WO2016107326A1 - 一种基于搜索词进行搜索推荐的方法和装置 - Google Patents
一种基于搜索词进行搜索推荐的方法和装置 Download PDFInfo
- Publication number
- WO2016107326A1 WO2016107326A1 PCT/CN2015/095016 CN2015095016W WO2016107326A1 WO 2016107326 A1 WO2016107326 A1 WO 2016107326A1 CN 2015095016 W CN2015095016 W CN 2015095016W WO 2016107326 A1 WO2016107326 A1 WO 2016107326A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- search
- recommendation
- anonymous
- network topology
- search term
- 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.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Definitions
- the present invention relates to the field of Internet technologies, and in particular, to a method and apparatus for performing search recommendation based on search terms.
- search recommendation technology came into being, search recommendation technology The point is to guide users to find the information and information they need more quickly and accurately.
- the existing search recommendation descriptions are generally referred to as "related xx", where xx is a category with no information, such as "person”, “movie”, etc., and the recommendation results of different categories are mixed together; for example, at present There is a search engine searching for "Transformers”, and the search engine will appear to recommend "related characters", which include both movie actors and comic characters.
- the recommendation results of each dimension are mixed together, making it difficult for users to distinguish and find, and the user experience is more difference.
- the existing search recommendations are mostly based on Collaborative Filtering or Association Rules, which is characterized by a small number of results and insufficient focus, which leads to the recommendation results and user intent.
- the search engine will appear "relevant attractions” recommendation, basically some of the national hot spots with little relationship with "Jiuzhaigou”, such search recommendation scheme does not meet the user's search needs, It makes little sense to the user.
- the present invention has been made in order to provide a method and apparatus for performing search recommendation based on search terms that overcomes the above problems or at least partially solves the above problems.
- a method for performing search recommendation based on a search term comprising:
- a recommendation result associated with the search term is selected from the anonymous behavioral network topology.
- an apparatus for performing search recommendation based on a search term comprising:
- a building unit adapted to construct an anonymous behavior network topology based on anonymous search logs
- the recommendation unit is adapted to select a recommendation result related to the search term from the anonymous behavior network topology in response to an event of the user inputting the search term.
- a computer program comprising computer readable code, when said computer readable code is run on a computing device, causing said computing device to perform said search term based on said Search for recommended methods.
- a computer readable medium wherein the computer program described above is stored.
- the technical solution provided by the present invention completes the process of searching and recommending based on the search term by constructing an anonymous behavior network topology and selecting a recommendation result related to the search term input by the user. Because the anonymous behavior network topology reflects the behavior rules of the user in the statistical sense, the selected recommendation results are more in line with the user's search habits and search requirements.
- FIG. 1 shows a flow chart of a method for performing search recommendations based on search terms, in accordance with one embodiment of the present invention
- FIG. 2 shows a schematic diagram of an apparatus for performing search recommendations based on search terms, in accordance with one embodiment of the present invention
- FIG. 3 is a schematic diagram of another apparatus for performing search recommendation based on search terms according to another embodiment of the present invention.
- FIG. 4A shows a schematic diagram of a search result page in accordance with one embodiment of the present invention
- FIG. 4B shows a schematic diagram of a search result page in accordance with another embodiment of the present invention.
- Figure 5 shows schematically a block diagram of a computing device for performing the method according to the invention
- Fig. 6 schematically shows a storage unit for holding or carrying program code implementing the method according to the invention.
- FIG. 1 shows a flow chart of a method for performing search recommendations based on search terms, in accordance with one embodiment of the present invention. As shown in Figure 1, the method includes:
- Step S110 constructing an anonymous behavior network topology according to the anonymous search log.
- an anonymous search log is used as a data source, and different methods can be used to construct an anonymous behavior network topology.
- the anonymous behavioral network topology built is capable of correlating the common interests of most users.
- step S120 in response to the user inputting the event of the search term, the recommendation result related to the search term is selected from the anonymous behavior network topology.
- the method shown in FIG. 1 completes the process of searching and recommending based on the search term by constructing an anonymous behavior network topology and selecting a recommendation result related to the search term input by the user. Because the anonymous behavior network topology reflects the behavior rules of the user in the statistical sense, the selected recommendation results are more in line with the user's search habits and search requirements.
- the method shown in FIG. 1 further includes: clustering the recommendation results to obtain a plurality of classes, and sorting the plurality of classes.
- the above clustering of the recommendation results means that the recommendation results with the same dimension are clustered together; sorting the multiple classes means: by using some rules, such as correlation or search heat, benchmarking multiple classes Sort.
- the search term entered by the user is “Transformers”
- the recommended results selected from the anonymous behavior network topology include: “Megatron”, “Falling King Kong”, “Megen Fox”, “The Milky Way of Transformers” Force”, “Rickel Taylor”, “Transformers” “Carman's War”, etc., among them, “Megatron” and “Falling King Kong” are the characters of Transformers, “Megen Fox” and “Rickel Taylor” are all participating in the “Transformers” movie.
- the method shown in FIG. 1 further includes: filtering the recommendation result, filtering out the ambiguous and belonging to the spam content. Recommended results, get filtered recommendations. Then, the filtered recommendation results are clustered to obtain multiple classes, and multiple classes are sorted.
- the method further comprises: selecting an appropriate description for each class as the name of the class according to the knowledge map.
- the knowledge map focuses on exploring the attributes of these search words and the connection between them, and connecting different search words with the same search content, so knowledge The map reflects the complete body of knowledge of a content and the appropriate classification and name.
- constructing an anonymous behavior network topology according to an anonymous search log includes: using an anonymous user and a search term as a node according to content in the anonymous search log, and Click on the behavior side to build an anonymous behavior network topology.
- the search term "Transformers” is found by the click behavior of the A user, and the B user also clicks and searches for "Transformers”, then the user can be associated with the B user, and the B user also clicks and searches for the "Transformers of the Autobot Wars”. , then you can link to the Transformers related science fiction, and further can be linked to other users who clicked on the search for "Transformers of the Autobots", and so on.
- selecting a recommendation result related to the search term from an anonymous behavior network topology includes: performing randomization in an anonymous behavior network topology according to a random walk algorithm Walk around and select the predetermined number of recommendations that are most relevant to the search term; or, according to Pagerank, Personalized Pagerank, Random Walk with The Restart, or Metapath algorithm selects a predetermined number of recommendation results that are most relevant to the search term from the anonymous behavior network topology.
- the random walk algorithm and the Pagerank, Personalized Pagerank, Random Walk with Restart, and Metapath algorithms are all prior art, and are not specifically described herein.
- the method shown in FIG. 1 further includes: embedding the recommended result obtained by the final processing into the search result page for output.
- FIG. 4A is a schematic diagram of a search result page according to an embodiment of the present invention.
- the search result pages are sorted according to relevance and search popularity, showing the above.
- the three types of recommendations : Transformers characters, "Transformers” starring and science fiction films. Different types of recommendation results are divided and meaningful descriptions are made to meet the search needs of different users.
- FIG. 4B shows a schematic diagram of a search result page in accordance with another embodiment of the present invention.
- a predetermined number of recommendation results related to "Jiuzhaigou” are selected from the anonymous behavior network topology, and the recommendation results are filtered and clustered to obtain four types of recommendation results: relevant Sichuan attractions Relevant Sichuan cities; related Yunnan attractions and Yunnan cities; national tourist attractions. Sort the categories based on relevance, prioritize the relevant Sichuan attractions, then recommend a slightly divergent Sichuan city, followed by a more divergent nearby tourist city in Yunnan, and finally some other tourist attractions.
- the knowledge map select an appropriate description for each class as the name of the class, namely: Sichuan attractions, Sichuan cities, Yunnan administrative divisions and tourist attractions, and embed the four types of recommendation results finally processed into the search results page for output.
- the above-mentioned recommendation results may be arranged only on one side of the search result page (such as the right side) according to the category, and the other side of the page is arranged from top to bottom as a common search result. item.
- the apparatus 200 for performing search recommendation based on a search term includes:
- the building unit 210 is adapted to construct an anonymous behavior network topology according to the anonymous search log.
- the recommendation unit 220 is adapted to select a recommendation result related to the search term from the anonymous behavior network topology in response to the user inputting the event of the search term.
- the device shown in FIG. 2 constructs an anonymous behavior network topology by mutual cooperation of each unit, selects a recommendation result related to a search term input by the user, and completes a search based on the search term. Recommended process. Because the anonymous behavior network topology reflects the behavior rules of the user in the statistical sense, the selected recommendation results are more in line with the user's search habits and search requirements.
- the constructing unit 210 of the apparatus shown in FIG. 2 is adapted to construct an anonymous behavior network by using an anonymous user and a search term as nodes according to the content in the anonymous search log. Topology.
- the recommending unit 220 of the apparatus shown in FIG. 2 is adapted to randomly walk in an anonymous behavior network topology according to a random walk algorithm, and select a predetermined number of recommendation results most relevant to the search term. Or, suitable for selecting a predetermined number of recommendation results most relevant to the search term from the anonymous behavior network topology according to Pagerank, Personalized Pagerank, Random Walk with Restart, or Metapath algorithm.
- FIG. 3 illustrates a schematic diagram of another apparatus for performing search recommendations based on search terms, in accordance with another embodiment of the present invention.
- the apparatus 300 for performing search recommendation based on a search term includes: a construction unit 310, a recommendation unit 320, a clustering unit 330, a filtering unit 340, a description adding unit 350, and an output unit 360.
- the configuration unit 310 and the recommendation unit 320 are respectively the same as the construction unit 210 and the recommendation unit 220 of the device shown in FIG. 2, and details are not described herein again.
- the clustering unit 330 is adapted to cluster the recommendation results obtained by the recommendation unit 320 to obtain a plurality of classes, and sort the plurality of classes.
- the filtering unit 340 is adapted to filter the recommendation result obtained by the recommendation unit 320, filter out the ambiguous and recommendation results belonging to the spam content, and obtain filtering.
- the subsequent recommendation result is output to the clustering unit 330.
- the description adding unit 350 is adapted to select an appropriate description for each class clustered by the clustering unit 330 as the name of the class according to the knowledge map.
- the output unit 360 is adapted to embed the recommended result obtained by the final processing into the search result page for output.
- the technical solution provided by the present invention expands on the basis of user behavior, constructs an anonymous behavior network topology, and selects a recommendation result related to a search term input by a user by random walk or other algorithm, and completes the search term based on the search term.
- Anonymous behavior network extension In the statistical sense, it reflects the user's behavioral rules for searching, and deeply mines the few dimensions that users pay most attention to, overcomes the shortcomings of the recommendation in the prior art, and uses clustering to divide the recommendation results. Each class is accurately described by the knowledge map, and valuable information is provided, so that the selected recommendation results are more in line with the user's search habits and search requirements.
- modules in the devices of the embodiments can be adaptively changed and placed in one or more devices different from the embodiment.
- the modules or units or components of the embodiments may be combined into one module or unit or component, and further they may be divided into a plurality of sub-modules or sub-units or sub-components.
- any combination of the features disclosed in the specification, including the accompanying claims, the abstract and the drawings, and any methods so disclosed, or All processes or units of the device are combined.
- Each feature disclosed in this specification (including the accompanying claims, the abstract and the drawings) may be replaced by alternative features that provide the same, equivalent or similar purpose.
- the various component embodiments of the present invention may be implemented in hardware, or in a software module running on one or more processors, or in a combination thereof.
- a microprocessor or digital signal processor may be used in practice to implement some of some or all of the means for searching for recommendations based on search terms in accordance with embodiments of the present invention. Or all features.
- the invention can also be implemented as a device or device program (e.g., a computer program and a computer program product) for performing some or all of the methods described herein.
- Such a program implementing the invention may be stored on a computer readable medium or may be in the form of one or more signals. Such signals may be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.
- Figure 5 schematically illustrates a block diagram of a computing device for performing the method in accordance with the present invention.
- the computing device conventionally includes a processor 510 and a computer program product or computer readable medium in the form of a memory 520.
- the memory 520 may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), an EPROM, a hard disk, or a ROM.
- Memory 520 has a memory space 530 for program code 531 for performing any of the method steps described above.
- storage space 530 for program code may include various program code 531 for implementing various steps in the above methods, respectively.
- the program code can be read from or written to one or more computer program products.
- These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks. Such computer program products are typically portable or fixed storage units as described with reference to FIG.
- the storage unit may have storage segments, storage spaces, and the like that are similarly arranged to memory 520 in the computing device of FIG.
- the program code can be compressed, for example, in an appropriate form.
- the storage unit comprises computer readable code 531 ' for performing the steps of the method according to the invention, ie code that can be read by a processor such as 510, which when executed by the computing device causes the calculation The device performs the various steps in the methods described above.
- the present invention is applicable to computer systems/servers that can operate with numerous other general purpose or special purpose computing system environments or configurations.
- Examples of well-known computing systems, environments, and/or configurations suitable for use with computer systems/servers include, but are not limited to, personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, based on Microprocessor systems, set-top boxes, programmable consumer electronics, networked personal computers, small computer systems, mainframe computer systems, and distributed cloud computing technology environments including any of the above, and the like.
- the computer system/server can be described in the general context of computer system executable instructions (such as program modules) being executed by a computer system.
- program modules may include routines, programs, target programs, components, logic, data structures, and the like that perform particular tasks or implement particular abstract data types.
- the computer system/server can be implemented in a distributed cloud computing environment where tasks are performed by remote processing devices that are linked through a communication network.
- program modules may be located on a local or remote computing system storage medium including storage devices.
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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
本发明公开了一种基于搜索词进行搜索推荐的方法和装置,该方法包括:根据匿名搜索日志构建匿名行为网络拓扑;响应于用户输入搜索词的事件,从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果。本发明提供的技术方案在用户行为基础上进行扩展,将用户最关注的少数维度进行深入挖掘,克服了现有技术中推荐结果过于泛化的缺点,并利用聚类对推荐结果进行划分,并通过知识图谱对每个类进行准确的描述,提供有价值的信息,使得选取出的推荐结果更加符合用户的搜索习惯和搜索需求。
Description
本发明涉及互联网技术领域,具体涉及一种基于搜索词进行搜索推荐的方法和装置。
随着Web技术的不断发展,互联网信息的创建和分享变得越来越容易,信息的极度爆炸使得人们对于需要的信息的寻找变得越来越难,搜索推荐技术应运而生,搜索推荐技术的意义在于引导用户更快更准确的找到所需要的信息和资讯。
现有搜索推荐描述,一般都以“相关xx”其中xx为一些没有信息量的类别,如“人物”、“影片”等等,而且不同的类别的推荐结果都混合在一起;例如,在现有搜索引擎中搜索“变形金刚”,搜索引擎将出现推荐“相关人物”,其中既有电影演员,又有漫画角色,各维度的推荐结果混杂在一起,使得用户难以区分和查找,用户体验较差。
不仅如此,现有搜索推荐多基于Collaborative Filtering或Association Rules,特点是结果数量较少,且不够专注,导致推荐结果和用户意图不符。例如在现有搜索引擎中搜索“九寨沟”,搜索引擎将出现“相关景点”推荐,基本都是和“九寨沟”关系不大的一些全国热门景点,这样的搜索推荐方案不符合用户的搜索需求,对用户来说几乎没有意义。
发明内容
鉴于上述问题,提出了本发明以便提供一种克服上述问题或者至少部分地解决上述问题的一种基于搜索词进行搜索推荐的方法和装置。
依据本发明的一个方面,提供了一种基于搜索词进行搜索推荐的方法,该方法包括:
根据匿名搜索日志构建匿名行为网络拓扑;
响应于用户输入搜索词的事件,从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果。
依据本发明的另一个方面,提供了一种基于搜索词进行搜索推荐的装置,该装置包括:
构建单元,适于根据匿名搜索日志构建匿名行为网络拓扑;
推荐单元,适于响应于用户输入搜索词的事件,从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果。
根据本发明的又一个方面,提出了一种计算机程序,包括计算机可读代码,当所述计算机可读代码在计算设备上运行时,导致所述计算设备执行上文所述的基于搜索词进行搜索推荐的方法。
根据本发明的再一个方面,提出了一种计算机可读介质,其中存储了上述的计算机程序。
由上述可知,本发明提供的技术方案通过构建匿名行为网络拓扑,选取出与用户输入的搜索词相关的推荐结果,完成了基于搜索词进行搜索推荐的过程。由于匿名行为网络拓扑在统计意义上反映了用户进行搜索的行为规则,使得选取出的推荐结果更加符合用户的搜索习惯和搜索需求。
上述说明仅是本发明技术方案的概述,为了能够更清楚了解本发明的技术手段,而可依照说明书的内容予以实施,并且为了让本发明的上述和其它目的、特征和优点能够更明显易懂,以下特举本发明的具体实施方式。
通过阅读下文优选实施方式的详细描述,各种其他的优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式的目的,而并不认为是对本发明的限制。而且在整个附图中,用相同的参考符号表示相同的部件。在附图中:
图1示出了根据本发明一个实施例的一种基于搜索词进行搜索推荐的方法的流程图;
图2示出了根据本发明一个实施例的一种基于搜索词进行搜索推荐的装置的示意图;
图3示出了根据本发明另一个实施例的另一种基于搜索词进行搜索推荐的装置的示意图;
图4A示出了根据本发明一个实施例的搜索结果页面的示意图;
图4B示出了根据本发明另一个实施例的搜索结果页面的示意图;
图5示意性地示出了用于执行根据本发明的方法的计算设备的框图;以及
图6示意性地示出了用于保持或者携带实现根据本发明的方法的程序代码的存储单元。
下面将参照附图更详细地描述本公开的示例性实施例。虽然附图中显示了本公开的示例性实施例,然而应当理解,可以以各种形式实现本公开而不应被这里阐述的实施例所限制。相反,提供这些实施例是为了能够更透彻地理解本公开,并且能够将本公开的范围完整的传达给本领域的技术人员。
图1示出了根据本发明一个实施例的一种基于搜索词进行搜索推荐的方法的流程图。如图1所示,该方法包括:
步骤S110,根据匿名搜索日志构建匿名行为网络拓扑。
在本发明的实施例中的以匿名搜索日志为数据源,可以采用不同的方法来构建匿名行为网络拓扑。所构建的匿名行为网络拓扑在能够关联大多数用户的共同兴趣。
步骤S120,响应于用户输入搜索词的事件,从匿名行为网络拓扑中选取出与搜索词相关的推荐结果。
可见,图1所示的方法通过构建匿名行为网络拓扑,选取出与用户输入的搜索词相关的推荐结果,完成了基于搜索词进行搜索推荐的过程。由于匿名行为网络拓扑在统计意义上反映了用户进行搜索的行为规则,使得选取出的推荐结果更加符合用户的搜索习惯和搜索需求。
在本发明的一个实施例中,图1所示的方法进一步包括:对推荐结果进行聚类得到多个类,对多个类进行排序。
上述对推荐结果进行聚类是指:将具有相同维度的推荐结果聚类到一起;对多个类进行排序是指:通过一些规则,如以相关度或搜索热度为基准,对多个类进行排序。例如,用户输入的搜索词为“变形金刚”,从匿名行为网络拓扑中选取出的推荐结果包括:“威震天”、“堕落金刚”、“梅根·福克斯”、“变形金刚之银河之力”、“瑞切尔·泰勒”、“变形金刚之汽
车人战记”等,其中,“威震天”、“堕落金刚”均为变形金刚的角色,“梅根·福克斯”、“瑞切尔·泰勒”均为参演《变形金刚》电影的主演,“变形金刚之银河之力”、“变形金刚之汽车人战记”均为变形金刚相关的科幻片,因此,对上述推荐结果聚类得到三个类:变形金刚的角色、主演《变形金刚》的演员和变形金刚相关的科幻片。以相关度和搜索热度为基准,对这三个排序。
为了保证推荐结果的纯净性,在本发明的一个实施例中,在对推荐结果进行聚类之前,图1所示的方法进一步包括:对推荐结果进行过滤,过滤掉歧义的和属于垃圾内容的推荐结果,得到过滤后的推荐结果。再对过滤后的推荐结果进行聚类得到多个类,对多个类进行排序。
进一步地,在本发明的一个实施例中,在对多个类进行排序后,该方法进一步包括:根据知识图谱,为每个类选择一个恰当的描述作为类的名称。
不同用户在对同一种内容进行搜索时,输入的搜索词各不相同,知识图谱专注于探索这些搜索词的属性及彼此之间的连接,将不同搜索词与同一个搜索内容连接起来,因此知识图谱中体现了一个内容的完整知识体系和恰当的分类及名称。
在上述用户搜索“变形金刚”的例子中,根据知识图谱,为三个类各选择一个恰当的描述作为类的名称,分别为:变形金刚角色,《变形金刚》的主演和科幻片。
依据本发明的一个实施例,图1所示方法的步骤S110中,根据匿名搜索日志构建匿名行为网络拓扑包括:根据匿名搜索日志中的内容,以匿名的用户和搜索词为节点,以用户的点击行为边,构建匿名行为网络拓扑。这样通过A用户对的点击行为找到搜索词“变形金刚”,而B用户也点击搜索过“变形金刚”,那么可以关联到B用户,B用户还点击搜索过“变形金刚之汽车人战记”,那么可以关联到该变形金刚相关科幻小说,再进一步可以关联到点击搜索过“变形金刚之汽车人战记”的其他用户,以此类推。
依据本发明的一个实施例,图1所示方法的步骤S120中,从匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果包括:根据随机游走算法在匿名行为网络拓扑中进行随机游走,选取出与搜索词最相关的预定数量的推荐结果;或者,根据Pagerank、Personalized Pagerank、Random Walk with
Restart、或Metapath算法从匿名行为网络拓扑中选取出与搜索词最相关的预定数量的推荐结果。这里,随机游走算法以及Pagerank、Personalized Pagerank、Random Walk with Restart和Metapath算法均为现有技术,这里不再具体进行描述。
基于上述各实施例,图1所示的方法进一步包括:将最终处理得到的推荐结果嵌入搜索结果页面中输出。
图4A示出了根据本发明一个实施例的搜索结果页面的示意图,如图4A所示,当用户搜索“变形金刚”时,搜索结果页面按照相关度和搜索热度排序,示出了上文中得到的三类推荐结果:变形金刚角色,《变形金刚》的主演和科幻片。将不同类的推荐结果划分开,并进行有意义的描述,符合不同用户的搜索需求。
图4B示出了根据本发明另一个实施例的搜索结果页面的示意图。当用户搜索“九寨沟”时,根据一定算法从匿名行为网络拓扑中选取出与“九寨沟”相关的预定数量的推荐结果,对推荐结果进行过滤和聚类,得到四类推荐结果:相关的四川景点;相关的四川城市;相关的云南景点和云南城市;全国范围内的旅游景点。基于相关性对类进行排序,优先推荐相关的四川景点,然后推荐稍微发散一些的四川城市,接下来是更发散的附近的云南的旅游城市,最后是一些其他旅游景点。根据知识图谱,为每个类选择一个恰当的描述作为类的名称,分别为:四川景点、四川城市、云南行政区划和旅游景点,将最终处理得到的四类推荐结果嵌入搜索结果页面中输出,如图4B所示。另外除上述图4A、4B所示的方式外,上述推荐结果还可以按照类别仅排布在搜索结果页的一侧(比如右侧),页中另一侧由上到下排列是普通搜索结果项。
图2示出了根据本发明一个实施例的一种基于搜索词进行搜索推荐的装置的示意图。如图2所示,该基于搜索词进行搜索推荐的装置200包括:
构建单元210,适于根据匿名搜索日志构建匿名行为网络拓扑。
推荐单元220,适于响应于用户输入搜索词的事件,从匿名行为网络拓扑中选取出与搜索词相关的推荐结果。
可见,图2所示的装置通过各单元的相互配合,构建匿名行为网络拓扑,选取出与用户输入的搜索词相关的推荐结果,完成了基于搜索词进行搜索推
荐的过程。由于匿名行为网络拓扑在统计意义上反映了用户进行搜索的行为规则,使得选取出的推荐结果更加符合用户的搜索习惯和搜索需求。
在本发明的一个实施例中,图2所示装置的构建单元210,适于根据匿名搜索日志中的内容,以匿名的用户和搜索词为节点,以用户的点击行为边,构建匿名行为网络拓扑。
在本发明的一个实施例中,图2所示装置的推荐单元220,适于根据随机游走算法在匿名行为网络拓扑中进行随机游走,选取出与搜索词最相关的预定数量的推荐结果;或者,适于根据Pagerank、Personalized Pagerank、Random Walk with Restart、或Metapath算法从匿名行为网络拓扑中选取出与搜索词最相关的预定数量的推荐结果。
图3示出了根据本发明另一个实施例的另一种基于搜索词进行搜索推荐的装置的示意图。如图3所示,该基于搜索词进行搜索推荐的装置300包括:构建单元310、推荐单元320、聚类单元330、过滤单元340、描述添加单元350和输出单元360。
其中,构建单元310、推荐单元320分别与图2所示装置的构建单元210、推荐单元220对应相同,在此不再赘述。
聚类单元330,适于对推荐单元320得到的推荐结果进行聚类得到多个类,对多个类进行排序。
在本发明的一个实施例中,为了保证推荐结果的纯净性,过滤单元340,适于对推荐单元320得到的所述推荐结果进行过滤,过滤掉歧义的和属于垃圾内容的推荐结果,得到过滤后的推荐结果后输出给聚类单元330。
描述添加单元350,适于根据知识图谱,为聚类单元330聚类得到的每个类选择一个恰当的描述作为类的名称。
输出单元360,适于将最终处理得到的推荐结果嵌入搜索结果页面中输出。
图2与图3所示装置所执行的过程,上文中以用户搜索“变形金刚”的情况和用户搜索“九寨沟”的情况为例,已进行详细说明,在此不再赘述。
综上所述,本发明提供的技术方案在用户行为基础上进行扩展,构建匿名行为网络拓扑,通过随机游走或其他算法选取出与用户输入的搜索词相关的推荐结果,完成了基于搜索词进行搜索推荐的过程。由于匿名行为网络拓
扑在统计意义上反映了用户进行搜索的行为规则,将用户最关注的少数维度进行深入挖掘,克服了现有技术中推荐结果过于泛化的缺点,并利用聚类对推荐结果进行划分,并通过知识图谱对每个类进行准确的描述,提供有价值的信息,使得选取出的推荐结果更加符合用户的搜索习惯和搜索需求。
需要说明的是:
在此提供的算法和显示不与任何特定计算机、虚拟装置或者其它设备固有相关。各种通用装置也可以与基于在此的示教一起使用。根据上面的描述,构造这类装置所要求的结构是显而易见的。此外,本发明也不针对任何特定编程语言。应当明白,可以利用各种编程语言实现在此描述的本发明的内容,并且上面对特定语言所做的描述是为了披露本发明的最佳实施方式。
在此处所提供的说明书中,说明了大量具体细节。然而,能够理解,本发明的实施例可以在没有这些具体细节的情况下实践。在一些实例中,并未详细示出公知的方法、结构和技术,以便不模糊对本说明书的理解。
类似地,应当理解,为了精简本公开并帮助理解各个发明方面中的一个或多个,在上面对本发明的示例性实施例的描述中,本发明的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而,并不应将该公开的方法解释成反映如下意图:即所要求保护的本发明要求比在每个权利要求中所明确记载的特征更多的特征。更确切地说,如下面的权利要求书所反映的那样,发明方面在于少于前面公开的单个实施例的所有特征。因此,遵循具体实施方式的权利要求书由此明确地并入该具体实施方式,其中每个权利要求本身都作为本发明的单独实施例。
本领域那些技术人员可以理解,可以对实施例中的设备中的模块进行自适应性地改变并且把它们设置在与该实施例不同的一个或多个设备中。可以把实施例中的模块或单元或组件组合成一个模块或单元或组件,以及此外可以把它们分成多个子模块或子单元或子组件。除了这样的特征和/或过程或者单元中的至少一些是相互排斥之外,可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有特征以及如此公开的任何方法或者设备的所有过程或单元进行组合。除非另外明确陈述,本说明书(包括伴随的权利要求、摘要和附图)中公开的每个特征可以由提供相同、等同或相似目的的替代特征来代替。
此外,本领域的技术人员能够理解,尽管在此所述的一些实施例包括其它实施例中所包括的某些特征而不是其它特征,但是不同实施例的特征的组合意味着处于本发明的范围之内并且形成不同的实施例。例如,在下面的权利要求书中,所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本发明的各个部件实施例可以以硬件实现,或者以在一个或者多个处理器上运行的软件模块实现,或者以它们的组合实现。本领域的技术人员应当理解,可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本发明实施例的一种基于搜索词进行搜索推荐的装置中的一些或者全部部件的一些或者全部功能。本发明还可以实现为用于执行这里所描述的方法的一部分或者全部的设备或者装置程序(例如,计算机程序和计算机程序产品)。这样的实现本发明的程序可以存储在计算机可读介质上,或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到,或者在载体信号上提供,或者以任何其他形式提供。
例如,图5示意性地示出了用于执行根据本发明的方法的计算设备的框图。该计算设备传统上包括处理器510和以存储器520形式的计算机程序产品或者计算机可读介质。存储器520可以是诸如闪存、EEPROM(电可擦除可编程只读存储器)、EPROM、硬盘或者ROM之类的电子存储器。存储器520具有用于执行上述方法中的任何方法步骤的程序代码531的存储空间530。例如,用于程序代码的存储空间530可以包括分别用于实现上面的方法中的各种步骤的各个程序代码531。这些程序代码可以从一个或者多个计算机程序产品中读出或者写入到这一个或者多个计算机程序产品中。这些计算机程序产品包括诸如硬盘,紧致盘(CD)、存储卡或者软盘之类的程序代码载体。这样的计算机程序产品通常为如参考图6所述的便携式或者固定存储单元。该存储单元可以具有与图5的计算设备中的存储器520类似布置的存储段、存储空间等。程序代码可以例如以适当形式进行压缩。通常,存储单元包括用于执行根据本发明的方法步骤的计算机可读代码531’,即可以由例如诸如510之类的处理器读取的代码,这些代码当由计算设备运行时,导致该计算设备执行上面所描述的方法中的各个步骤。
应该注意的是上述实施例对本发明进行说明而不是对本发明
进行限制,并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出替换实施例。在权利要求中,不应将位于括号之间的任何参考符号构造成对权利要求的限制。单词“包含”不排除存在未列在权利要求中的元件或步骤。位于元件之前的单词“一”或“一个”不排除存在多个这样的元件。本发明可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中,这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。
此外,还应当注意,本说明书中使用的语言主要是为了可读性和教导的目的而选择的,而不是为了解释或者限定本发明的主题而选择的。因此,在不偏离所附权利要求书的范围和精神的情况下,对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。对于本发明的范围,对本发明所做的公开是说明性的,而非限制性的,本发明的范围由所附权利要求书限定。
本发明可以应用于计算机系统/服务器,其可与众多其它通用或专用计算系统环境或配置一起操作。适于与计算机系统/服务器一起使用的众所周知的计算系统、环境和/或配置的例子包括但不限于:个人计算机系统、服务器计算机系统、瘦客户机、厚客户机、手持或膝上设备、基于微处理器的系统、机顶盒、可编程消费电子产品、网络个人电脑、小型计算机系统、大型计算机系统和包括上述任何系统的分布式云计算技术环境,等等。
计算机系统/服务器可以在由计算机系统执行的计算机系统可执行指令(诸如程序模块)的一般语境下描述。通常,程序模块可以包括例程、程序、目标程序、组件、逻辑、数据结构等等,它们执行特定的任务或者实现特定的抽象数据类型。计算机系统/服务器可以在分布式云计算环境中实施,分布式云计算环境中,任务是由通过通信网络链接的远程处理设备执行的。在分布式云计算环境中,程序模块可以位于包括存储设备的本地或远程计算系统存储介质上。
本文中所称的“一个实施例”、“实施例”或者“一个或者多个实施例”意味着,结合实施例描述的特定特征、结构或者特性包括在本发明的至少一个实施例中。此外,请注意,这里“在一个实施例中”的词语例子不一定全
指同一个实施例。
Claims (18)
- 一种基于搜索词进行搜索推荐的方法,其中,该方法包括:根据匿名搜索日志构建匿名行为网络拓扑;响应于用户输入搜索词的事件,从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果。
- 如权利要求1所述的方法,其中,该方法进一步包括:对所述推荐结果进行聚类得到多个类,对所述多个类进行排序。
- 如权利要求1-2任一项所述的方法,其中,所述根据匿名搜索日志构建匿名行为网络拓扑包括:根据匿名搜索日志中的内容,以匿名的用户和搜索词为节点,以用户的点击行为边,构建匿名行为网络拓扑。
- 如权利要求1-3任一项所述的方法,其中,该方法在对所述推荐结果进行聚类得到多个类之前进一步包括:对所述推荐结果进行过滤,过滤掉歧义的和属于垃圾内容的推荐结果,得到过滤后的推荐结果;对过滤后的推荐结果进行聚类得到多个类。
- 如权利要求1-4任一项所述的方法,其中,在对所述多个类进行排序后,该方法进一步包括:根据知识图谱,为每个类选择一个恰当的描述作为类的名称。
- 如权利要求1-5任一项所述的方法,其中,所述从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果包括:根据随机游走算法在所述匿名行为网络拓扑中进行随机游走,选取出与所述搜索词最相关的预定数量的推荐结果。
- 如权利要求1-6任一项所述的方法,其中,所述从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果包括:根据Pagerank、Personalized Pagerank、Random Walk with Restart、或Metapath算法从所述匿名行为网络拓扑中选取出与所述搜索词最相关的预定数量的推荐结果。
- 如权利要求1-7中任一项所述的方法,其中,该方法进一步包括:将最终处理得到的推荐结果嵌入搜索结果页面中输出。
- 一种基于搜索词进行搜索推荐的装置,其中,该装置包括:构建单元,适于根据匿名搜索日志构建匿名行为网络拓扑;推荐单元,适于响应于用户输入搜索词的事件,从所述匿名行为网络拓扑中选取出与所述搜索词相关的推荐结果。
- 如权利要求9所述的装置,其中,该装置进一步包括:聚类单元,适于对所述推荐单元得到的所述推荐结果进行聚类得到多个类,对所述多个类进行排序。
- 如权利要求9-10任一项所述的装置,其中,所构建单元,适于根据匿名搜索日志中的内容,以匿名的用户和搜索词为节点,以用户的点击行为边,构建匿名行为网络拓扑。
- 如权利要求9-11所述的装置,其中,该装置进一步包括:过滤单元,适于对所述推荐单元得到的所述推荐结果进行过滤,过滤掉歧义的和属于垃圾内容的推荐结果,得到过滤后的推荐结果后输出给所述聚类单元;
- 如权利要求9-12所述的装置,其中,该装置进一步包括:描述添加单元,适于根据知识图谱,为所述聚类单元聚类得到的每个类选择一个恰当的描述作为类的名称。
- 如权利要求9-13所述的装置,其中,所述推荐单元,适于根据随机游走算法在所述匿名行为网络拓扑中进行随机游走,选取出与所述搜索词最相关的预定数量的推荐结果。
- 如权利要求9-14所述的装置,其中,所述推荐单元,适于根据Pagerank、Personalized Pagerank、Random Walk with Restart、或Metapath算法从所述匿名行为网络拓扑中选取出与所述搜索词最相关的预定数量的推荐结果。
- 如权利要求9-15中任一项所述的装置,其中,该装置进一步包括:输出单元,适于将最终处理得到的推荐结果嵌入搜索结果页面中输出。
- 一种计算机程序,包括计算机可读代码,当所述计算机可读代码在计算设备上运行时,导致所述计算设备执行根据权利要求1-8中的任一项所述的基于搜索词进行搜索推荐的方法。
- 一种计算机可读介质,其中存储了如权利要求17所述的计算机程序。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410843509.2A CN104699751A (zh) | 2014-12-30 | 2014-12-30 | 一种基于搜索词进行搜索推荐的方法和装置 |
| CN201410843509.2 | 2014-12-30 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2016107326A1 true WO2016107326A1 (zh) | 2016-07-07 |
Family
ID=53346872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2015/095016 Ceased WO2016107326A1 (zh) | 2014-12-30 | 2015-11-19 | 一种基于搜索词进行搜索推荐的方法和装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN104699751A (zh) |
| WO (1) | WO2016107326A1 (zh) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110321482A (zh) * | 2019-06-11 | 2019-10-11 | 阿里巴巴集团控股有限公司 | 一种信息的推荐方法、装置及设备 |
| CN111859147A (zh) * | 2020-07-31 | 2020-10-30 | 中国工商银行股份有限公司 | 对象推荐方法、对象推荐装置和电子设备 |
| CN112269882A (zh) * | 2020-10-12 | 2021-01-26 | 西安工程大学 | 一种面向知识图谱的旅游景点推荐方法 |
| CN112732923A (zh) * | 2020-11-13 | 2021-04-30 | 哈尔滨工业大学 | 一种基于知识图谱的快件物流服务语义提取方法 |
| CN117112797A (zh) * | 2023-07-18 | 2023-11-24 | 昆明理工大学 | 基于旅游知识图谱的搜索和推荐方法、系统 |
| CN118520042A (zh) * | 2024-05-31 | 2024-08-20 | 四川省交通勘察设计研究院有限公司 | 基于环境敏感因素的公路可视化选线方法、装置及系统 |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104699751A (zh) * | 2014-12-30 | 2015-06-10 | 北京奇虎科技有限公司 | 一种基于搜索词进行搜索推荐的方法和装置 |
| CN105138664B (zh) * | 2015-09-02 | 2019-03-22 | 中国地质大学(武汉) | 一种隐私保护的大数据推荐方法及系统 |
| CN106919577A (zh) * | 2015-12-24 | 2017-07-04 | 北京奇虎科技有限公司 | 基于搜索词进行搜索推荐的方法、装置和搜索引擎 |
| CN105786977B (zh) * | 2016-02-05 | 2020-03-03 | 北京百度网讯科技有限公司 | 基于人工智能的移动搜索方法和装置 |
| CN106776981B (zh) * | 2016-12-06 | 2020-12-15 | 广州同构科技有限公司 | 一种基于经验知识的智能检索方法 |
| CN110309404A (zh) * | 2018-03-08 | 2019-10-08 | 优酷网络技术(北京)有限公司 | 内容推荐方法及装置 |
| CN109614603A (zh) * | 2018-12-12 | 2019-04-12 | 北京百度网讯科技有限公司 | 用于生成信息的方法和装置 |
| CN110348940A (zh) * | 2019-05-28 | 2019-10-18 | 成都美美臣科技有限公司 | 一种电子商务网站搜索建议的方法 |
| CN110377829B (zh) * | 2019-07-24 | 2021-10-08 | 中国工商银行股份有限公司 | 应用于电子设备的功能推荐方法和装置 |
| CN115221423A (zh) * | 2021-04-16 | 2022-10-21 | 北京三快在线科技有限公司 | 一种信息推荐方法、装置、电子设备及存储介质 |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103455576A (zh) * | 2013-08-22 | 2013-12-18 | 西安交通大学 | 一种基于思维地图的网络学习资源推荐方法 |
| CN104699751A (zh) * | 2014-12-30 | 2015-06-10 | 北京奇虎科技有限公司 | 一种基于搜索词进行搜索推荐的方法和装置 |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101359332A (zh) * | 2008-09-02 | 2009-02-04 | 浙江大学 | 具有语义分类功能的可视化搜索界面的设计方法 |
| CN103218719B (zh) * | 2012-01-19 | 2016-12-07 | 阿里巴巴集团控股有限公司 | 一种电子商务网站导航方法及系统 |
| CN103365839B (zh) * | 2012-03-26 | 2017-12-12 | 深圳市世纪光速信息技术有限公司 | 一种搜索引擎的推荐搜索方法和装置 |
| CN103425643B (zh) * | 2012-05-14 | 2018-07-31 | 深圳市世纪光速信息技术有限公司 | 一种相关搜索请求串推荐方法和系统 |
-
2014
- 2014-12-30 CN CN201410843509.2A patent/CN104699751A/zh active Pending
-
2015
- 2015-11-19 WO PCT/CN2015/095016 patent/WO2016107326A1/zh not_active Ceased
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103455576A (zh) * | 2013-08-22 | 2013-12-18 | 西安交通大学 | 一种基于思维地图的网络学习资源推荐方法 |
| CN104699751A (zh) * | 2014-12-30 | 2015-06-10 | 北京奇虎科技有限公司 | 一种基于搜索词进行搜索推荐的方法和装置 |
Non-Patent Citations (1)
| Title |
|---|
| CHENG, SHUYU: "Study of the Personalized Recommendation System Based on Collaborative Filtering Algorithm", CHINA MASTER'S THESES FULL-TEXT DATABASE (ELECTRONIC JOURNALS, 15 March 2012 (2012-03-15), pages 6 - 38, ISSN: 1674-0246 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110321482A (zh) * | 2019-06-11 | 2019-10-11 | 阿里巴巴集团控股有限公司 | 一种信息的推荐方法、装置及设备 |
| CN110321482B (zh) * | 2019-06-11 | 2023-04-18 | 创新先进技术有限公司 | 一种信息的推荐方法、装置及设备 |
| CN111859147A (zh) * | 2020-07-31 | 2020-10-30 | 中国工商银行股份有限公司 | 对象推荐方法、对象推荐装置和电子设备 |
| CN111859147B (zh) * | 2020-07-31 | 2023-08-22 | 中国工商银行股份有限公司 | 对象推荐方法、对象推荐装置和电子设备 |
| CN112269882A (zh) * | 2020-10-12 | 2021-01-26 | 西安工程大学 | 一种面向知识图谱的旅游景点推荐方法 |
| CN112269882B (zh) * | 2020-10-12 | 2022-10-18 | 西安工程大学 | 一种面向知识图谱的旅游景点推荐方法 |
| CN112732923A (zh) * | 2020-11-13 | 2021-04-30 | 哈尔滨工业大学 | 一种基于知识图谱的快件物流服务语义提取方法 |
| CN112732923B (zh) * | 2020-11-13 | 2022-07-12 | 哈尔滨工业大学 | 一种基于知识图谱的快件物流服务语义提取方法 |
| CN117112797A (zh) * | 2023-07-18 | 2023-11-24 | 昆明理工大学 | 基于旅游知识图谱的搜索和推荐方法、系统 |
| CN118520042A (zh) * | 2024-05-31 | 2024-08-20 | 四川省交通勘察设计研究院有限公司 | 基于环境敏感因素的公路可视化选线方法、装置及系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104699751A (zh) | 2015-06-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109074358B (zh) | 提供与用户兴趣有关的地理位置 | |
| CN104699751A (zh) | 一种基于搜索词进行搜索推荐的方法和装置 | |
| Jiang et al. | Learning query and document relevance from a web-scale click graph | |
| US9922125B1 (en) | Selecting content using entity properties | |
| US9311406B2 (en) | Discovering trending content of a domain | |
| US20110202533A1 (en) | Dynamic Search Interaction | |
| US10824806B2 (en) | Counterintuitive recommendations based upon temporary conditions | |
| CN107480158A (zh) | 基于相似性得分评估内容项目与图像的匹配的方法和系统 | |
| CN106462807A (zh) | 根据大规模非结构化数据学习多媒体语义 | |
| CN105339933A (zh) | 通过查询扩展的新闻结果 | |
| CN111522886B (zh) | 一种信息推荐方法、终端及存储介质 | |
| CN107463592B (zh) | 用于将内容项目与图像匹配的方法、设备和数据处理系统 | |
| CN105593851A (zh) | 用于跟踪与通过关联的文本和图像可识别的实体相关的微博消息的方法和装置 | |
| CN108241646B (zh) | 一种搜索匹配方法和装置、推荐方法和装置 | |
| CN106021418B (zh) | 新闻事件的聚类方法及装置 | |
| CN108959323B (zh) | 视频分类方法和装置 | |
| US9846751B2 (en) | Takepart action platform for websites | |
| CN108509497A (zh) | 信息推荐方法、装置和电子设备 | |
| WO2020155877A1 (zh) | 信息推荐 | |
| KR102215082B1 (ko) | Cnn 기반 이미지 검색 방법 및 장치 | |
| US20190019217A1 (en) | Group formation and recommendations based on trigger events | |
| CN110609959B (zh) | 基于项目生命周期的检索方法、存储介质及电子设备 | |
| Zhao et al. | Detecting profilable and overlapping communities with user-generated multimedia contents in LBSNs | |
| Choi et al. | Multimodal location estimation of consumer media: Dealing with sparse training data | |
| CN106682146A (zh) | 一种根据关键词检索景区评价的方法及系统 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15875005 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 15875005 Country of ref document: EP Kind code of ref document: A1 |