CN107748801A - News recommendation method, device, terminal device and computer-readable storage medium - Google Patents

Abstract

The invention provides a news recommendation method, a news recommendation device, terminal equipment and a computer readable storage medium, wherein the method comprises the following steps: acquiring news to be recommended; determining a target space to which the news to be recommended belongs by using a preset space division method; and determining first alternative news related to the news to be recommended according to the target space to which the recommended news belongs. Therefore, the first alternative news belonging to the same space with the news to be recommended is recommended to the user, the semantic correlation between the first alternative news and the news to be recommended is guaranteed, the recommendation result is more accurate, and the user experience is better.

Description

News recommendation method, device, terminal device and computer-readable storage medium

technical field

The present invention relates to the field of computer technology, in particular to a news recommendation method, device, terminal equipment and computer-readable storage medium.

Background technique

With the rapid development of information technology and the Internet, network news has become more and more popular and has become a major way for people to obtain information in their daily lives. In recent years, with the development of the mobile Internet, in addition to the well-known major websites and network self-media platforms (such as Weibo, etc.), many news content providers also use the application software that can be applied to mobile communication products ( APP) provides users with a large amount of news information.

In the existing news recommendation method, after obtaining the user's search sentence, or based on the current news, the news index is mainly carried out through the inverted index, so as to recommend the recalled news to the user. However, the inverted index method is only based on word frequency statistics, ignoring the relevance of the content contained in the rest of the recommended news to the search sentence or the current news. For example, when recalling the search phrase "Army Day", only the frequency of occurrence of "Jianjun" and "Jie" in various news will be included in the statistics, while other words related to "Army Day" such as "August 1 ", "Military Parade", "Nanchang Uprising" and other words are not included in the statistics, so that the news recommended for users has poor semantic correlation with the search sentence or current news, and the accuracy is low, and the recommendation effect and user experience are not good.

Contents of the invention

The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

Therefore, the present invention proposes a news recommendation method. By recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, thereby making the recommendation result more accurate. Accurate, better user experience.

The invention also proposes a news recommendation device.

The invention also proposes a terminal device.

The invention also proposes a computer-readable storage medium.

The embodiment of the first aspect of the present invention proposes a news recommendation method, including: obtaining news to be recommended; using a preset space division method to determine the target space to which the news to be recommended belongs; Space, determine the first candidate news associated with the news to be recommended.

In the news recommendation method of the embodiment of the present invention, after the news to be recommended is acquired, the target space to which the news to be recommended belongs is determined by using a preset space division method, so as to determine the information associated with the news to be recommended according to the target space to which the news to be recommended belongs. The first alternative news for . Therefore, by recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, so that the recommendation result is more accurate and the user experience is better.

The embodiment of the second aspect of the present invention proposes a news recommendation device, including: an acquisition module, used to acquire news to be recommended; a first determination module, used to use a preset space division method to determine the news to be recommended belongs to Target space; a second determining module, configured to determine the first candidate news associated with the news to be recommended according to the target space to which the news to be recommended belongs.

The news recommendation device of the embodiment of the present invention, after obtaining the news to be recommended, determines the target space to which the news to be recommended belongs by using the preset space division method, so as to determine the information related to the news to be recommended according to the target space to which the news to be recommended belongs. The first alternative news for . Therefore, by recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, so that the recommendation result is more accurate and the user experience is better.

The embodiment of the third aspect of the present invention proposes a terminal device, including:

A memory, a processor, and a computer program stored on the memory and operable on the processor, wherein the processor implements the news recommendation method as described in the first aspect when executing the program.

The embodiment of the fourth aspect of the present invention provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the news recommendation method as described in the first aspect is implemented.

Description of drawings

The above and/or additional aspects and advantages of the present invention will become apparent and easy to understand from the following description of the embodiments in conjunction with the accompanying drawings, wherein:

Fig. 1 is the flowchart of the news recommendation method of an embodiment of the present invention;

Fig. 2 is a flowchart of a news recommendation method according to another embodiment of the present invention;

FIG. 3 is a schematic structural diagram of a news recommendation device according to an embodiment of the present invention;

Fig. 4 is a schematic structural diagram of a news recommendation device according to another embodiment of the present invention;

Fig. 5 is a schematic structural diagram of a terminal device according to an embodiment of the present invention.

Detailed ways

Embodiments of the present invention are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference numerals designate the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

Specifically, each embodiment of the present invention aims at the existing news recommendation method. After obtaining the user's search sentence, or on the basis of the current news, the news index is mainly carried out through the inverted index, so that the recalled news is recommended For users, however, the inverted index method is only based on word frequency statistics to recall, ignoring the relevance of the content contained in the rest of the recommended news and the search sentence or current news, so that the news recommended by the user is related to the search sentence or the current news To solve the problems of poor semantic correlation, low accuracy, poor recommendation effect and user experience of news, a news recommendation method is proposed.

In the news recommendation method provided by the embodiment of the present invention, after obtaining the news to be recommended, the target space to which the news to be recommended belongs is determined by using a preset space division method, so as to determine the news to be recommended according to the target space to which the news to be recommended belongs. The associated first alternative news. Therefore, by recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, so that the recommendation result is more accurate and the user experience is better.

Fig. 1 is a flowchart of a news recommendation method according to an embodiment of the present invention.

As shown in Figure 1, the news recommendation method includes:

Step 101, obtaining news to be recommended.

Wherein, the execution body of the news recommendation method provided by the embodiment of the present invention is the news recommendation device provided by the embodiment of the present invention, and the device can be configured in any terminal device for news recommendation.

Wherein, the news to be recommended may be any news in the news library, may also be a search sentence input by the user, or may be current news.

Step 102, using a preset space division method, to determine the target space to which the news to be recommended belongs.

Wherein, the space division method may be an approximate nearest neighbor algorithm (Approximate Nearest Neighbor, ANN for short) or other methods, which are not limited in this application.

Specifically, after obtaining the news to be recommended, the target space to which the news to be recommended belongs can be determined according to the bag-of-words vector corresponding to the news to be recommended.

That is, before step 102, it may also include:

Determine the bag-of-words vector corresponding to the news to be recommended.

Wherein, the bag-of-words vector may be a bag-of-words vector corresponding to the title or abstract of the news to be recommended, or may be a bag-of-words vector corresponding to the specific content of the news to be recommended. In addition, the elements in the bag of words vector can be the frequency of occurrence of each word in the title, abstract or specific content of the news to be recommended, or the term frequency-inverse document frequency (Term Frequency-Inverse Document Frequency, TF-IDF for short) of each word ) values, etc., are not limited here.

Specifically, a dictionary can be preset according to each news in the news database, and a position or index can be set for each word in the dictionary, so that after obtaining the news to be recommended, the index number of each word can be used to determine the news to be recommended The corresponding bag-of-words vector.

For example, assuming that the elements in the bag of words vector are the frequency of occurrence of each word in the title of the recommended news, the dictionary is Dictionary={1. "Bob", 2. "likes", 3. "to", 4. "play" ,5. "basketball", 6. "also", 7. "football", 8. "games", 9. "Jim", 10. "too"}, where "1" to "10" are the words the corresponding index. When the news to be recommended is "Bob likes to play basketball, Jim likes too", since "likes" appears twice in the news to be recommended, "Bob", "to", "play", "basketball", "Jim" , "too" all appear once in the news to be recommended, and "also", "football" and "games" do not appear in the news to be recommended, then according to the index number of the dictionary, the word bag vector corresponding to the news to be recommended can be determined is [1,2,1,1,1,0,0,0,1,1].

Correspondingly, step 102 may specifically include:

According to the bag-of-words vector corresponding to the news to be recommended, the approximate nearest neighbor algorithm is used to determine the target space to which the news to be recommended belongs.

Wherein, the approximate nearest neighbor algorithm may be algorithms such as locality-sensitive hashing, random projection forest, and k-dimensional (k-dimensional tree, k-d) tree for short.

Specifically, when using the approximate nearest neighbor algorithm to determine the target space to which the news to be recommended belongs, the division conditions of the space can be set in advance, so that after the bag-of-words vector corresponding to the news to be recommended is determined, the to-be-recommended news can be determined according to the preset conditions. The target space that the recommended news belongs to.

For example, assuming that the space is divided according to whether the weight of each element in the word bag vector is greater than 5, the word bag vectors corresponding to the news A, B, C, D, E, and F in the news library are [1,8 ,6,7,0,1], [4,3,1,0,0,1], [2,1,1,1,0,0], [6,3,1,4,2,0 ], [7,6,2,0,0,1], [8,1,1,0,0,1]. Since the weight of the first element in the bag-of-words vector corresponding to news A, B, and C is less than 5, and the weight of the first element in the bag-of-words vector corresponding to news D, E, and F is greater than 5, it can be News A, B, and C are divided into space a, and news D, E, and F are divided into space b.

Further, the news in spaces a and b may be further divided into spaces. Since the weight of the second element in the bag-of-words vector corresponding to news A is greater than 5, and the weight of the second element in the bag-of-words vector corresponding to news B and C is less than 5, news A in space a can be divided into In space a1, news B and C in space a are divided into space a2. Since the weight of the second element in the bag-of-words vector corresponding to news D and F is less than 5, and the weight of the second element in the bag-of-words vector corresponding to news E is greater than 5, news D, F is divided into space b1, and news E in space b is divided into space b2.

Using the approximate nearest neighbor algorithm, when the news library contains a large number of news, it can still quickly determine the target space to which the news to be recommended belongs.

Step 103, according to the target space to which the news to be recommended belongs, determine the first candidate news associated with the news to be recommended.

Wherein, the first candidate news is the news semantically related to the news to be recommended in the news database.

It should be noted that the first candidate news may be one piece of news, or may be multiple pieces of news.

It can be understood that, through the above space division method, semantically related news can be divided into the same space. However, the degree of correlation of news in the same space may be different. In the embodiment of the present invention, the degree of correlation between any news in the same space and other news can also be determined, so as to determine the first news associated with any news according to the degree of correlation. One alternative news.

Specifically, step 103 may include:

Determine the similarity between each news in the target space and the news to be recommended;

According to each degree of similarity, the first candidate news is determined.

Among them, the similarity is used to represent the degree of correlation between each news in the target space and the news to be recommended. The greater the similarity, the greater the degree of correlation between each news and the news to be recommended, and vice versa.

Specifically, by calculating the cosine similarity and Pearson correlation coefficient between each news in the target space and the bag-of-words vectors corresponding to the news to be recommended, the similarity between each news and the news to be recommended can be determined, and based on the similarity The degree of each news in the target space is sorted, so that the news with the largest similarity is determined as the first candidate news associated with the news to be recommended.

It can be understood that, using the above method, the semantically related news in the news base can be divided into the same space, and according to the similarity between any news and other news in the same space, the first news associated with any news can be determined. One alternative news. When any news in the news library is the current news, if the user’s recommendation request is obtained, the space to which the current news belongs can be determined, and then the first alternative news associated with it can be selected from the news belonging to the same space. Recommended to users, so as to realize semantically relevant news jumps. When the search sentence input by the user is obtained, the space to which the search sentence belongs can also be determined through the above method, and the first candidate news related to the search sentence is selected from the news belonging to the same space, and recommended to the user.

Specifically, in the embodiment of the present invention, since the space is divided according to each word in each news, the news semantics in the same space are related, so from the news belonging to the same space as the news to be recommended, the selected news associated with the news to be recommended The first candidate news, the semantic relevance and accuracy of the news to be recommended are higher, and the recommendation effect is better.

In the news recommendation method of the embodiment of the present invention, after the news to be recommended is acquired, the target space to which the news to be recommended belongs is determined by using a preset space division method, so as to determine the information associated with the news to be recommended according to the target space to which the news to be recommended belongs. The first alternative news for . Therefore, by recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, so that the recommendation result is more accurate and the user experience is better.

From the above analysis, we can know that the semantically related news in the news base can be divided into the same space through the preset space division method, so that when recommending news for users, we can select from the news that belongs to the same space as the current news or search sentence. In , select alternative news related to its semantics and recommend it to the user. In practical application, the user may also wish to obtain recommended news that belongs to the same topic as the current news or search statement but has different content. The above situation will be described in detail below in conjunction with FIG. 2 .

Fig. 2 is a flowchart of a news recommendation method according to another embodiment of the present invention.

As shown in Figure 2, the news recommendation method provided by the embodiment of the present invention may further include, after obtaining the recommended news:

Step 201, according to the word frequency-reverse document frequency of the news to be recommended, determine the category to which the news to be recommended belongs.

Specifically, after obtaining the news to be recommended, the words with better category discrimination ability in the news to be recommended can be determined according to the word frequency of the news to be recommended - the reverse file frequency, so as to determine the category of the news to be recommended according to the category to which the determined words belong. kind.

Step 202, spatially divide the news belonging to the same category as the news to be recommended, and determine the news set belonging to a different space from the news to be recommended.

Step 203, according to the similarity between each news in the news collection and the news to be recommended, determine the second candidate news associated with the news to be recommended.

Wherein, the second candidate news is the news related to the news event to be recommended in the news database.

It should be noted that the second candidate news may be one piece of news, or may be multiple pieces of news.

Specifically, by spatially dividing the news that belongs to the same category as the news to be recommended, determining the news set that belongs to a different space from the news to be recommended, and determining the similarity between each news in the news set and the news to be recommended, it is possible to determine the A second alternative for news that belongs to the same category but in a different space.

During specific implementation, step 203 can be implemented in the following ways:

Determine the news in the news collection that has the greatest similarity with the news to be recommended and belongs to different spaces as the second candidate news.

Specifically, it is possible to determine the similarity between each news in the news collection and the news to be recommended, and sort the news in the news collection according to the size of the similarity, so that the news with the largest similarity and belonging to a different space from the news to be recommended The news of is determined as the second alternative news.

For the specific space division process and similarity determination process, reference may be made to the relevant descriptions in the foregoing embodiments, which will not be repeated here.

It can be understood that, through the above method, the news in the news database can be hierarchically clustered, so that the news belonging to the same topic can be divided into the same category, and by spatially dividing the news belonging to the same category, it can be divided into The news belonging to the same topic but belonging to different events is divided into different spaces, and then according to the similarity between each news and the news belonging to the same category but belonging to different spaces, the second backup associated with each news is determined. Choose news.

When any news in the news database is the current news, if a jump or recommendation request is obtained, the category of the current news can be determined according to the word frequency of the current news - the reverse file frequency, and then from the same category as the current news, However, in the news collections belonging to different spaces, according to the similarity between the current news and each news in the news collection, a second candidate news related to the current news is selected and recommended to the user. When the search sentence entered by the user is obtained, the above method can also be used to determine the category of the search sentence according to the word frequency of the search sentence - the reverse file frequency, and then from the news collection that belongs to the same category as the search sentence but belongs to a different space. , according to the similarity between the search statement and each news in the news collection, select the second candidate news associated with the search statement, and recommend it to the user.

Since the second candidate news and the news to be recommended belong to the same category, but belong to different spaces, and the news in the same category belong to the same topic, and the news in different spaces belong to different events, thus jumping from the current news or search sentence to the first 2. When selecting news, jumping to news of different events within the topic can be realized.

In the news recommendation method of the embodiment of the present invention, after the news to be recommended is obtained, the news that belongs to the same category as the news to be recommended is first space-divided, and the news set that belongs to a different space from the news to be recommended is determined, and then according to each news set in the news set, The similarity between the news and the news to be recommended determines the second candidate news associated with the news to be recommended, thus, by recommending the second candidate news that belongs to the same category as the news to be recommended and belongs to a different space to the user, it ensures The second candidate news and the news to be recommended belong to the same topic, and the excessive similarity in words and content between the two is avoided, so that the recommendation results are richer and the user experience is better.

Fig. 3 is a schematic structural diagram of a news recommendation device according to an embodiment of the present invention.

As shown in Figure 3, the news recommendation device includes:

An acquisition module 31, configured to acquire news to be recommended;

The first determination module 32 is configured to determine the target space to which the news to be recommended belongs by using a preset space division method;

The second determination module 33 is configured to determine the first candidate news associated with the news to be recommended according to the target space to which the news to be recommended belongs.

Specifically, the news recommendation device provided in the embodiment of the present invention can execute the news recommendation method provided in the embodiment of the present invention, and the device can be configured in any terminal device to generate hot news commentary articles.

Wherein, the first candidate news is news semantically related to the news to be recommended.

In a possible implementation form of the embodiment of the present application, the above-mentioned second determining module 33 is specifically used for:

Determine the similarity between each news in the target space and the news to be recommended;

According to each degree of similarity, the first candidate news is determined.

It should be noted that the foregoing explanations for the news recommendation method embodiment are also applicable to the news recommendation device of this embodiment, and will not be repeated here.

The news recommendation device of the embodiment of the present invention, after obtaining the news to be recommended, determines the target space to which the news to be recommended belongs by using the preset space division method, so as to determine the information related to the news to be recommended according to the target space to which the news to be recommended belongs. The first alternative news for . Therefore, by recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, so that the recommendation result is more accurate and the user experience is better.

Fig. 4 is a schematic structural diagram of a news recommendation device according to another embodiment of the present invention.

As shown in Figure 4, on the basis of Figure 3, the news recommendation device also includes:

The third determination module 41 is configured to determine the bag-of-words vector corresponding to the news to be recommended.

The fourth determination module 42 is configured to determine the category of the news to be recommended according to the word frequency-reverse document frequency of the news to be recommended.

The fifth determining module 43 is configured to space-divide the news belonging to the same category as the news to be recommended, and determine a news set belonging to a different space from the news to be recommended.

The sixth determination module 44 is configured to determine the second candidate news associated with the news to be recommended according to the similarity between each news in the news collection and the news to be recommended.

Wherein, the second candidate news is news related to the news event to be recommended.

In a possible implementation form of the embodiment of the present application, the above-mentioned first determination module 32 is specifically used for:

According to the bag-of-words vector corresponding to the news to be recommended, the approximate nearest neighbor algorithm is used to determine the target space to which the news to be recommended belongs.

In another possible implementation form of the embodiment of the present application, the above-mentioned sixth determination module 44 is specifically used for:

Determine the news in the news collection that has the greatest similarity with the news to be recommended and belongs to different spaces as the second candidate news.

It should be noted that the foregoing explanations for the news recommendation method embodiment are also applicable to the news recommendation device of this embodiment, and will not be repeated here.

The news recommendation device of the embodiment of the present invention, after obtaining the news to be recommended, determines the target space to which the news to be recommended belongs by using the preset space division method, so as to determine the information related to the news to be recommended according to the target space to which the news to be recommended belongs. The first alternative news for . Therefore, by recommending the first candidate news belonging to the same space as the news to be recommended to the user, it is ensured that the first candidate news is semantically related to the news to be recommended, so that the recommendation result is more accurate and the user experience is better.

FIG. 5 is a schematic structural diagram of a terminal device provided by an embodiment of the present invention.

As shown in Figure 5, the terminal equipment includes:

A memory 51 , a processor 52 , and a computer program stored in the memory 51 and executable on the processor 52 .

The processor 52 implements the news recommendation method provided in the above-mentioned embodiments when executing the program.

Wherein, the terminal device may be a computer, a mobile phone, a wearable device, and the like.

Further, the terminal equipment also includes:

The communication interface 53 is used for communication between the memory 51 and the processor 52 .

The memory 51 is used to store computer programs that can run on the processor 52 .

The memory 51 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one magnetic disk memory.

The processor 52 is configured to implement the news recommendation method described in the above embodiment when executing the program.

If the memory 51 , the processor 52 and the communication interface 53 are implemented independently, the communication interface 53 , the memory 51 and the processor 52 may be connected to each other through a bus to complete mutual communication. The bus may be an Industry Standard Architecture (Industry Standard Architecture, ISA for short) bus, a Peripheral Component Interconnect (PCI for short) bus, or an Extended Industry Standard Architecture (EISA for short) bus. The bus can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is shown in FIG. 5 , but it does not mean that there is only one bus or one type of bus.

Optionally, in specific implementation, if the memory 51, the processor 52 and the communication interface 53 are integrated on one chip, the memory 51, the processor 52 and the communication interface 53 can communicate with each other through the internal interface.

Processor 52 may be a central processing unit (Central Processing Unit, referred to as CPU), or a specific integrated circuit (Application Specific Integrated Circuit, referred to as ASIC), or is configured to implement one or more integrated circuits of the embodiments of the present invention .

The embodiment of the fourth aspect of the present invention provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the news recommendation method in the foregoing embodiments is implemented.

The embodiment of the fifth aspect of the present invention provides a computer program product. When the instructions in the computer program product are executed by a processor, the news recommendation method in the foregoing embodiments is executed.

In the description of this specification, descriptions referring to the terms "one embodiment", "some embodiments", "example", "specific examples", or "some examples" mean that specific features described in connection with the embodiment or example , structure, material or characteristic is included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the described specific features, structures, materials or characteristics may be combined in any suitable manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples and features of different embodiments or examples described in this specification without conflicting with each other.

In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include at least one of these features. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

Any process or method descriptions in flowcharts or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing custom logical functions or steps of a process , and the scope of preferred embodiments of the invention includes alternative implementations in which functions may be performed out of the order shown or discussed, including substantially concurrently or in reverse order depending on the functions involved, which shall It is understood by those skilled in the art to which the embodiments of the present invention pertain.

The logic and/or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced listing of executable instructions for implementing logical functions, can be embodied in any computer-readable medium, For use with instruction execution systems, devices, or devices (such as computer-based systems, systems including processors, or other systems that can fetch instructions from instruction execution systems, devices, or devices and execute instructions), or in conjunction with these instruction execution systems, devices or equipment used. For the purposes of this specification, a "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use in or in conjunction with an instruction execution system, device or device. More specific examples (non-exhaustive list) of computer-readable media include the following: electrical connection with one or more wires (electronic device), portable computer disk case (magnetic device), random access memory (RAM), Read Only Memory (ROM), Erasable and Editable Read Only Memory (EPROM or Flash Memory), Fiber Optic Devices, and Portable Compact Disc Read Only Memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program can be printed, since the program can be read, for example, by optically scanning the paper or other medium, followed by editing, interpretation or other suitable processing if necessary. The program is processed electronically and stored in computer memory.

It should be understood that various parts of the present invention can be realized by hardware, software, firmware or their combination. In the embodiments described above, various steps or methods may be implemented by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented by any one or combination of the following techniques known in the art: Discrete logic circuits, ASICs with suitable combinational logic gates, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

Those of ordinary skill in the art can understand that all or part of the steps carried by the methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. During execution, one or a combination of the steps of the method embodiments is included.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing module, each unit may exist separately physically, or two or more units may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware or in the form of software function modules. If the integrated modules are realized in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium.

The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, and the like. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention, those skilled in the art can make the above-mentioned The embodiments are subject to changes, modifications, substitutions and variations.

Claims (14)

1. a kind of news recommends method, it is characterised in that including：

Obtain news to be recommended；

Using default space-division method, the object space belonging to the news to be recommended is determined；

According to the object space belonging to the news to be recommended, it is determined that the first alternative news associated with the news to be recommended.

2. the method as described in claim 1, it is characterised in that the object space determined belonging to the news to be recommended it Before, in addition to：

Determine bag of words vector corresponding to the news to be recommended；

The object space determined belonging to the news to be recommended, including：

According to bag of words vector corresponding to the news to be recommended, using approximate KNN algorithm, the news institute to be recommended is determined The object space of category.

3. the method as described in claim 1, it is characterised in that it is described determine to associate with the news to be recommended it is first alternative News, including：

Determine each similarity between each news and the news to be recommended in the object space；

According to each similarity, the described first alternative news is determined.

4. the method as described in claim 1-3 is any, it is characterised in that after the acquisition news to be recommended, in addition to：

According to the word frequency of the news to be recommended-reverse document-frequency, the class belonging to the news to be recommended is determined；

Pair belonging to mutually similar news with the news to be recommended carries out space division, it is determined that belonging to not with the news to be recommended The news agregator of isospace；

According to the similarity of each news in the news agregator and the news to be recommended, it is determined that the with the association to be recommended Two alternative news.

5. method as claimed in claim 4, it is characterised in that the determination is alternative new with the second of the association to be recommended Hear, including：

Determine maximum with the news similarity to be recommended in the news agregator and belong to the news of different spaces for each second Alternative news.

6. method as claimed in claim 4, it is characterised in that the first alternative news is semantic with the news to be recommended Related news, the second alternative news is the news related to the media event to be recommended.

A kind of 7. news recommendation apparatus, it is characterised in that including：

Acquisition module, for obtaining news to be recommended；

First determining module, for utilizing default space-division method, determine the object space belonging to the news to be recommended；

Second determining module, for the object space according to belonging to the news to be recommended, it is determined that being closed with the news to be recommended First alternative news of connection.

8. device as claimed in claim 7, it is characterised in that also include：

3rd determining module, for determining bag of words vector corresponding to the news to be recommended；

First determining module, is specifically used for：

According to bag of words vector corresponding to the news to be recommended, using approximate KNN algorithm, the news institute to be recommended is determined The object space of category.

9. device as claimed in claim 7, it is characterised in that second determining module, be specifically used for：

Determine each similarity between each news and the news to be recommended in the object space；

According to each similarity, the described first alternative news is determined.

10. the device as described in claim 7-9 is any, it is characterised in that also include：

4th determining module, for the word frequency according to the news to be recommended-reverse document-frequency, determine the news to be recommended Affiliated class；

5th determining module, belong to mutually similar news progress space division with the news to be recommended for, it is determined that and institute State the news agregator that news to be recommended belongs to different spaces；

6th determining module, for the similarity according to each news in the news agregator and the news to be recommended, it is determined that with Second alternative news of the association to be recommended.

11. device as claimed in claim 10, it is characterised in that the 6th determining module, be specifically used for：

Determine maximum with the news similarity to be recommended in the news agregator and belong to the news of different spaces for each second Alternative news.

12. device as claimed in claim 10, it is characterised in that the first alternative news be and the news language to be recommended Adopted related news, the second alternative news is the news related to the media event to be recommended.

13. a kind of terminal device, including：

Memory, processor and the computer program that can be run on the memory and on the processor is stored in, it is special Sign is, the news recommendation method as described in any in claim 1-6 is realized during the computing device described program.

14. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that described program is processed Realize that the news as described in any in claim 1-6 recommends method when device performs.