CN108241631B - Method and device for pushing information - Google Patents
Method and device for pushing information Download PDFInfo
- Publication number
- CN108241631B CN108241631B CN201611206433.8A CN201611206433A CN108241631B CN 108241631 B CN108241631 B CN 108241631B CN 201611206433 A CN201611206433 A CN 201611206433A CN 108241631 B CN108241631 B CN 108241631B
- Authority
- CN
- China
- Prior art keywords
- pushed
- sentence
- identification information
- statement
- sentences
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Machine Translation (AREA)
Abstract
The application discloses a method and a device for pushing information. One embodiment of the method comprises: generating sequence characteristics for each statement to be pushed in a statement set to be pushed obtained in advance by using a depth sequence model trained in advance; using the sequence characteristics of the sentences to be pushed and the context information between the words contained in the sentences to be pushed as the input characteristics of a pre-trained conditional random field model for sequence labeling, thereby obtaining the identification information in the sentences to be pushed; and matching the obtained identification information of each statement to be pushed with the identification information in the identification information set of the user obtained in advance, and selecting the statement to be pushed from the statement set to be pushed according to the matching result and pushing the statement to be pushed to the terminal. The embodiment can accurately identify the identification information in the sentence to be pushed, so that the user can effectively avoid legal risks caused by the identification information contained in the promotion sentence.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to the field of internet technologies, and in particular, to a method and an apparatus for pushing information.
Background
Search Engine Marketing (SEM) is to deliver Marketing information to target users as much as possible using the user's opportunity to retrieve information based on the way the user uses the Search Engine. With the rapid development of internet technology, more enterprises begin to conduct business promotion through search engine marketing, and in the promotion process, keywords or sentences set by the enterprises through a search engine marketing system directly influence the promotion effect.
In order to enable an enterprise to describe the promoted business more accurately and comprehensively and improve the promotion effect, the search engine marketing system recommends keywords or sentences for the enterprise to be selected and purchased by a user. However, if an enterprise purchases a keyword or a sentence containing other enterprise identification information (e.g., brand name, company name, etc.), legal risk is often brought to the enterprise, and the search experience of the network user is also reduced.
Disclosure of Invention
The present application aims to provide a method and an apparatus for pushing information, so as to solve the technical problems mentioned in the above background section.
In a first aspect, the present application provides a method for pushing information, including: generating sequence characteristics for each statement to be pushed in a statement set to be pushed obtained in advance by using a depth sequence model trained in advance; using the sequence characteristics of the sentences to be pushed and the context information between words contained in the sentences to be pushed as input characteristics of a pre-trained conditional random field model for sequence labeling, thereby obtaining identification information in the sentences to be pushed; and matching the obtained identification information of each statement to be pushed with the identification information in the identification information set of the user obtained in advance, and selecting the statement to be pushed from the statement set to be pushed according to the matching result and pushing the statement to be pushed to the terminal.
In some embodiments, the set of identification information of the user is obtained by: generating sequence characteristics for each sentence in at least one sentence preset by the user by using the depth sequence model; and performing sequence labeling by using the sequence characteristics of the sentences and the context information between the words contained in the sentences as the input characteristics of the conditional random field, so as to obtain the identification information in at least one sentence preset by the user, and forming an identification information set by using the obtained identification information.
In some embodiments, the matching the obtained identification information of each to-be-pushed statement with the identification information in the pre-obtained identification information set of the user, and selecting the to-be-pushed statement from the to-be-pushed statement set according to a matching result to push the to-be-pushed statement to the terminal includes: for each statement to be pushed in the statement set to be pushed, executing the following steps: comparing the obtained identification information of the sentence to be pushed with each identification information in the identification information set of the user; if the identification information of the statement to be pushed is different from the identification information in the identification information set of the user, deleting the statement to be pushed from the statement set to be pushed; and pushing the remaining sentences to be pushed in the sentence set to the terminal.
In some embodiments, the set of statements to be pushed is obtained by: and obtaining a sentence set to be pushed according to at least one sentence preset by the user through the terminal.
In some embodiments, the obtaining a set of sentences to be pushed according to at least one sentence preset by the user through the terminal includes: and searching in a preset sentence set by using at least one sentence set by the user, and forming a sentence set to be pushed by using the searched sentences.
In some embodiments, the depth sequence model is trained by: selecting sample sentences used for model training from an information base; and training and generating a depth sequence model for generating sequence features by using the selected sample sentences.
In some embodiments, the conditional random field model is trained by: generating sequence features for each training sentence in a training sentence set by using the depth sequence model, wherein each training sentence in the training sentence set is labeled in advance; and training and generating a conditional random field model for sequence labeling by using the sequence features of each training sentence and context information among words contained in each training sentence.
In a second aspect, the present application provides an apparatus for pushing information, comprising: the generating unit is used for generating sequence characteristics for each sentence to be pushed in a sentence set to be pushed, which is obtained in advance, by using a depth sequence model trained in advance; the labeling unit is used for performing sequence labeling by using the sequence characteristics of the sentences to be pushed and the context information among the words contained in the sentences to be pushed as the input characteristics of a pre-trained conditional random field model so as to obtain the identification information in the sentences to be pushed; and the pushing unit is used for matching the obtained identification information of each statement to be pushed with the identification information in the pre-obtained identification information set of the user, and selecting the statement to be pushed from the statement set to be pushed according to the matching result and pushing the statement to be pushed to the terminal.
In some embodiments, the apparatus further includes an identification information set obtaining unit, where the identification information set obtaining unit is configured to: generating a sequence feature for each sentence in at least one sentence preset by the user by using the depth sequence model; and performing sequence labeling by using the sequence characteristics of each sentence and the context information between words contained in each sentence as the input characteristics of the conditional random field to obtain the identification information in at least one sentence preset by the user, and forming an identification information set by using the obtained identification information.
In some embodiments, the pushing unit is further configured to: for each statement to be pushed in the statement set to be pushed, executing the following steps: comparing the obtained identification information of the sentence to be pushed with each identification information in the identification information set of the user; if the identification information of the statement to be pushed is different from the identification information in the identification information set of the user, deleting the statement to be pushed from the statement set to be pushed; and pushing the remaining sentences to be pushed in the sentence set to be pushed to the terminal.
In some embodiments, the apparatus further includes a statement set to be pushed obtaining unit, where the statement set to be pushed obtaining unit is configured to: and obtaining a sentence set to be pushed according to at least one sentence preset by the user through the terminal.
In some embodiments, the statement set to be pushed obtaining unit is further configured to: and searching in a preset sentence set by using at least one sentence set by the user, and forming a sentence set to be pushed by using the searched sentences.
In some embodiments, the depth sequence model is trained by: selecting sample sentences used for model training from an information base; and training and generating a depth sequence model for generating sequence features by using the selected sample sentences.
In some embodiments, the conditional random field model is trained by: generating sequence features for each training sentence in a training sentence set by using the depth sequence model, wherein each training sentence in the training sentence set is labeled in advance; and training and generating a conditional random field model for sequence labeling by using the sequence features of each training sentence and context information among words contained in each training sentence.
According to the method and the device for pushing the information, the depth sequence model is used for generating sequence characteristics for each sentence to be pushed in the sentence set to be pushed, then the sequence characteristics of each sentence to be pushed and context information between words contained in each sentence to be pushed are used as input characteristics of the conditional random field model for sequence marking, identification information in each sentence to be pushed is obtained, finally the obtained identification information of each sentence to be pushed is matched with the identification information in the identification information set of the user, the sentence to be pushed is selected from the sentence set to be pushed according to a matching result and pushed to the terminal, and therefore the identification information in the sentence to be pushed is accurately identified, and legal risks caused by the identification information contained in the promotion sentences are effectively avoided for the user.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
FIG. 1 is an exemplary system architecture diagram in which the present application may be applied;
FIG. 2 is a flow diagram of one embodiment of a method for pushing information, according to the present application;
FIG. 3 is a schematic diagram of an application scenario of a method for pushing information according to the present application;
FIG. 4 is a schematic block diagram illustrating one embodiment of an apparatus for pushing information according to the present application;
fig. 5 is a schematic structural diagram of a computer system suitable for implementing the terminal device or the server according to the embodiment of the present application.
Detailed Description
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the related invention are shown in the drawings.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
Fig. 1 shows an exemplary system architecture 100 to which embodiments of the present method for pushing information or apparatus for pushing information may be applied.
As shown in fig. 1, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the terminal devices 101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.
A user may use terminal devices 101, 102, 103 to interact with a server 105 over a network 104 to receive or send messages or the like. The terminal devices 101, 102, 103 may have installed thereon various communication client applications, such as a web browser application, a shopping-type application, a search-type application, and the like.
The terminal devices 101, 102, 103 may be various electronic devices having a display screen and selecting and purchasing keywords or sentences, including, but not limited to, a smart phone, a tablet computer, an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III, Moving Picture Experts compression standard Audio Layer 3), an MP4 player (Moving Picture Experts Group Audio Layer IV, Moving Picture Experts compression standard Audio Layer 4), a laptop portable computer, a desktop computer, and the like.
The server 105 may be a server providing various services, such as a background server pushing statements to be pushed for the terminal devices 101, 102, 103. The background server may select a statement to be pushed from the statement set to be pushed and push the selected statement to the terminal devices 101, 102, and 103.
It should be noted that the method for pushing information provided by the embodiment of the present application is generally performed by the server 105, and accordingly, the apparatus for pushing information is generally disposed in the server 105.
It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for an implementation.
With continued reference to fig. 2, a flow 200 of one embodiment of a method for pushing information in accordance with the present application is shown. The method for pushing the information comprises the following steps:
In this embodiment, an electronic device (for example, the server 105 shown in fig. 1) on which the method for pushing information operates may generate a sequence feature for each statement to be pushed in a statement set to be pushed using a depth sequence model, where the depth sequence model may be various machine learning models obtained by pre-training and used for generating the sequence feature of the statement to be pushed, such as a pre-trained LSTM (Long Short-Term Memory) model, an RNN (Recurrent neural Network) model, and so on. For each statement to be pushed, the electronic device may use the statement to be pushed as an input of a depth sequence model, so as to generate a sequence feature corresponding to the statement to be pushed, where when the statement to be pushed is composed of a plurality of words or words, the sequence feature corresponding to the statement to be pushed is also composed of a sequence of the plurality of words or words, and the sequence feature may be in the form of a digital vector. Here, the electronic device may obtain a statement set to be pushed in advance through various manners, where the statement set to be pushed includes at least one statement to be pushed for the terminal device. For example, the electronic device may obtain the statement set to be pushed according to a service to be promoted by an enterprise user who uses search engine marketing to commercially promote, for example, the statement set to be pushed may be formed by statements set by other enterprise users who operate the same service as the enterprise user.
In some optional implementations of the present embodiment, the depth sequence model may be trained by: first, a sample sentence for model training may be selected from an information base, where the information base may include materials of an enterprise user who performs business promotion by using search engine marketing (for example, keywords, sentences, creatives, promotion regions, and the like set by the enterprise user); then, the selected sample sentences are used for training to generate a depth sequence model for generating sequence features, for example, taking training of an LSTM model as an example, for each sample sentence, the sample sentence is firstly segmented into a sequence of words or characters, and then the sequence of words or characters included in the sample sentence is input into the LSTM model for training, so as to obtain the depth sequence model for generating the sequence features.
In some optional implementation manners of this embodiment, the statement set to be pushed may be obtained by: and the electronic equipment obtains a sentence set to be pushed according to at least one sentence preset by the user through the terminal. For example, the electronic device may use an existing word expansion tool to expand at least one sentence preset by a user through a terminal, so as to obtain a sentence set to be pushed. Here, the user may refer to an enterprise user who uses search engine marketing to make a business promotion, and the at least one sentence may refer to a sentence set by the enterprise user for the business promotion.
In some optional implementation manners, the obtaining a sentence set to be pushed according to at least one sentence preset by the user through the terminal may specifically include: and searching in a preset sentence set by using at least one sentence set by the user, and forming a sentence set to be pushed by using the searched sentences. Due to the limitation of knowledge and experience of enterprise users, at least one statement set by the enterprise users often cannot accurately and comprehensively describe the promoted service, so that more statements can be retrieved from a preset statement set to form a statement set to be pushed based on the at least one statement set by the enterprise users. The sentence set may be an industry sentence library obtained in various manners (e.g., manually set manners), and the industry sentence library may include a large number of sentences capable of accurately describing the industry service.
In this embodiment, the electronic device may use the sequence features of the sentences to be pushed generated in step 201 and context information between the words included in the sentences to be pushed as input features of a pre-trained Conditional Random Field (CRF) model for sequence tagging, that is, use the conditional random field model to tag the sequence features corresponding to the sentences to be pushed and tag whether each word included in the sentences to be pushed is identification information, so as to obtain identification information in each sentence to be pushed. The context information between words included in the statement to be pushed may refer to information for describing a context relationship between words included in the statement to be pushed, for example, the statement to be pushed is "ABC housekeeping", and the context information of the statement may be used for describing a word following the word "ABC" as "housekeeping". Here, the identification information may be information such as a brand name and a company name. It should be noted that some statements to be pushed may not contain identification information. At this time, only the identification information in the statement to be pushed, which contains the identification information, needs to be identified.
In some optional implementations of this embodiment, the conditional random field model may be trained by: the electronic device or other electronic devices for training the conditional random field model may first generate sequence features for each training sentence in a training sentence set using the depth sequence model, where each training sentence in the training sentence set is labeled in advance, for example, each training sentence may be manually labeled to mark identification information in the training sentence, and if the training sentence is "miss family service", the brand name "miss" may be labeled as a brand, and the words "home" and "service" are labeled as common words; the conditional random field model for sequence labeling may then be trained using the sequence features of each training sentence and the context information between the words contained in each training sentence.
And 203, matching the obtained identification information of each statement to be pushed with the identification information in the pre-obtained identification information set of the user, and selecting the statement to be pushed from the statement set to be pushed according to the matching result to push to the terminal.
In this embodiment, the electronic device may obtain an identification information set of the user in advance, match the identification information of each to-be-pushed sentence obtained in step 202 with the identification information in the identification information set of the user, and finally select a to-be-pushed sentence from the to-be-pushed sentence set according to a matching result and push the selected to-be-pushed sentence to the terminal. Here, the user may refer to an enterprise user who uses search engine marketing to implement business promotion, the enterprise user may set a plurality of statements for describing a business to be promoted through a search engine marketing account, and identification information included in the statements set by the enterprise user may form an identification information set of the enterprise user. The terminal may refer to a terminal used by an enterprise user.
In some optional implementations of this embodiment, the set of identification information of the user may be obtained by: firstly, the electronic device may generate sequence features for each of at least one sentence preset by the user by using the depth sequence model, where the at least one sentence preset by the user may be at least one sentence which is set by an enterprise user through a search engine marketing account and used for describing a service to be promoted, and the at least one sentence may include identification information such as a brand name, a company name, and an enterprise name, for example, an enterprise a selling flowers has three brands "a 1", "a 2", and "A3", and may set sentences such as "a flower express delivery", "a 1 flower express delivery", "a 2 flower", "A3 flower type", "what the flower express delivery is good" and the like through the search engine marketing account; then, the electronic device may perform sequence labeling using the sequence features of the sentences and the context information between the words included in the sentences as the input features of the conditional random field, thereby obtaining the identification information in at least one sentence preset by the user, and compose an identification information set using the obtained identification information.
In some optional implementations of this embodiment, the step 203 may specifically include: first, for each statement to be pushed in the statement to be pushed set, the electronic device may perform the following steps: comparing the obtained identification information of the sentence to be pushed with each identification information in the identification information set of the user; if the identification information of the statement to be pushed is different from the identification information in the identification information set of the user, deleting the statement to be pushed from the statement set to be pushed; then, the electronic device may push the remaining statements to be pushed in the statement to be pushed set to the terminal.
With continued reference to fig. 3, fig. 3 is a schematic diagram of an application scenario of the method for pushing information according to the present embodiment. In the application scenario of fig. 3, an enterprise user a selling flowers owns three brands "a 1", "a 2", and "A3", and the enterprise user a purchases statements "a flower express delivery", "a 1 flower express delivery", "a 2 flower", "how the flower 3 flower looks" and which the flower express delivery is good "for commercial promotion through a search engine marketing system, and first, the server generates sequence features for each to-be-pushed statement in a to-be-pushed statement set by using a depth sequence model, where the to-be-pushed statement set is obtained according to the statements purchased by the enterprise user a and used for commercial promotion; then, the server uses the sequence characteristics of the sentences to be pushed and the context information between the words contained in the sentences to be pushed as the input characteristics of the conditional random field model for sequence labeling, so as to obtain the identification information in the sentences to be pushed; finally, matching the obtained identification information of each to-be-pushed sentence with the identification information "a", "a 1", "a 2" and "A3" in the identification information set of the enterprise user a, if the identification information of the to-be-pushed sentence is different from the identification information in the identification information set of the enterprise user a, deleting the to-be-pushed sentence, and pushing the remaining to-be-pushed sentences "how flowers are delivered", "flower is delivered", "online booking", and delivery in 1-2 hours "in the to-be-pushed sentence set to the terminal used by the enterprise user a for the enterprise user to select and purchase, as shown in fig. 3, the enterprise user can purchase the selected sentence by clicking a button 301.
According to the method provided by the embodiment of the application, the identification information of the sentences to be pushed can be accurately identified by using the depth sequence model and the conditional random field model, so that the sentences to be pushed containing identification information which is not matched with the enterprise user are eliminated, the enterprise user effectively avoids legal risks caused by the non-matching identification information contained in the popularization sentences, and meanwhile, the search experience of the network user is improved.
With further reference to fig. 4, as an implementation of the methods shown in the above-mentioned figures, the present application provides an embodiment of an apparatus for pushing information, where the embodiment of the apparatus corresponds to the embodiment of the method shown in fig. 2, and the apparatus may be applied to various electronic devices.
As shown in fig. 4, the apparatus 400 for pushing information according to this embodiment includes: a generation unit 401, a labeling unit 402 and a pushing unit 403. The generation unit 401 is configured to generate, by using a depth sequence model trained in advance, sequence features for each to-be-pushed sentence in a to-be-pushed sentence set obtained in advance; the labeling unit 402 is configured to perform sequence labeling by using the sequence features of the sentences to be pushed and context information between words included in the sentences to be pushed as input features of a pre-trained conditional random field model, so as to obtain identification information in the sentences to be pushed; the pushing unit 403 is configured to match the obtained identification information of each to-be-pushed sentence with identification information in a pre-obtained identification information set of the user, and select a to-be-pushed sentence from the to-be-pushed sentence set according to a matching result and push the selected to-be-pushed sentence to the terminal.
In this embodiment, the detailed descriptions of step 201, step 202, and step 203 in the embodiment of fig. 2 may be referred to for the specific processing of the generating unit 401, the labeling unit 402, and the pushing unit 403, and are not repeated herein.
In some optional implementations of this embodiment, the apparatus further includes an identification information set obtaining unit (not shown), where the identification information set obtaining unit is configured to: generating a sequence feature for each sentence in at least one sentence preset by the user by using the depth sequence model; and performing sequence labeling by using the sequence characteristics of each sentence and the context information between words contained in each sentence as the input characteristics of the conditional random field to obtain the identification information in at least one sentence preset by the user, and forming an identification information set by using the obtained identification information. For the implementation, reference may be made to the detailed description of the corresponding implementation in the embodiment corresponding to fig. 2, which is not described herein again.
In some optional implementations of this embodiment, the pushing unit 403 is further configured to: for each statement to be pushed in the statement set to be pushed, executing the following steps: comparing the obtained identification information of the sentence to be pushed with each identification information in the identification information set of the user; if the identification information of the statement to be pushed is different from the identification information in the identification information set of the user, deleting the statement to be pushed from the statement set to be pushed; and pushing the remaining sentences to be pushed in the sentence set to the terminal. For the implementation, reference may be made to the detailed description of the corresponding implementation in the corresponding embodiment of fig. 2, which is not described herein again.
In some optional implementation manners of this embodiment, the apparatus further includes a statement set to be pushed obtaining unit (not shown), where the statement set to be pushed obtaining unit is configured to: and obtaining a sentence set to be pushed according to at least one sentence preset by the user through the terminal. For the implementation, reference may be made to the detailed description of the corresponding implementation in the embodiment corresponding to fig. 2, which is not described herein again.
In some optional implementation manners of this embodiment, the to-be-pushed statement set obtaining unit is further configured to: and searching in a preset sentence set by using at least one sentence set by the user, and forming a sentence set to be pushed by using the searched sentences. For the implementation, reference may be made to the detailed description of the corresponding implementation in the embodiment corresponding to fig. 2, which is not described herein again.
In some optional implementations of this embodiment, the depth sequence model is obtained by training in the following manner: selecting sample sentences used for model training from an information base; and training and generating a depth sequence model for generating sequence features by using the selected sample sentences. For the implementation, reference may be made to the detailed description of the corresponding implementation in the embodiment corresponding to fig. 2, which is not described herein again.
In some optional implementations of this embodiment, the conditional random field model is trained by: generating sequence features for each training sentence in a training sentence set by using the depth sequence model, wherein each training sentence in the training sentence set is labeled in advance; and training and generating a conditional random field model for sequence labeling by using the sequence features of the training sentences and context information between words contained in the training sentences. For the implementation, reference may be made to the detailed description of the corresponding implementation in the corresponding embodiment of fig. 2, which is not described herein again.
Referring now to FIG. 5, a block diagram of a computer system 500 suitable for use in implementing a terminal device or server of an embodiment of the present application is shown.
As shown in fig. 5, the computer system 500 includes a Central Processing Unit (CPU)501 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)502 or a program loaded from a storage section 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data necessary for the operation of the system 500 are also stored. The CPU 501, ROM 502, and RAM 503 are connected to each other via a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.
The following components are connected to the I/O interface 505: an input portion 506 including a keyboard, a mouse, and the like; an output portion 507 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage portion 508 including a hard disk and the like; and a communication section 509 including a network interface card such as a LAN card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. A drive 510 is also connected to the I/O interface 505 as needed. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as necessary, so that a computer program read out therefrom is mounted into the storage section 508 as necessary.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 509, and/or installed from the removable medium 511. The computer program performs the above-described functions defined in the method of the present application when executed by the Central Processing Unit (CPU) 501.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units described in the embodiments of the present application may be implemented by software or hardware. The described units may also be provided in a processor, which may be described as: a processor includes a generation unit, a labeling unit, and a pushing unit. The names of the units do not form a limitation on the units themselves in some cases, for example, the generating unit may also be described as a unit that generates sequence features for each sentence to be pushed in a pre-obtained sentence set to be pushed by using a pre-trained deep sequence model.
As another aspect, the present application also provides a non-volatile computer storage medium, which may be the non-volatile computer storage medium included in the apparatus in the above-described embodiments; or it may be a non-volatile computer storage medium that exists separately and is not incorporated into the terminal. The non-transitory computer storage medium stores one or more programs that, when executed by a device, cause the device to: generating sequence characteristics for each statement to be pushed in a statement set to be pushed obtained in advance by using a depth sequence model trained in advance; using the sequence characteristics of the sentences to be pushed and the context information between the words contained in the sentences to be pushed as the input characteristics of a pre-trained conditional random field model for sequence labeling, thereby obtaining the identification information in the sentences to be pushed; and matching the obtained identification information of each statement to be pushed with the identification information in the identification information set of the user obtained in advance, and selecting the statement to be pushed from the statement set to be pushed according to the matching result and pushing the statement to be pushed to the terminal.
The above description is only a preferred embodiment of the application and is illustrative of the principles of the technology employed. It will be appreciated by a person skilled in the art that the scope of the invention according to the present application is not limited to the specific combination of the above-mentioned features, but also covers other embodiments where any combination of the above-mentioned features or their equivalents is made without departing from the inventive concept. For example, the above features may be replaced with (but not limited to) features having similar functions disclosed in the present application.
Claims (14)
1. A method for pushing information, the method comprising:
generating sequence characteristics for each sentence to be pushed in a sentence set to be pushed obtained in advance by using a depth sequence model trained in advance;
using sequence features of sentences to be pushed and context information among the words contained in the sentences to be pushed as input features of a pre-trained conditional random field model for sequence labeling, so as to obtain identification information in the sentences to be pushed, wherein the context information among the words contained in the sentences to be pushed is used for describing the front-back relation among the words contained in the sentences to be pushed, and the identification information is used for indicating whether the words contained in the sentences to be pushed belong to at least one of brand names and company names;
and matching the obtained identification information of each statement to be pushed with the identification information in the pre-obtained identification information set of the user, and selecting the statement to be pushed from the statement set to be pushed according to the matching result and pushing the selected statement to the terminal.
2. The method of claim 1, wherein the set of identification information of the user is obtained by:
generating sequence characteristics for each statement in at least one statement preset by a user by using the depth sequence model;
and using the sequence characteristics of each sentence and the context information between words contained in each sentence as the input characteristics of the conditional random field for sequence labeling, thereby obtaining the identification information in at least one sentence preset by the user, and using the obtained identification information to form an identification information set.
3. The method according to claim 1, wherein the matching of the obtained identification information of each sentence to be pushed with the identification information in the user identification information set obtained in advance, and the selecting of the sentence to be pushed from the sentence set to be pushed according to the matching result to be pushed to the terminal comprises:
for each statement to be pushed in the statement set to be pushed, executing the following steps: comparing the obtained identification information of the sentence to be pushed with each identification information in the identification information set of the user; if the identification information of the statement to be pushed is different from the identification information in the identification information set of the user, deleting the statement to be pushed from the statement set to be pushed;
and pushing the remaining sentences to be pushed in the sentence set to the terminal.
4. The method according to claim 1, wherein the set of statements to be pushed is obtained by:
and obtaining a sentence set to be pushed according to at least one sentence preset by the user through the terminal.
5. The method according to claim 4, wherein obtaining a set of sentences to be pushed according to at least one sentence preset by the user through a terminal comprises:
and searching in a preset sentence set by using at least one sentence set by the user, and forming a sentence set to be pushed by using the searched sentences.
6. The method of claim 1, wherein the depth sequence model is trained by:
selecting sample sentences used for model training from an information base;
and training and generating a depth sequence model for generating sequence features by using the selected sample sentences.
7. The method of claim 6 wherein the conditional random field model is trained by:
generating sequence features for each training sentence in a training sentence set by using the depth sequence model, wherein each training sentence in the training sentence set is labeled in advance;
and training and generating a conditional random field model for sequence labeling by using the sequence features of the training sentences and context information between words contained in the training sentences.
8. An apparatus for pushing information, the apparatus comprising:
the generation unit is used for generating sequence characteristics for each sentence to be pushed in a sentence set to be pushed, which is obtained in advance, by using a depth sequence model trained in advance;
the system comprises a labeling unit, a judging unit and a judging unit, wherein the labeling unit is used for performing sequence labeling by using sequence characteristics of sentences to be pushed and context information among words contained in the sentences to be pushed as input characteristics of a pre-trained conditional random field model so as to obtain identification information in the sentences to be pushed, the context information among the words contained in the sentences to be pushed is used for describing the front-back relation among the words contained in the sentences to be pushed, and the identification information is used for indicating whether the words contained in the sentences to be pushed belong to at least one of brand names and company names;
and the pushing unit is used for matching the obtained identification information of each statement to be pushed with the identification information in the pre-obtained identification information set of the user, and selecting the statement to be pushed from the statement set to be pushed according to the matching result and pushing the statement to be pushed to the terminal.
9. The apparatus according to claim 8, wherein the apparatus further comprises an identification information set acquisition unit configured to:
generating sequence characteristics for each statement in at least one statement preset by a user by using the depth sequence model;
and using the sequence characteristics of each sentence and the context information between words contained in each sentence as the input characteristics of the conditional random field for sequence labeling, thereby obtaining the identification information in at least one sentence preset by the user, and using the obtained identification information to form an identification information set.
10. The apparatus of claim 8, wherein the pushing unit is further configured to:
for each statement to be pushed in the statement set to be pushed, executing the following steps: comparing the obtained identification information of the sentence to be pushed with each identification information in the identification information set of the user; if the identification information of the statement to be pushed is different from the identification information in the identification information set of the user, deleting the statement to be pushed from the statement set to be pushed;
and pushing the remaining sentences to be pushed in the sentence set to be pushed to a terminal.
11. The apparatus according to claim 8, further comprising a to-be-pushed statement set obtaining unit, wherein the to-be-pushed statement set obtaining unit is configured to:
and obtaining a sentence set to be pushed according to at least one sentence preset by the user through the terminal.
12. The apparatus of claim 11, wherein the to-be-pushed sentence set obtaining unit is further configured to:
and searching in a preset sentence set by using at least one sentence set by the user, and forming a sentence set to be pushed by using the searched sentences.
13. The apparatus of claim 8, wherein the depth sequence model is trained by:
selecting sample sentences used for model training from an information base;
and training and generating a depth sequence model for generating sequence features by using the selected sample sentences.
14. The apparatus of claim 8 wherein the conditional random field model is trained by:
generating sequence features for each training sentence in a training sentence set by using the depth sequence model, wherein each training sentence in the training sentence set is labeled in advance;
and training and generating a conditional random field model for sequence labeling by using the sequence features of each training sentence and context information among words contained in each training sentence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611206433.8A CN108241631B (en) | 2016-12-23 | 2016-12-23 | Method and device for pushing information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611206433.8A CN108241631B (en) | 2016-12-23 | 2016-12-23 | Method and device for pushing information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108241631A CN108241631A (en) | 2018-07-03 |
| CN108241631B true CN108241631B (en) | 2022-09-30 |
Family
ID=62703479
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611206433.8A Active CN108241631B (en) | 2016-12-23 | 2016-12-23 | Method and device for pushing information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108241631B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103544496A (en) * | 2012-07-12 | 2014-01-29 | 同济大学 | Method for recognizing robot scenes on basis of space and time information fusion |
| CN103988158A (en) * | 2011-12-12 | 2014-08-13 | 国际商业机器公司 | System, device, and method for processing of three-dimensional time series data |
| CN104573054A (en) * | 2015-01-21 | 2015-04-29 | 杭州朗和科技有限公司 | Information pushing method and equipment |
| CN105868193A (en) * | 2015-01-19 | 2016-08-17 | 富士通株式会社 | Device and method used to detect product relevant information in electronic text |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103778142A (en) * | 2012-10-23 | 2014-05-07 | 南开大学 | Conditional random fields (CRF) based acronym expansion explanation recognition method |
| CN103778407A (en) * | 2012-10-23 | 2014-05-07 | 南开大学 | Gesture recognition algorithm based on conditional random fields under transfer learning framework |
| US9779087B2 (en) * | 2013-12-13 | 2017-10-03 | Google Inc. | Cross-lingual discriminative learning of sequence models with posterior regularization |
| CN105404632B (en) * | 2014-09-15 | 2020-07-31 | 深港产学研基地 | System and method for carrying out serialized annotation on biomedical text based on deep neural network |
| CN104598510A (en) * | 2014-10-16 | 2015-05-06 | 苏州大学 | Event trigger word recognition method and device |
| CN104679885B (en) * | 2015-03-17 | 2018-03-30 | 北京理工大学 | A kind of user's search string organization names recognition method based on semantic feature model |
| CN105701155B (en) * | 2015-12-30 | 2019-05-31 | 百度在线网络技术（北京）有限公司 | Information-pushing method and device |
| CN105976056A (en) * | 2016-05-03 | 2016-09-28 | 成都数联铭品科技有限公司 | Information extraction system based on bidirectional RNN |
-
2016
- 2016-12-23 CN CN201611206433.8A patent/CN108241631B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103988158A (en) * | 2011-12-12 | 2014-08-13 | 国际商业机器公司 | System, device, and method for processing of three-dimensional time series data |
| CN103544496A (en) * | 2012-07-12 | 2014-01-29 | 同济大学 | Method for recognizing robot scenes on basis of space and time information fusion |
| CN105868193A (en) * | 2015-01-19 | 2016-08-17 | 富士通株式会社 | Device and method used to detect product relevant information in electronic text |
| CN104573054A (en) * | 2015-01-21 | 2015-04-29 | 杭州朗和科技有限公司 | Information pushing method and equipment |
Non-Patent Citations (3)
| Title |
|---|
| Spoken language understanding using long short-term memory neural networks;Kaisheng Yao 等;《 2014 IEEE Spoken Language Technology Workshop (SLT)》;20150802;189-194 * |
| 面向商品评价的情感要素抽取;冯仓龙 等;《沈阳航空航天大学学报》;20160630;第33卷(第6期);71-76 * |
| 面向自然语言处理的深度学习研究;奚雪峰 等;《自动化学报》;20161015;第42卷(第10期);1445-1465 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108241631A (en) | 2018-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109460513B (en) | Method and apparatus for generating click rate prediction model | |
| CN107577807B (en) | Method and device for pushing information | |
| CN107241260B (en) | News pushing method and device based on artificial intelligence | |
| CN110069698B (en) | Information pushing method and device | |
| CN106407381B (en) | A kind of method and apparatus of the pushed information based on artificial intelligence | |
| CN113592535B (en) | Advertisement recommendation method and device, electronic equipment and storage medium | |
| CN108121814B (en) | Search result ranking model generation method and device | |
| US20170287023A1 (en) | Blacklisting Based on Image Feature Analysis and Collaborative Filtering | |
| CN112925973A (en) | Data processing method and device | |
| CN112818224A (en) | Information recommendation method and device, electronic equipment and readable storage medium | |
| CN111695041B (en) | Method and device for recommending information | |
| CN108038172B (en) | Search method and device based on artificial intelligence | |
| CN110059172B (en) | Method and device for recommending answers based on natural language understanding | |
| CN110910178A (en) | Method and device for generating advertisement | |
| CN109978594B (en) | Order processing method, device and medium | |
| US20220198487A1 (en) | Method and device for processing user interaction information | |
| CN110555747A (en) | method and device for determining target user | |
| CN111125502B (en) | Method and device for generating information | |
| CN113450172A (en) | Commodity recommendation method and device | |
| CN111026849A (en) | Data processing method and device | |
| CN111782850A (en) | Object searching method and device based on hand drawing | |
| CN108241631B (en) | Method and device for pushing information | |
| CN111787042A (en) | Method and device for pushing information | |
| CN110881056A (en) | Method and device for pushing information | |
| WO2022271661A1 (en) | Training data generation, model serving, and machine learning techniques for advanced frequency management |
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 |