement of real-time performance, the pull-down prompt of the keywords is usually to pre-establish a list of hot keywords under each prefix, and when a character requested by a user arrives, the hot keywords corresponding to the input character of the user are directly returned.

However, in order to correct the Keyword input by the user, the form that the user may input errors is often stored in advance, which consumes a huge storage space, and meanwhile, some Long Tail keywords (Long Tail keywords) are not set in the form of errors, so that the error correction cannot be performed.

Disclosure of Invention

The embodiment of the invention provides a keyword prompting method and device, computer equipment and a storage medium, and aims to solve the problems of reducing storage space occupied by keyword error correction and covering long-tail keywords under the condition that the pull-down prompt of the keywords meets the real-time property.

In a first aspect, an embodiment of the present invention provides a keyword prompting method, including:

determining an ordered tree, wherein nodes in the ordered tree represent characters and are associated with keywords for searching, and the keywords use the characters from a root node to the nodes as prefixes;

receiving characters which are sent by a client and are currently used for searching, and using the characters as original character strings;

searching the ordered tree for characters which are the same as and/or similar to the original character string to form a target character string;

querying keywords with the target character strings as prefixes in the ordered tree as candidate keywords;

extracting part of the candidate keywords as target keywords;

and sending the target keywords to the client side for display.

In a second aspect, an embodiment of the present invention further provides a keyword prompting apparatus, including:

the ordered tree determining module is used for determining an ordered tree, nodes in the ordered tree represent characters and are associated with keywords for searching, and the keywords take the characters from a root node to the nodes as prefixes;

the system comprises an original character string receiving module, a searching module and a searching module, wherein the original character string receiving module is used for receiving characters which are sent by a client and are currently used for searching and used as an original character string;

the target character string searching module is used for searching characters which are the same as and/or similar to the original character string in the ordered tree to form a target character string;

a candidate keyword query module, configured to query, in the ordered tree, a keyword using the target character string as a prefix, as a candidate keyword;

the target keyword extraction module is used for extracting part of the candidate keywords as target keywords;

and the target keyword pushing module is used for sending the target keywords to the client side for displaying.

In a third aspect, an embodiment of the present invention further provides a computer device, where the computer device includes:

one or more processors;

a memory for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement the method of hinting for keywords as described in the first aspect.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when executed by a processor, the computer program implements the method for prompting the keyword according to the first aspect.

In this embodiment, an ordered tree is determined, nodes in the ordered tree represent characters and keywords associated with search, the keywords use characters from a root node to the nodes as prefixes, receive characters currently used for search and sent by a client as original character strings, search characters identical to and/or similar to the original character strings in the ordered tree to form target character strings, query keywords with the target character strings as prefixes in the ordered tree, use the keywords as candidate keywords, extract part of the candidate keywords as target keywords, send the target keywords to the client for display, on one hand, use the ordered tree to carry the characters of the keywords, can quickly match the original character strings requested by the user, the ordered tree provides quick search service, can quickly locate the target character strings, can ensure the real-time performance of pushing, and the average response speed can be within 5ms, the method can show the target keyword before the user inputs the next character, on the other hand, the target character string similar to the original character string is searched in the ordered tree and pushed to the client, error correction can be carried out on the error input of the user, the target character string is formed by characters of the keyword and is recorded in the ordered tree, the error input form of the user does not need to be stored in advance, the occupation of storage space is reduced, and meanwhile, some long-tail keywords can be covered to carry out error correction on the long-tail keywords.

Drawings

Fig. 1 is a flowchart of a keyword prompting method according to an embodiment of the present invention;

FIG. 2 is an exemplary diagram of an ordered tree according to an embodiment of the present invention;

fig. 3 is a flowchart of a keyword prompting method according to a second embodiment of the present invention;

FIG. 4 is a diagram illustrating an example of displaying keywords according to a second embodiment of the present invention;

fig. 5 is a schematic structural diagram of a keyword prompting device according to a third embodiment of the present invention;

fig. 6 is a schematic structural diagram of a computer device according to a fourth embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.

Example one

Fig. 1 is a flowchart of a keyword prompting method according to an embodiment of the present invention, where the embodiment is applicable to a case of generating an ordered tree for a keyword, the method may be executed by a keyword prompting device, the keyword prompting device may be implemented by software and/or hardware, and may be configured in a computer device, such as a server, a workstation, a personal computer, and the like, and specifically includes the following steps:

step 101, obtaining a keyword for searching.

Currently, a user may access a search engine from various electronic devices, which may specifically include mobile devices, such as a mobile phone, a PDA (Personal Digital Assistant), a laptop computer, a palmtop computer, an intelligent wearable device (such as an intelligent watch and an intelligent glasses), and the like, and may also include fixed devices, such as a Personal computer, an intelligent television, and the like.

These electronic devices may support operating systems including Android (Android), IOS, windows, etc., and may typically run a browser to access web pages over a network or an application with built-in browsing components (e.g., WebView), such as short video applications, live applications, shopping applications, blogging applications, and so forth.

In one case, a user may open a web page in which a search engine is located in a browser or an application with a built-in browsing component, where the web page would typically include a search box in which the user may enter a keyword to request the search engine to search for business information (e.g., web page, short video, anchor, live program, blog, etc.) related to the keyword.

In another case, a browser or an application with a built-in browsing component is installed with a search plug-in (plug-ins) which can add a search function in the browser or the application with the built-in browsing component by interacting with a search engine, and the search plug-in provides a search box in which a user can input a keyword to request the search engine to search for service information (such as a web page, a short video, a main broadcast, a live program, a blog, etc.) related to the keyword.

The search engine provides a quick and highly relevant information search service for the user by relying on technologies such as a web crawler technology, a retrieval sorting technology, a web page processing technology, a big data processing technology, a natural language processing technology and the like, namely relevant service information is searched from the network and fed back to the user.

Part of search engines are comprehensive search and transverse search, the search information quantity is large, the search accuracy and the correlation quality are difficult to be considered, and the requirements of pursuing accurate personalized and specialized search are difficult to meet.

Therefore, part of the Search engines are Vertical Search (Vertical Search), that is, specialized Search engines in a certain business field, for example, Search for short videos, Search for live broadcasts, Search for commodity information, and the like, the Vertical Search is a subdivision and extension of the Search engines, and is a process of integrating specific information in a web library at one time, and extracting relevant business information in a targeted sub-field and returning the extracted business information to the user.

In this embodiment, the related information generated when the user requests the search engine to perform the search may be recorded in the log file, so as to extract the keyword from the log files of different users.

It should be noted that the keywords may include words, phrases, symbols, audio data, and other forms of various languages, which is not limited in this embodiment.

Step 102, nodes representing characters are determined in the ordered tree.

A keyword contains one or more characters that are the smallest constituent units of the keyword, e.g., the letters "a" - "Z", "a" - "Z" for english, the characters "0" - "9" for arabic, the characters may be single chinese characters for chinese, etc.

For applying the embodiment, an ordered tree may be generated for the keyword, where the ordered tree has a plurality of nodes, and an edge between the nodes is a directed edge, so that each character in the keyword is carried by the node in the ordered tree according to the directed edge, that is, one node of the ordered tree is used to represent one character of the keyword, and the directed edge of the node represents the order of the characters.

In general, in an ordered tree, there is and only one node as a root node, the root node does not represent a character, no directed edge points to the node, any node other than the root node can be used to represent a character, and there is and only one directed edge points to the node, i.e., there is a directed path from the root node to any non-root node.

In order to make the skilled person better understand the embodiments of the present invention, in this specification, a Tire tree (also called a prefix tree, a dictionary tree) is described as an example of an ordered tree.

The Trie is an index structure, which is actually a Deterministic Finite Automaton (DFA), in which each node corresponds to a DFA state, and each (directed) marked edge pointing from a parent node to a child node corresponds to a DFA transition.

In the specific implementation, the characters in the keywords are sequentially extracted according to the sequence, and the nodes representing the characters are searched from the root node according to the sequence of the directed edges.

If the characters are found, continuing to obtain the next character, if the characters are not found, inserting a node representing the characters below the current node, generating a directed edge pointing to the node from the current node, and continuing to obtain the next character until all the characters of the keywords are traversed.

And 103, associating the keywords with the nodes.

In this embodiment, each node is located in a stack structure, and if a character of a keyword matches a certain node (i.e., the content is the same and the sequence is the same), the keyword may be recorded in the stack structure of the node, so that the nodes in the ordered tree may be associated with the keyword for search in addition to representing the character, and the keyword is prefixed by the character from the root node to the node.

For example, as shown in fig. 2, assuming that a certain keyword is "news", it is possible to search for a node indicating "n" in the direction indicated by the root node (R) from the root node (R), associate "news" with the stack structure of the nodes indicating "n" when the node indicating "n" is found, search for a node indicating "e" in the direction indicated by the node indicating "n", associate "news" with the stack structure of the nodes indicating "e" when the node indicating "e" is found, search for a node indicating "w" in the direction indicated by the node indicating "e" when the node indicating "e" is found, associate "news" with the stack structure of the nodes indicating "w", search for a node indicating "s" in the direction indicated by the node indicating "w", generate a node indicating "s" when the node indicating "s" is not found, "news" is associated into the heap structure of the node representing "s" by pointing along the node representing "w" to the node representing "s".

In one embodiment of the present invention, step 103 may comprise the steps of:

and step 1031, inquiring a channel during searching the recall keyword as a recall channel.

The channel to which the query client requests to search for a certain keyword is referred to as a recall channel for the convenience of distinction.

Wherein a channel is an abstraction of one or more features in the process of recalling a keyword.

In one example, the channel may include the geographic location where the client is located, and in this example, the geographic location from which the queryable keyword originated at the time of the search is used as the recall channel.

For example, if the application program serves users in different languages, the geographic location may be divided by using the language as a dimension, so that users in the same geographic location use one language, and since different search intentions may possibly exist for keywords in different languages, the geographic location is used as a recall channel, so that different search intentions may be distinguished, and thus the accuracy of subsequently pushing the keywords is improved.

Of course, the above-mentioned recall channel is only an example, and when implementing the embodiment of the present invention, other recall channels may be set according to actual situations, for example, a version of the client, a system platform where the client is located, an ISP (Internet Service Provider) to which the client accesses, and the like, which are not limited in this embodiment of the present invention. In addition, besides the above judgment processing method, a person skilled in the art may also adopt other judgment processing methods according to actual needs, and the embodiment of the present invention is not limited thereto.

And step 1032, taking the recall channel as a statistic dimension, and counting the user behaviors fed back by aiming at the keywords to obtain a search index.

If the search engine searches the business information related to the keyword, the business information is sent to the client as a search result to be displayed, and a user can trigger corresponding user behaviors aiming at the business information at the client.

In this embodiment, for keywords having the same recall channel, one or more user behaviors of the keyword under the recall channel and fed back by the client may be counted, and a value of the keyword or a value after fusion is set as a search index.

Generally, the user behavior counted in this embodiment is a behavior expressing a forward direction, for example, the geographic location may be used as a dimension of the statistics, the number of the service information corresponding to the click keyword (i.e., the click amount) may be counted as a search index, the operation of counting the click amount is simple, the calculation resource may be saved, the click amount may more comprehensively express the quality of the service information searched by the keyword in the search engine, and if a subsequent user uses the pushed keyword to perform a search, the high-quality service information may be quickly located.

Of course, the search indexes are only examples, and when the embodiment of the present invention is implemented, other search indexes may be set according to actual situations, for example, browsing time, frequency of searching for the keyword, conversion rate, and the like, which is not limited in the embodiment of the present invention. In addition, besides the above search indexes, those skilled in the art may also adopt other search indexes according to actual needs, and the embodiment of the present invention is not limited to this.

And step 1033, inserting the keywords into a recall channel under the node.

As shown in FIG. 2, in the ordered tree, the heap structure of the non-root node includes recall channels (e.g., channel A, channel B, etc.), and the keywords are inserted under the corresponding recall channels of the current node.

And 1034, applying m keywords with the highest search indexes under the recall channel.

As shown in fig. 2, the keywords of each recall channel in the node are sorted according to the search index, and m (m is a positive integer) keywords with the highest search index are validated, so that the m (m is a positive integer) keywords with the highest search index are applied to the process of pushing to the client, and other keywords are retained in the stack structure of the node, but are not deleted, but are not applied to the process of pushing to the client.

In this embodiment, a recall channel is used as a dimension of statistics, user behaviors fed back by aiming at keywords are counted, search indexes are obtained, the keywords are inserted into the recall channel under a node, and m keywords with the highest search indexes are applied to the recall channel, so that the search quality of the keywords under the recall channel can be ensured, the probability of the keywords being selected by a user when the keywords are pushed is improved, and the operation cost of searching by inputting the keywords by the user is reduced.

Example two

Fig. 3 is a flowchart of a keyword prompting method according to a second embodiment of the present invention, where this embodiment is applicable to a case where an ordered tree fuzzy search keyword is applied, and specifically includes the following steps:

and 301, determining the ordered tree.

In this embodiment, an ordered tree, such as a Trie tree, may be generated offline, and loaded when keywords are pushed to a client online.

The nodes in the ordered tree represent characters and keywords used for searching in a related mode, and the keywords use the characters from the root node to the nodes as prefixes.

Step 302, receiving characters which are sent by the client and currently used for searching as an original character string.

The method comprises the steps that a search bar is provided in a page of a client side such as a browser and an application program with a built-in browsing component, a user moves a cursor into the search bar through mouse, touch and other operations, an input method can be called to input characters in the search bar, and when the user determines to search, the client side sends the characters to a search engine as keywords.

In the process of inputting characters by a user, the client detects the characters in the search bar in real time, if the characters are changed, such as adding characters, deleting characters, modifying characters and the like, the characters in the current search bar can be uploaded, and the characters can be called as original character strings for convenience of distinguishing.

For example, as shown in fig. 4, if the user sequentially inputs "new" in the search bar, the client may sequentially upload "n", "ne", and "new".

Step 303, searching the ordered tree for the same and/or similar characters as the original character string to form the target character string.

To support fuzzy searching, the nodes may be traversed in the ordered tree in order of directed edges, searching for characters that are the same and/or similar to the original string, and these searched characters may constitute the target string.

Traversal starts from the root node, from head to tail, determines the next state from each character of the original string, and the edges marked with the same character are selected for movement. Each such movement consumes one character from the original string and walks to the next node of the ordered tree until each character of the original string is traversed.

If a node representing a character in the original string does not exist, adjacent branches may be traversed to find similar characters.

In one embodiment of the present invention, step 303 may comprise the steps of:

step 3031, determining the next node pointed by the current node in the ordered tree.

Step 3032, matching the next node with the original character string.

In this embodiment, a finite state automaton, such as LA (Levenshtein Automata, LA, edit distance Automata), may be created, and two parameters are maintained in advance, one of which is a state set and the other is a state transition matrix.

Wherein, the state set comprises a matchable state and a non-matching state, the matchable state can represent that the ordered tree continues to walk, and the non-matching state can represent that the ordered tree stops to walk.

The state transition matrix describes that in one state a jump to another state can be made by a certain character.

For the ordered tree, all possible state sets, and all possible jump cases for each state, can be traversed in advance and stored in the state transition matrix.

When traversing the ordered tree aiming at the original character string, inputting each node of the ordered tree into a finite state automaton (such as LA) to jump to the next node, forming characters between the root node and the next node into a candidate character string, and calculating the edit distance between the original character string and the candidate character string.

The edit distance may refer to the number of times the minimum change (e.g., insertion, deletion, replacement, etc.) is required from the candidate string to the original string.

For example, if the candidate character string is "abcd" and the original character string is "acdf", the "b" is deleted from "abcd" to obtain "acd", and "f" is inserted into "acd" to obtain "acdf", so that the edit distance between "abcd" and "acdf" is 2.

Generally, to improve the matching accuracy, the edit distance between the original character string currently being matched and the candidate character string in the same digit number can be calculated, for example, if the original character string is "acdf", the "acd" is currently being matched, and the candidate character string is "abc", the edit distance between "acd" and "abc" can be calculated.

And if the editing distance is smaller than or equal to the preset threshold value, determining that the next node is successfully matched with the original character string.

And if the editing distance is larger than a preset threshold value, determining that the matching between the next node and the original character string fails.

Of course, the matching manner is only an example, and when the embodiment of the present invention is implemented, other matching manners may be set according to actual situations, for example, a hamming distance between an original character string and a candidate character string is calculated, and when the hamming distance is less than or equal to a preset threshold, it is determined that a next node is successfully matched with the original character string, otherwise, it is determined that the next node is unsuccessfully matched with the original character string; or, a cosine value between the word vector of the original character string and the word vector of the candidate character string is calculated, when the cosine value is greater than or equal to a preset threshold value, it is determined that the next node is successfully matched with the original character string, otherwise, it is determined that the next node is unsuccessfully matched with the original character string, and so on. In addition, besides the above matching methods, those skilled in the art may also adopt other matching methods according to actual needs, and the embodiment of the present invention is not limited thereto.

Step 3033, if the matching is successful, jumping from the current node to the next node, and returning to execute the step 3031.

If the finite state automaton (e.g., LA) determines that the matching of the next node to the original string is successful, the state of the next node is output as a matchable state, and the search can continue.

Step 3034, if the matching fails, determining that the characters between the root node and the current node are the same as and/or similar to the original character string as the target character string.

If the finite state automaton (such as LA) determines that the matching of the next node and the original character string fails, the state of the next node is output to be a non-matching state, the search can be stopped, and the current character is output to be the target character string.

The process of walking in the ordered tree belongs to fuzzy search, so that the original character string input by the user can be corrected.

For example, assuming that the distance threshold is 1, the original character string input by the user is "nes", and the ordered tree shown in fig. 2 is walked to search out the candidate character string "new", since the edit distance between "new" and "nes" is 1, that is, the edit distance is equal to the distance threshold, the candidate character string "new" can be identified as the target character string.

And step 304, inquiring keywords with the target character string as a prefix in the ordered tree as candidate keywords.

And querying related keywords in a stack structure of the nodes successfully matched with the ordered tree, wherein the keywords take the target character string as a prefix, and the keywords can be called as candidate keywords for distinguishing conveniently.

In an embodiment of the present invention, if the keywords are distinguished by the recall channels, a character located at the last position in the target character string may be queried as a target character, and keywords in all the recall channels associated with the target character are extracted in a stack structure of the target character as candidate keywords.

The keywords of all the recall channels are extracted, so that the condition that the keywords are not found in a certain recall channel and are not returned can be avoided, and the number of the keywords is ensured.

And 305, extracting partial candidate keywords as target keywords.

Because the number of the candidate keywords is large and the area displayed by the client is limited, when the keywords are pushed to the client, part of the keywords can be recalled from the candidate keywords by using different recall strategies according to different service requirements (such as recalling high-quality keywords, recalling keywords meeting the personalized requirements of the user and the like) to serve as target keywords.

In particular implementations, recall policies include, but are not limited to:

a subscription recall (recall keywords associated with a user's subscribed listings (e.g., a game, a restaurant, etc.), a recall in the same geographic location (recall keywords in the same geographic location as the user belongs to), a recall in the same language (recall keywords in the same language as the user used), a collaborative filtering recall (recall keywords using a collaborative filtering algorithm), and a preference recall (recall keywords in the same preference as the user).

In one embodiment of the present invention, step 305 comprises the steps of:

step 3051, determining similarity between the original character string and the candidate keyword.

In this embodiment, the similarity between the original character string and the candidate keyword may be directly calculated in a hamming distance, an editing distance, and the like, or the similarity between the original character string and the candidate character string to which the candidate keyword belongs may be multiplexed to save the calculation resources, which is not limited in this embodiment.

In the specific implementation, the candidate character string to which the candidate keyword belongs is inquired, the editing distance between the candidate character string and the original character string is inquired, and the editing distance is set as the similarity between the original character string and the candidate keyword.

Step 3052, calculating a quality value for the candidate character string based on at least the similarity.

In this embodiment, a candidate character string calculation quality value indicating the quality of a candidate character string for a search may be calculated with the degree of similarity as a parameter for calculation.

Generally, the quality value is positively correlated with the similarity, i.e., the higher the similarity is, the larger the quality value is, whereas the lower the similarity is, the smaller the quality value is.

In one example, the similarity is expressed in terms of edit distance, and a quality value may be calculated for the candidate character strings based on at least the similarity, where the quality value is inversely related to the edit distance, i.e., the smaller the edit distance, the higher the similarity and the greater the quality value, and vice versa, the larger the edit distance, the lower the similarity and the smaller the quality value.

Further, if the recall channel record keyword is distinguished when the ordered tree is generated and the keyword is validated according to the search index, the recall channel and the search index (such as click rate) of the target keyword can be searched in the ordered tree.

And inquiring the channel (such as the geographic position and the like) to which the client belongs from the current request of the client as a target channel.

And comparing the recall channel with the target channel to configure channel weight for the target keyword, wherein the channel weight when the recall channel is the same as the target channel is greater than the channel weight when the recall channel is not the same as the target channel.

Therefore, when calculating the quality value, the quality value may be calculated for the candidate character string based on the search index, the similarity, and the channel weight, where the quality value is positively correlated with the search index and the channel weight, that is, the higher the search index is, the higher the channel weight is, the larger the quality value is, and the lower the search index is, the lower the channel weight is, the smaller the quality value is, in addition to maintaining the positive correlation between the quality value and the similarity.

In one example, the mass value score is calculated as follows:

score＝log(click+1.0)+alpha*log(edit_distance+1.0)+region_boost。

wherein log () is a logarithmic function, click is a search index (e.g., click amount), alpha is a penalty coefficient, and belongs to a constant, edit _ distance is a similarity (e.g., edit distance), and region _ boost is a channel weight.

It should be noted that, when the ordered tree is generated, the error correction may be performed on the keyword input by the user, or the error correction may not be performed on the keyword input by the user, that is, the keyword input by the user is maintained.

Step 3053, selecting the target keyword from the candidate keywords according to the quality value.

In the present embodiment, a partial keyword is selected from the candidate keywords as the target keyword with the quality value as a reference.

In a manner of selecting the target keyword, the candidate keywords may be sorted according to the quality value in a descending order, that is, the higher the quality value is, the earlier the ranking is, otherwise, the lower the quality value is, the later the ranking is, k (k is a positive integer and belongs to an adjustable parameter, such as 8) candidate keywords with the highest ranking are selected as the target keyword

Of course, in addition to selecting the k candidate keywords with the highest quality values, the target keywords may also be selected in other manners, for example, selecting the candidate keywords with the quality values greater than or equal to the preset quality threshold as the target keywords, filtering out the candidate keywords with the quality values lower than the preset quality threshold, and selecting the k candidate keywords with the highest quality values from the retained candidate keywords as the target keywords, which is not limited in this embodiment.

And step 306, sending the target keywords to the client side for displaying.

After determining the target keyword, the target keyword may be sent to the client, and the client displays the target keyword in the page.

In general, as shown in fig. 2, the client may display the target keyword under the search bar, so as to facilitate the user to browse the target keyword and select the desired target keyword.

If the user clicks a certain target keyword, the search bar can be emptied, and the target keyword is written into the search bar.

Further, since the keyword input by the user is a dynamic process, the input characters may change in real time, such as adding characters, deleting characters, modifying characters, and the like, during a keyword input process, the client may send a plurality of original character strings, and in order to keep the pushed target keyword matched with the currently input character of the user, the client may carry a timestamp when pushing the target keyword.

And when the client receives the target keyword, comparing the timestamps and displaying the target keyword corresponding to the latest (namely, the largest numerical value) timestamp.

It should be noted that, for simplicity of description, the method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the illustrated order of acts, as some steps may occur in other orders or concurrently in accordance with the embodiments of the present invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred and that no particular act is required to implement the invention.

EXAMPLE III

Fig. 5 is a block diagram of a keyword prompting apparatus according to a third embodiment of the present invention, which may specifically include the following modules:

an ordered tree determining module 501, configured to determine an ordered tree, where nodes in the ordered tree represent characters and are associated with keywords used for searching, and the keywords use characters from a root node to the nodes as prefixes;

an original character string receiving module 502, configured to receive a character currently used for searching and sent by a client as an original character string;

a target character string searching module 503, configured to search the ordered tree for characters that are the same as and/or similar to the original character string to form a target character string;

a candidate keyword query module 504, configured to query, in the ordered tree, a keyword with the target character string as a prefix, as a candidate keyword;

a target keyword extraction module 505, configured to extract a part of the candidate keywords as target keywords;

and a target keyword pushing module 506, configured to send the target keyword to the client for display.

In one embodiment of the present invention, the ordered tree determination module 501 comprises:

the keyword acquisition sub-module is used for acquiring a keyword for searching, wherein the keyword comprises one or more characters;

a node determination submodule for determining nodes representing the characters in the ordered tree;

and the keyword association submodule is used for associating the keyword with the node.

In an embodiment of the present invention, the keyword association sub-module includes:

a recall channel query unit, configured to query a channel used when the keyword is recalled for search, as a recall channel;

the search index counting unit is used for counting the user behaviors fed back by aiming at the keywords by taking the recall channel as a counting dimension to obtain a search index;

a keyword insertion unit for inserting the keyword into the recall channel under the node;

and the keyword application unit is used for applying the m keywords with the highest search indexes under the recall channel.

In one embodiment of the present invention, the recall channel query unit includes:

the geographic position inquiry subunit is used for inquiring the geographic position of the source of the keyword during searching and taking the geographic position as a recall channel;

the search index statistical unit includes:

and the click quantity counting subunit is used for counting the quantity of the business information corresponding to the clicked keyword by taking the geographic position as a counting dimension, and taking the counted quantity as a search index.

In one embodiment of the present invention, the target character string searching module 503 includes:

a direction determining submodule, configured to determine, in the ordered tree, a next node to which the current node points;

the node matching submodule is used for matching the next node with the original character string;

the node skip submodule is used for skipping from the current node to the next node and returning to the direction determination submodule if the matching is successful;

and the target character string output sub-module is used for determining that the characters between the root node and the current node are the same as and/or similar to the original character string to serve as the target character string if the matching fails.

In one embodiment of the present invention, the node matching sub-module includes:

a candidate character string composing unit, configured to compose characters from a root node to a next node into a candidate character string;

an edit distance calculation unit for calculating an edit distance between the original character string and the candidate character string;

a matching success determining unit, configured to determine that the next node and the original character string are successfully matched if the editing distance is smaller than or equal to a preset threshold;

and the matching failure determining unit is used for determining that the next node is failed to be matched with the original character string if the editing distance is greater than a preset threshold value.

In one embodiment of the present invention, the candidate keyword query module 504 comprises:

the target character query submodule is used for querying a character positioned at the last position in the target character string to be used as a target character;

and the recall channel extraction submodule is used for extracting the keywords in all recall channels associated with the target character to serve as candidate keywords.

In one embodiment of the present invention, the target keyword extraction module 505 comprises:

the similarity determining submodule is used for determining the similarity between the original character string and the candidate keyword;

a quality value calculation sub-module for calculating a quality value for the candidate string based at least on the similarity, the quality value being positively correlated with the similarity;

and the quality value selection sub-module is used for selecting the target keyword from the candidate keywords according to the quality value.

In one embodiment of the present invention, the similarity determination submodule includes:

the candidate character string query unit is used for querying the candidate character string to which the candidate keyword belongs;

the editing distance query unit is used for querying the editing distance between the candidate character string and the original character string;

an edit distance setting unit configured to set the edit distance as a similarity between the original character string and the candidate keyword;

accordingly, the quality value calculation sub-module is further configured to:

a quality value is calculated for the candidate string based at least on the edit distance, the quality value being inversely related to the edit distance.

In an embodiment of the present invention, the target keyword extraction module 505 further includes:

the search parameter query submodule is used for querying a recall channel and a search index of the target keyword in the ordered tree;

the target channel query submodule is used for querying a channel to which the client belongs and taking the channel as a target channel;

a channel weight configuration submodule, configured to compare the recall channel with the target channel, so as to configure a channel weight for the target keyword, where a channel weight when the recall channel is the same as the target channel is greater than a channel weight when the recall channel is not the same as the target channel;

accordingly, the quality value calculation sub-module is further configured to:

and calculating a quality value of the candidate character string based on the search index, the similarity and the channel weight, wherein the quality value is positively correlated with the search index and the channel weight.

In one embodiment of the invention, the quality value selection sub-module comprises:

the keyword sorting unit is used for sorting the candidate keywords in a descending order according to the quality values;

and the ranking selection unit is used for selecting the k candidate keywords with the highest ranking as the target keywords.

The keyword prompting device provided by the embodiment of the invention can execute the keyword prompting method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

Example four

Fig. 6 is a schematic structural diagram of a computer device according to a fourth embodiment of the present invention. FIG. 6 illustrates a block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention. The computer device 12 shown in FIG. 6 is only an example and should not bring any limitations to the functionality or scope of use of embodiments of the present invention.

As shown in FIG. 6, computer device 12 is in the form of a general purpose computing device. The components of computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16.

Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, micro-channel architecture (MAC) bus, enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

Computer device 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by computer device 12 and includes both volatile and nonvolatile media, removable and non-removable media.

The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)30 and/or cache memory 32. Computer device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, and commonly referred to as a "hard drive"). Although not shown in FIG. 6, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.

A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of the described embodiments of the invention.

Computer device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), with one or more devices that enable a user to interact with computer device 12, and/or with any devices (e.g., network card, modem, etc.) that enable computer device 12 to communicate with one or more other computing devices. Such communication may be through an input/output (I/O) interface 22. Also, computer device 12 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the Internet) via network adapter 20. As shown, network adapter 20 communicates with the other modules of computer device 12 via bus 18. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with computer device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

The processing unit 16 executes various functional applications and data processing, such as implementing a method of prompting for a keyword provided by an embodiment of the present invention, by executing a program stored in the system memory 28.

EXAMPLE five

The fifth embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program implements each process of the keyword prompting method, and can achieve the same technical effect, and in order to avoid repetition, the detailed description is omitted here.

A computer readable storage medium may include, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A keyword prompting method is characterized by comprising the following steps:

extracting part of the candidate keywords as target keywords;

and sending the target keywords to the client side for display.

2. The method of claim 1, wherein determining the ordered tree comprises:

acquiring a keyword for searching, wherein the keyword comprises one or more characters;

determining nodes representing the characters in the ordered tree;

and associating the key words with the nodes.

3. The method of claim 2, wherein associating the keyword with the node comprises:

inquiring a channel when the keyword is recalled for searching to be used as a recall channel;

taking the recall channel as a statistic dimension, and counting the user behaviors fed back by aiming at the keywords to obtain a search index;

inserting the keyword into the recall channel under the node;

and applying m keywords with the highest search indexes under the recall channel.

4. The method of claim 3,

the inquiring the channel of the keyword during searching as a recall channel comprises the following steps:

inquiring the geographic position of the source of the keyword during searching to serve as a recall channel;

the counting the user behaviors fed back by aiming at the keywords by taking the recall channel as a statistic dimension to obtain a search index, and the method comprises the following steps:

and taking the geographic position as a statistical dimension, and counting the number of the service information corresponding to the clicked keyword as a search index.

5. The method according to any one of claims 1-4, wherein said searching for the same and/or similar characters in the ordered tree as the original string to form a target string comprises:

determining a next node pointed to by the current node in the ordered tree;

matching the next node with the original character string;

if the matching is successful, jumping from the current node to the next node, returning to the ordered tree, and determining the next node pointed by the current node;

and if the matching fails, determining that the characters between the root node and the current node are the same as and/or similar to the original character string as a target character string.

6. The method of claim 5, wherein said matching the next said node to the original string comprises:

forming a candidate character string by characters from a root node to a next node;

calculating an edit distance between the original character string and the candidate character string;

if the editing distance is smaller than or equal to a preset threshold value, determining that the next node is successfully matched with the original character string;

7. The method according to any one of claims 1-4, wherein said querying keywords prefixed by said target string in said ordered tree as candidate keywords comprises:

inquiring a character positioned at the last position in the target character string to be used as a target character;

and extracting the keywords in all recall channels associated with the target characters to serve as candidate keywords.

8. The method according to any one of claims 1-4, wherein said extracting part of said candidate keywords as target keywords comprises:

determining the similarity between the original character string and the candidate keywords;

calculating a quality value for the candidate string based at least on the similarity, the quality value positively correlated with the similarity;

and selecting the target keyword from the candidate keywords according to the quality value.

9. The method of claim 8,

the determining the similarity between the original character string and the candidate keyword comprises:

querying a candidate character string to which the candidate keyword belongs;

inquiring the editing distance between the candidate character string and the original character string;

setting the editing distance as the similarity between the original character string and the candidate keyword;

the calculating a quality value for the candidate string based at least on the similarity includes:

10. The method according to claim 8, wherein said extracting part of said candidate keywords as target keywords further comprises:

inquiring a recall channel and a search index of the target keyword in the ordered tree;

inquiring a channel to which the client belongs as a target channel;

comparing the recall channel with the target channel to configure channel weights for the target keywords, wherein the channel weights of the recall channel and the target channel which are the same are greater than the channel weights of the recall channel and the target channel which are not the same;

11. The method of claim 8, wherein selecting the target keyword from the candidate keywords according to the quality values comprises:

sorting the candidate keywords in a descending order according to the quality values;

and selecting the k candidate keywords with the highest ranking as target keywords.

12. A keyword presentation apparatus, comprising:

13. A computer device, characterized in that the computer device comprises:

one or more processors;

a memory for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement a method of hinting for keywords as recited in any of claims 1-11.

14. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out a method of hinting for a keyword as claimed in any one of claims 1 to 11.