WO2008131698A1

WO2008131698A1 - An information searching method and corresponding device

Info

Publication number: WO2008131698A1
Application number: PCT/CN2008/070847
Authority: WO
Inventors: Nianpeng Tang; Qiang Li; Liang Sun
Original assignee: Tencent Technology (Shenzhen) Company Limited
Priority date: 2007-04-29
Filing date: 2008-04-29
Publication date: 2008-11-06
Also published as: CN101072194B; CN101072194A

Abstract

An information searching method and corresponding device are provided. The method comprises: extracting the query string from the entered information by a first user; obtaining the knowledge level on said query string for each user based on each user's information; providing at least one user's contact information to the first user based on the knowledge level. The users with needed information can be recommended to the user performing the search with the method and corresponding device, so that users can achieve needed information in more ways.

Description

Method and device for searching information

Technical field

The invention relates to the field of search engine, in particular to a method and device for searching information. Background of the invention

Search engine technology refers to a database created by extracting information from various websites (mainly webpage text) from the Internet, and searching related records matching user query conditions (also called query strings, search strings), and then pressing A certain sort order gives the results back to the user.

Internet information to be processed by search engines has characteristics such as massive, heterogeneous, and hypermedia. It uses a variety of computer science theories and cornerstones, such as artificial intelligence, machine learning, data mining, information retrieval, information extraction, text processing, natural language. Processing, information theory, data compression, architecture, parallel computing, database technology, etc.

However, although the content of the Internet is already very rich, the knowledge acquired by human beings is more abundant, and there is a lot of knowledge in people's brains. For various reasons, it may not necessarily be on the Internet; other information is even on the Internet. However, due to the limitations of current search technology, people may not be able to easily find the information they want, such as the download address of a movie, and the search results in front may not be connected. In short, current methods and systems for information search rely too much on the knowledge provided on the web (mainly web pages) to fully satisfy people's search needs. Summary of the invention

The embodiment of the invention provides a method and a device for searching for information, which can recommend other users who can provide the required information to the user who performs the information search, so that the user can obtain the required information through more channels. To solve the above technical problem, an embodiment of the present invention provides a method for information search, including:

Extracting a query string from the information input by the first user, obtaining a cognitive level of each query string for each user according to the information of each user, and providing at least one user contact to the first user according to the cognitive level information.

An embodiment of the present invention further provides an information search device, including:

a cognitive level setting device, configured to extract a query string from information input by the first user, and acquire, by each user information, a cognitive level of each query string by the user; and

And a recommending device, configured to provide the first user with contact information of the at least one user according to the cognitive level acquired by the cognitive level setting device.

The method and device for information search according to the embodiment of the present invention, by determining the cognitive level of each user for a certain query string, and then recommending to the user who performs the information search a user with a higher cognition level of the query string of the searched information. Therefore, users who conduct information search can obtain the required knowledge through more ways. BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flowchart of a method for information search provided by an embodiment of the present invention.

FIG. 2 is a flowchart of step A in the method for information search provided by the embodiment of the present invention. FIG. 3 is a flowchart of step B in the method for information search provided by the embodiment of the present invention. FIG. 4 is a schematic diagram of an apparatus for information search provided by an embodiment of the present invention. Mode for carrying out the invention

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. An embodiment of the present invention mainly obtains a query string from the information input by the first user, and acquires a cognitive level of each query string by each user according to the information of each user, according to the cognitive level. A user provides contact information for one or more other users.

Here, the query string refers to a series of combinations of words, symbols, and numbers used for searching (generally text-based, which can contain symbols and numbers), such as "Olympics", "Olympic Torch", "2008 Olympics", etc. Query string.

Cognitive level is used to indicate the user's understanding of a certain aspect of knowledge. The higher the level of cognition, the more knowledge the user knows about this aspect. A hierarchical level of cognition can be used, which divides the level of cognition into several levels, such as primary, intermediate, advanced, etc., and can also use numerical cognitive levels, etc., here no longer - for example.

For example, the system may provide the user who performs the information search with contact information of one or more other users, such as the forum identity or email address of the one or more users, etc., so that the user can associate with the one or Communicate with multiple users. In order to enable the user who conducts the information search to contact the one or more users more quickly and conveniently, the instant communication account of the one or more users may be provided to the user who performs the information search, and the user can immediately perform the information through the instant communication system. communicate with. The system can ask the user to provide contact information such as their email address, instant messaging account, mobile number, etc. when the user registers, and save the information.

In another embodiment, the user can log in to the instant messaging system to search for information through the search interface provided by the instant messaging system. Instant messaging system means that users can communicate in real time through various Internet devices and mobile networks through various terminal devices, and immediately send text, image, video, audio and email information to designated recipients. The recipient can reply immediately or reply later. . To use an instant messaging system, users must first apply for an instant messaging system account. Today, instant messaging systems have become an important means of communication for the public to communicate in real time. The existing instant messaging system provides the function of information search, and must pass the information search. The account is logged into the instant messaging system. After the user inputs the keyword of the information to be searched, the search engine searches for the matching information in the database of the Internet according to the input keyword, and provides the result to the user.

Referring to FIG. 1 to FIG. 3, a method for information search provided by an embodiment of the present invention includes the following steps:

A. Determine the level of cognition of each query string for each user, where the user uses his or her unique identifier, such as an instant messaging system account, to identify. This step may specifically include the following steps.

Al, the user logs in to the system, for example, logs in to the instant messaging system through the account of the instant messaging system, logs in to the web forum through the web page, and the like.

A2. The user inputs the search information, and the system extracts a query string from the search information input by the user (hereinafter referred to as a query string input by the user). The search information can be input in the webpage page, or input in the search interface of the instant messaging system client tool, or input into the instant messaging system through the webpage, and input in the webpage page provided by the instant messaging system. Here, the user input search information can be used to search for a query string or register a query string. The user's cognitive level can be changed in the above two ways. One is to increase the number of searches by a conventional search, and the like, and the other is to register the query string by the above. When a user registers a query string, the system can increase the user's level of awareness of the query string.

A3. Calculate the matching degree of the query string by using a vector space model or by string comparison, that is, the similarity calculation is performed between the query string input by the user and the query string in the database. Among them, the vector space model is a kind of search model. The feature item of the vector is generally the weight of the morpheme. It can be calculated by tf*idf, and the similarity between the two query strings is determined by the similarity calculation of the vector space. The vector space model calculates the formula of two query strings in general. Next:

, where "" is the vector of the query string saved by the database, ^, ^, ^, ^ is the vector of the query string input by the user;

The degree of matching is calculated by string comparison by calculating the sum of the proportions of the same string to calculate the similarity. There are many ways to implement string comparison algorithms, such as

Sim(d , q) U - U ( -^)□-□ (^)

Calculate the formula as follows: ^{2 2} , where L _d is the length of the query string in the database, which is the length of the string completely repeated between the query string in the database and the query string input by the user, and L _q is the length of the query string input by the user. L ₂ is the length of the string that the user-entered query string completely repeats with the query string in the database. If the two strings are completely repeated, the value of Sim(d, q) is 1.

There are many ways to calculate the matching degree between the query string input by the user and the query string in the database, and here is no longer - enumeration.

A4. After searching for the most relevant query string, store the number of queries of the user and the user identifier, for example, the account of the instant messaging system, into the database; if there is no related query string, save the query string input by the user to In the database. If the user is to register a query string, the information of the user's registration query string is saved. A user can register multiple query strings, indicating that the user knows more about these query strings.

A5. Determine the level of cognition of the most relevant query string query string for each user (e.g., the user identified by the account of the instant messaging system).

According to the static information of the user, or according to the dynamic information of the user, the cognitive level of the query string of the instant messaging system account may be determined according to the static information and the dynamic information of the user. User's static information and dynamic information can be served by The server device is recorded in the database and modified according to the operation of the user's modified data or the user's use of the system. The static information of the user is user information with a relatively low frequency of change, such as registration information submitted by the user when registering the account, personal information of the user, such as the age, interest of the user, whether the invited expert, the registered query string information, etc. You can modify this information as needed. User dynamic information refers to user information that changes frequently, such as the number of queries for a query string, the span of query time, the number of times to answer questions, and online conditions. For example, for a person who has a large number of queries, he must be interested in this question. If he can't find the answer on the Internet, he may also get the answer by consulting other materials and asking for advice, so according to the number of queries and the time of the query. It can be determined that its cognitive level is another ij; of course, it is also possible to increase the user's cognitive level by voting by other users. The system needs to update the user's dynamic information according to certain rules, such as when the user logs in or logs out of the system, or at regular intervals, or in batches. A hierarchical level of cognition can be used, which divides the level of cognition into different levels, such as low, medium, and high. A digital cognitive level can also be used, and the user's cognitive level can be a specific number. Of course, the system can save the number of the user's cognitive level, but set different thresholds to display the user's cognitive level as a different level. For example, the cognitive level of 8.5 or higher is displayed as "advanced" and the like. There are other various implementations, and will not be described here.

The calculation of the cognitive level may include the following specific parameters.

User's query count factor. The query number factor can be calculated by selecting an algebraic expression containing the number of user queries as needed. When calculating a user's cognitive level of a certain query string, we also need to consider the behavior of some users cheating in order to get recommendations, such as too many queries in a small time period, etc., you can choose to use Wa=^, etc. To calculate the number of queries, where n is the number of queries, which can alleviate the shadow caused by the above multiple short queries. ring.

Time weight, which indicates how long the user pays attention to a problem. Assuming that the earliest query time is T1 and the most recent query time is T2, the time weight can be W ^{= lQgCr2} - x. This formula indicates that the user pays attention to a problem for a long time, and the number of times of attention is high, then the user's understanding of the problem is Higher.

You can also consider the relevance weight Wb of the registration information and the query, the user level weight Wc (for example, according to the user's points, etc.), the user online weight Wd, the age weight We, because the adult is more capable of solving problems, faster Therefore, for a relatively old age, and a longer network age, a certain weighting, and the user's interest weight Wf, may be based on the user's account registration information, or may be determined according to the user's browsing habits. If the user frequently browses the computer hardware channel, the user may have a more professional hardware configuration of the computer, and may also be based on the keyword characteristics and categories searched by the user, and the web page classification characteristics of the search results. If the user's interest matches the query string, the weight can be increased. If the user has registered the query string, a higher weight value for the query string can be set for the user. Other factors can also be considered to get other weight values, which are not listed here.

Factors that need to be considered can be selected to determine the user's level of awareness, as needed. For example, you can sum the weights of ownership, such as ^ ^{= +} ^ ^{+ +} ^ ^{+ +} ^ , or select one or some of the weight values to calculate, or you can set different weighting factors for different weights for each weight. Values are weighted and summed to reflect the importance of different weights, etc. This calculation can be done in a variety of ways, just to name a few, no longer for other calculation methods - enumeration.

In addition, the timing of calculating the cognitive level of the user may also be selected according to needs, and may be any manner, for example, when the user ends the information search, the user is searched for all the query strings or all the search strings of the user. Cognitive level, or the system performs cognitive level calculations for all users on a regular and/or batch basis, etc. For newly registered users, or Users who have not queried a query string can default their cognition level for the query string to the lowest level or value.

After determining the level of awareness of each user on the most relevant query string, you can also take steps

A6.

A6. Establish an index of all query strings and users in the database according to the number of times the user queries and the instant messaging account.

B. Recommend one or more users to the user who performs the information search according to the query string cognition level of the search information, which specifically includes the following steps.

Bl. The user inputs the search information in the instant messaging system, and can input in the search interface of the client tool of the instant messaging system, or log in to the account of the instant messaging system through the webpage, and input in the webpage page.

B2, extracting a query string (such as a keyword, etc.) from the above search information, such as removing some meaningless interjections and auxiliary words, and of course, directly using the search information input by the user as the query string.

B3. Search the database for the query string most relevant to the query string.

B4. Calculate the matching degree by using a vector space model or by string comparison, and calculate the similarity between the search result and the query string.

B5. Select one or more users from the account of all the instant messaging systems according to the cognitive level of the most relevant query string, and recommend to the user who performs information search, for example, a user with a higher cognitive level may be selected to contact the user. Information, such as an instant messaging account, is provided to users who search for information. For the case where the relevant query string is not searched, the system may also select a query string that is the same or similar to the domain of the query string, or provide one or more fields or one or more query strings that are closer to the user for the user. Selecting, determining a query string as a related query string according to the user's selection result.

The selected one or more users may include the user who performs the information search, or One or more users are selected among other users than the user who performs the information search. The selection strategy here is also diverse, such as selecting the user with the highest cognitive level, or selecting all users with a cognitive level of a certain level, or selecting a cognitive level higher than a certain value, or providing all users with For users who search for information. When one or more users' contact information is provided, the users can also be arranged in a certain order, for example, in descending order of cognitive level, in descending order of age, etc., here no longer - for example. When users with higher cognitive levels are recommended, only the users with the highest cognitive level can be provided according to the system settings or according to the instructions of the users. The contact information of one or several levels of users can also be provided, for example, the cognitive level is "Advanced". Users with "and higher level" can also set or specify the threshold of the cognitive level by the user, and provide all users whose cognitive level is higher than the threshold. Users who provide higher cognitive levels can also follow the user's cognitive level. Sorted from high to low and displayed to users who search for information. You can also provide only current online users according to system settings or user instructions. Any feasible implementation can be applied here, so here is no longer - an example.

Step B5 can be implemented in two ways. The first method is to determine, by step A5, the cognitive level of each user to the most relevant query string, and directly select one or more users according to the cognitive level according to the cognitive level. The contact information of one or more users is provided to the user who searches for information; the second is that after performing step A6, the users listed in the index are obtained from the most relevant query string and the index of the user, and the users are obtained. The cognitive level, then select one or more users according to a certain strategy, and provide the contact information of the one or more users to the user who performs the information search.

Referring to FIG. 4, an apparatus for performing information search by using an instant messaging system according to an embodiment of the present invention may include:

a cognitive level setting device, configured to determine a cognitive level of each instant messaging system account for a certain query string; And a recommending device, configured to select at least one user according to a cognitive level of the query string of the information to be searched, and provide the contact information to the user who performs the information search.

The cognitive level setting device may include:

The login module is configured to authenticate the user and allow the user to log in when the user logs in to the instant messaging system through the account of the instant messaging system;

a query string input module, configured to input, by the user, the searched information in the instant messaging system, and extract a query string from the information, wherein the user inputs the information in order to search for a query string or register a query string;

The query string similarity calculation module is configured to calculate a matching degree by using a vector space model or by string comparison, and perform similarity calculation between the query string and the query string in the database;

An information storage module, configured to store the number of times the user queries, the account of the instant messaging system, and the query string not in the database are stored in the database;

a cognitive level calculation module, configured to calculate a cognitive level of an account of each instant messaging system for a query string;

An index establishing module is configured to establish an index of a user who has queried the query string for each query string in the database according to the number of times of the user and the instant messaging system account.

The method and device for searching for information of the embodiment of the present invention can recommend a user with a higher level of cognition of the query string of the searched information to the user who performs the information search, so that the user who performs the information search obtains the required content through more channels. Knowledge can also enhance communication between people with the same interests, and the exchange of knowledge may produce new knowledge, so it has certain social benefits.

The above is only the preferred embodiment of the present invention, and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection of the present invention. Within the scope.

Claims

Claim

A method for searching for information, characterized in that it comprises:

2. The method of claim 1 further comprising: recording information for each user.

The method according to claim 1 or 2, wherein the information of the user comprises at least one of the following: static information and dynamic information.

4. The method according to claim 3, wherein the static information of the user comprises at least one of the following:

The registration information of the user, the personal information of the user, the age and interest of the user; the dynamic information of the user includes at least one of the following:

The number of times the user queries the query string, the query time span, the information of the user registration query string, the number of times the question is answered, the online status, and the number of votes cast by other users to the user.

The method of claim 4, wherein the recording the information of each user comprises:

Recording information of the user registration query string when receiving the request for the user to register a query string; and/or

The number of queries of the query string by the user is increased when the request for the user search query string is received.

The method according to any one of claims 1 to 5, further comprising: establishing an index of the query string and at least one user; The providing the first user with the contact information of the at least one user according to the cognitive level includes:

Selecting at least one user from the index of the query string in descending order of cognition level, and providing contact information of the selected at least one user to the first user.

7. A method according to any one of claims 1 to 6, further comprising:

Receiving information input by the first user from a search interface of a client of the instant messaging system, or

Receiving information input by the first user through a webpage.

The method according to any one of claims 1 to 7, wherein the contact information of the at least one user is an instant messaging account of the at least one user.

9. An information search device, comprising: a storage device for storing at least one query string, comprising:

10. The device according to claim 9, wherein the cognitive level setting device comprises:

An information storage module, configured to record information of each user;

And a cognitive level calculation module, configured to calculate, according to information of each user stored in the information storage module, a cognitive level of each query string by the user.

The device according to claim 9 or 10, wherein said cognitive level setting means further comprises:

An index establishing module, configured to establish an index of the query string and at least one user; The recommending device is configured to select at least one user from an index of the query string according to a cognitive level from high to low, and provide contact information of the selected at least one user to the user.