WO2017096896A1 - 资源搜索方法以及服务器 - Google Patents
资源搜索方法以及服务器 Download PDFInfo
- Publication number
- WO2017096896A1 WO2017096896A1 PCT/CN2016/089525 CN2016089525W WO2017096896A1 WO 2017096896 A1 WO2017096896 A1 WO 2017096896A1 CN 2016089525 W CN2016089525 W CN 2016089525W WO 2017096896 A1 WO2017096896 A1 WO 2017096896A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- resource
- words
- user
- resources
- searched
- 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
- G06F16/951—Indexing; Web crawling techniques
Definitions
- the embodiments of the present invention relate to the field of communications technologies, and in particular, to a resource search method and a server.
- keywords search such as web pages, video/audio and other media streams, images, applications, and the like.
- the search is generally one-way vertical search, for example, the app store only has app-related search, and the video player has only video-related search. Users searching for different resource types need to open different search engines, which causes inconvenience and poor user experience.
- the display is not performed according to the user's intention, so the user may need to spend some time to find the desired resource among the searched large resources, which also causes the user experience to be poor. .
- the present invention provides a resource search method, which includes: performing resource search in a resource database according to an entry entered by a user, in response to a technical problem that the existing resource search and display method in the prior art cannot meet the user's needs. Search and user intent recognition; and sorting the searched resources according to the results of the user's intent to identify.
- the embodiment of the present invention further provides a server, where the server includes: a resource database; and a processing module, configured to perform resource search in the resource database according to the entry entered by the user. And user intent identification; and sorting the searched resources for the results identified by the user's intent.
- the resource search method and the server provided by the embodiment of the present invention can perform resource search and user intention identification in a resource database according to an entry input by a user, and then sort the searched resources according to the result of the user's intention identification, thereby achieving All types of resources related to the entry entered by the user (for example, resources including video, application, audio, etc.) are sorted according to the user's intention. For example, if the user desires a video resource, it can be among a large number of searched resources. Video resources are arranged at the forefront, so users can easily and quickly find the resources they need, saving time and greatly improving the user experience.
- FIG. 1 is a schematic structural diagram of an example server according to an embodiment of the present invention.
- FIG. 2 is an example flow diagram of execution of a processing module of an example server in accordance with an embodiment of the present invention
- FIG. 3 is an example flow diagram of a resource search method in accordance with an embodiment of the present invention.
- the server may include: a resource database 10; and a processing module 20 for using the entry entered by the user in the resource database 10 Performing resource search and user intent recognition; and sorting the searched resources according to the results of the user's intention to identify.
- the server 100 first needs to establish a resource database 10, which may include at least one of the following words:
- a plurality of sensitive words and a resource type corresponding to each of the plurality of sensitive words wherein the sensitive words are words capable of identifying the type of the resource, for example, the "theatrical version” may be a sensitive word, and the The resource type corresponding to the sensitive word is video, that is, the "theatrical version, video" can be stored in the database 10;
- the video resource corresponding weight value is 5000
- the application (app) resource corresponding weight value is 2000
- the audio resource corresponding weight value is 1500, that is, The "flowers, video: 5000, app: 2000, audio: 1500" is stored in the database 10;
- "Wu Zetian” can be a named entity, and the named entity also corresponds to the weight values of various resource types, for example
- the video resource corresponding weight value is 6000
- the application (app) resource corresponding weight value is 1500
- the audio resource corresponding weight value is 500, that is, "Wu Zetian, video: 6000, app: 1500, audio: 500" can be stored in the number.
- the library 10 According to the library 10;
- the weight values of the plurality of words and the plurality of resource types corresponding to each of the plurality of words may be one word respectively, and the words respectively correspond to the plurality of resources
- the weight value of the type that is, "Daddy, Video: 6000, App: 1200, Audio: 3000", "Mahjong, Video: 500, App: 2500, Audio: 0" can be stored in the database 10.
- the greater the weight value of the above vocabulary indicates the more relevant to the user's intention, and the weight value is obtained according to the TF-IDF (term frequency–inverse document frequency) algorithm, and the calculation of the algorithm is similar to the prior art, in order not to confuse the present The scope of protection of the invention will not be described herein.
- TF-IDF term frequency–inverse document frequency
- the above vocabulary and the examples of weight values are only used to illustrate the idea of the present invention and are not intended to limit the scope of the present invention. Those skilled in the art can appropriately set the vocabulary and weight values according to actual conditions. The invention does not limit this. Moreover, the resource database provided by the present invention preferably includes the above three vocabulary, but those skilled in the art may also select any one of the above and combinations thereof according to actual conditions.
- FIG. 2 is an example flow diagram of execution of a processing module of an example server in accordance with an embodiment of the present invention, as shown in FIG. 2, specifically:
- the processing module 20 first performs step S1, that is, the term input by the user divides the term into a plurality of words according to a word segmentation algorithm.
- the word segmentation algorithm may be an appropriate word segmentation algorithm such as the Friso algorithm, and the calculation of the algorithm is similar to the prior art. In order not to obscure the scope of protection of the present invention, no further description is provided herein;
- At least one of the plurality of words after the word segmentation may be determined to be in the resource database. Whether at least one of the plurality of sensitive words matches, that is, the execution step S2;
- the processing module 20 may identify the user's intention as the first search in the resource database. a resource type corresponding to the matched sensitive word, and sorting the searched resource corresponding to the resource type;
- step S4 needs to be performed, that is, the plurality of words are determined. Whether at least one of the words matches at least one of the plurality of named entities in the resource database;
- the user intention is Identifying the weight values of the multiple resource types corresponding to the at least one named entity, and sorting the searched multiple resources corresponding to the multiple resource types according to the weight value, that is, performing step S5, and displaying the sorting After the resources.
- the processing module 20 may calculate a total weight value of the resources of the named entities, and The user intention is identified as a weight value of a plurality of resource types corresponding to the named entity having the largest resource total weight value, and the searched plurality of resources corresponding to the multiple resource types are sorted according to the weight value, wherein one named entity
- the total weight of the resource is the sum of the weight values of its corresponding multiple resource types;
- step S6 that is, the processing module is required. 20 need to determine at least one of the plurality of words and the plurality of resource databases Whether at least one of the words matches;
- the user intention is identified as a plurality of resource types corresponding to the at least one word
- the weight value is sorted, and the searched resource corresponding to the resource type is sorted according to the weight value, that is, step S7 is performed, and the sorted resource can be displayed.
- the processing module 20 may weight the weight values of the plurality of resource types corresponding to the words, and according to The weighted weight value displays the resource;
- step S8 is performed to sort the search in the default order. Resources related to multiple words.
- the processing module 20 can perform the above steps S1-S8 according to the user's input, and can display the searched resources according to the above sorting, and then can be published and pushed to the client device (such as a mobile phone, a computer, a tablet, etc.) through the webpage. This allows you to display a variety of searched resources sorted by user intent on the client device.
- the client device such as a mobile phone, a computer, a tablet, etc.
- the user inputs "**theatrical version", and the word “theatrical version” in the word obtained after the word segmentation is a sensitive word, and the processing module 20 of the server 100 can perform steps S1-S3, that is, the user intention is identified as a video resource, and therefore, the resource is searched. Thereafter, the processing module 20 arranges the video resources related to the "**theatrical version" at the forefront of the searched resource list and pushes it to the client device.
- the processing module 20 can identify the user's intent as the resource type corresponding to the theatrical version in the resource database, and sort the searched resources corresponding to the resource type, and vice versa.
- the user inputs "the fifth set of flowers and bones", and the word “flowers” obtained after the word segmentation is a named entity, and the processing module 20 of the server 100 can perform steps S1-S5, that is, the user's intention is identified as being the first video resource, and second. It is app, then audio. Therefore, after searching for resources, the processing module 20 arranges the resources related to the "Flower, App, and Audio" in the order of video, app, and audio and pushes them to the client device.
- the processing module 20 can calculate the total weights of the resources of the two named entities, that is, as described above,
- the weight values of the corresponding resource types in the resource database are video: 5000, app: 2000, audio: 1500, and the total weight of the resources is 8500.
- the weight values of the corresponding resource types in the resource database of Wu Zetian are respectively The video is 6000, app: 1500, audio: 500, and the total weight of the resource is 8000. Since the total weight of the resource of the flower bone is the largest, the processing module 20 identifies the user's intention as a plurality of resources corresponding to the flower bone. A weight value of the type, and sorting the searched plurality of resources corresponding to the plurality of resource types according to the weight value, and vice versa.
- the user inputs "Daddy Goes Home", and after the word segmentation, "Daddy” is the word "Daddy” (the word “going home” obtained at this time is not the word in the resource database 10), that is, the processing module 20 of the server 100 can perform the steps.
- S1-S7 that is, the user's intention is identified as the weight value corresponding to the word "Dad”.
- the user intent is identified as being a video resource first, followed by an app, then an audio. Therefore, after searching for the resource, the processing module 20 arranges and pushes resources related to "Daddy Go Home” in the order of video, app, and audio. To the client device.
- the processing module 20 can perform steps S1-S7, ie, the user intent is identified as a weighting of the weight values of the words "dad” and "mahjong.”
- the user intent is identified as being the video resource first, followed by the app, then the audio. Therefore, after searching for the resource, the processing module 20 arranges and pushes the resources related to the “Daddy Mahjong” in the order of video, app, and audio. Client device.
- the user inputs "Where are you", the word obtained by the word segmentation is neither a sensitive word, a named entity, nor a word, so the processing module 20 of the server 100 can perform steps S1-S8, that is, sorting the searched in the default order. Resources are pushed to the client device.
- FIG. 3 is a flow chart showing an example of a resource search method according to an embodiment of the present invention. As shown in FIG. 3, the method may include the following steps:
- Step 1001 Perform resource search and user intention identification in a resource database according to an entry entered by a user;
- Step 1002 Sort the searched resources according to the result of the user's intention identification.
- the performing resource search and user intent recognition in the resource database according to the entry input by the user comprises: dividing the term input by the user into the plurality of words according to a word segmentation algorithm; Multiple words are described for resource search and user intent recognition.
- the resource database includes a plurality of sensitive words and a resource type corresponding to each of the plurality of sensitive words, wherein the sensitive words are words capable of identifying the resource type.
- sorting the searched resources according to the result of the user's intention identification includes: if at least one of the plurality of words matches at least one of the plurality of sensitive words in the resource database, the user is The intent is identified as a resource type corresponding to the at least one sensitive word, and the searched resource corresponding to the resource type is sorted.
- the resource database includes a plurality of named entities and weight values of a plurality of resource types corresponding to each of the plurality of named entities, wherein the named entity is an entity identified by a name.
- sorting the searched resources according to the result of the user's intention to identify includes: if at least one of the plurality of words matches at least one of the plurality of named entities in the resource database, The user intention is identified as a weight value of a plurality of resource types corresponding to the at least one named entity, and the searched plurality of resources corresponding to the plurality of resource types are sorted according to the weight value.
- the resource database includes a plurality of words and weight values of a plurality of resource types corresponding to each of the plurality of words.
- sorting the searched resources according to the result of the user's intention to identify includes: in the case of at least one of the plurality of words and at least one of the plurality of words in the resource database, the user intent Identifying weight values of the plurality of resource types corresponding to the at least one word, and sorting the searched resources corresponding to the resource types according to the weight values.
- the resource search method and the server provided by the embodiment of the present invention can implement all types of resources related to the entry input by the user (for example, resources including video, application, audio, etc.) according to the intention of the user, for example, if the user If the video resource is expected, the video resources can be arranged at the forefront in a large number of searched resources, so that the user can find the resources that he needs conveniently and quickly, which saves time and greatly improves the user experience. .
- resources related to the entry input by the user for example, resources including video, application, audio, etc.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (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
一种资源搜索方法以及服务器,该方法包括:根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别(1001),按照用户意图识别的结果排序搜索到的资源(1002)。该方法能够实现将与用户输入的词条相关的所有类型的资源按照用户的意图排序,例如,如果用户期望的是视频资源,则可以在大量搜索到的资源中,将视频资源排列在最前端,这样用户可以方便、快速地找到自己需要的资源,在节约了时间的同时,也大大提高了用户体验。
Description
本申请要求于2015年12月07日提交中国专利局、申请号为201510887858.9的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本发明实施例涉及通信技术领域,具体地,涉及一种资源搜索方法、以及一种服务器。
随着网络的发展,用户可以通过关键词搜索在各种搜索引擎上搜索用户感兴趣的资源,例如网页、视频/音频等媒体流、图像、应用程序等。
但是,在现有技术中搜索一般为单向的垂直搜索,比如应用商店只有app相关搜索,视频播放器只有视频相关搜索。用户进行不同资源类型的搜索,需要打开不同的搜索引擎,这引起了不便,造成用户体验不佳。此外,现有技术中在搜索到资源后并不是按照用户意图进行显示的,因此用户很有可能需要花费一些时间才能在搜索到的大量资源中找到自己期望的资源,这也造成用户体验较差。
发明内容
针对现有技术中存在的现有资源搜索和显示方法不能满足用户需求的技术问题,本发明实施例提供了一种资源搜索方法,该方法包括:根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别;以及按照用户意图识别的结果排序搜索到的资源。
相对应地,本发明实施例还提供了一种服务器,该服务器包括:资源数据库;以及处理模块,用于根据用户输入的词条在资源数据库中进行资源搜
索和用户意图识别;以及用于按照用户意图识别的结果排序搜索到的资源。
采用本发明实施例提供的资源搜索方法以及服务器,通过根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别,以及之后按照用户意图识别的结果排序搜索到的资源,能够实现将与用户输入的词条相关的所有类型的资源(例如包括视频、应用程序、音频等资源)按照用户的意图排序,例如,如果用户期望的是视频资源,则可以在大量搜索到的资源中,将视频资源排列在最前端,这样用户可以方便、快速地找到自己需要的资源,在节约了时间的同时,也大大提高了用户体验。
本发明实施例的其它特征和优点将在随后的具体实施方式部分予以详细说明。
附图是用来提供对本发明的进一步理解,并且构成说明书的一部分,与下面的具体实施方式一起用于解释本发明,但并不构成对本发明的限制。在附图中:
图1是根据本发明的一种实施方式的示例服务器的结构示意图;
图2是根据本发明的一种实施方式的示例服务器的处理模块执行的示例流程图;以及
图3是根据本发明的一种实施方式的资源搜索方法的示例流程图。
附图标记说明
10 资源数据库 20 处理模块 100服务器
以下结合附图对本发明的具体实施方式进行详细说明。应当理解的是,
此处所描述的具体实施方式仅用于说明和解释本发明实施例,并不用于限制本发明。
下面将举例说明本发明的思想,但应当理解的这些示例为非局限性示例,本发明的保护范围不限于此:
为了更加清楚地说明本发明的思想,以本发明实施例所提供的示例服务器进行详细地说明。
图1是根据本发明的一种实施方式的示例服务器的结构示意图,如图1所示,该服务器可以包括:资源数据库10;以及处理模块20,用于根据用户输入的词条在资源数据库10中进行资源搜索和用户意图识别;以及用于按照用户意图识别的结果排序搜索到的资源。
为了实现本发明实施例的目的,服务器100首先需要建立资源数据库10,所述资源数据库10可以包括以下词汇中的至少一者:
(1)多个敏感词以及多个敏感词中的每一个敏感词所对应的资源类型,其中所述敏感词为能够标识资源类型的词语,例如“剧场版”可以为一个敏感词,并且该敏感词对应的资源类型为视频,即可以将“剧场版,视频”存储在数据库10中;
(2)多个命名实体以及多个命名实体中的每一个命名实体所对应的多种资源类型的权重值,其中所述命名实体为以名称为标识的实体,例如“花千骨”可以为一个命名实体,并且该命名实体还对应多种资源类型的权重值,例如,视频资源对应权重值为5000,应用程序(app)资源对应权重值为2000,音频资源对应权重值为1500,即可以将“花千骨,视频:5000,app:2000,音频:1500”存储在数据库10中;再例如“武则天”可以为一个命名实体,并且该命名实体还对应多种资源类型的权重值,例如,视频资源对应权重值为6000,应用程序(app)资源对应权重值为1500,音频资源对应权重值为500,即可以将“武则天,视频:6000,app:1500,音频:500”存储在数
据库10中;
(3)多个词语以及多个词语中的每一个词语所对应的多种资源类型的权重值,例如“爸爸”和“麻将”可以分别为一个词语,并且所述词语还分别对应多种资源类型的权重值,即可以将“爸爸,视频:6000,app:1200,音频:3000”、“麻将,视频:500,app:2500,音频:0”存储在数据库10中。其中,上述词汇的权重值越大表示与用户意图越相关,并且所述权重值根据TF-IDF(term frequency–inverse document frequency)算法获得,该算法的计算与现有技术相似,为了不混淆本发明的保护范围,在此不再赘述。
应当理解的是,上述词汇以及权重值的示例仅用于说明本发明的思想而不用于限定本发明的保护范围,本领域技术人员可以根据实际情况对词汇和权重值进行适当地设定,本发明对此不进行限定。并且,本发明提供的资源数据库优选地包括上述三种词汇,但本领域技术人员也可以根据实际情况选择上述中的任一者及其组合。
当用户通过客户端设备(例如移动电话、计算机、平板电脑等)输入一个词条时,服务器的处理模块20可以根据用户输入的词条在资源数据库10中进行资源搜索和用户意图识别;以及用于按照用户意图识别的结果排序搜索到的资源。图2是根据本发明的一种实施方式的示例服务器的处理模块执行的示例流程图,如图2所示,具体地:
所述处理模块20首先执行步骤S1,即将用户输入的所述词条根据分词算法将所述词条分为多个词。优选地,所述分词算法可以为Friso算法等适当地分词算法,所述算法的计算与现有技术相似,为了不混淆本发明的保护范围,在此不再赘述;
接着,由于上述三种词汇与用户意图之间的相关性是从高到低的,因此为了精确地识别用户意图,首先可以判断分词后的多个词中的至少一者与所述资源数据库中的多个敏感词中的至少一个敏感词是否匹配,即执行步骤
S2;
在所述多个词中的至少一者与所述资源数据库中的多个敏感词中的至少一个至少一者匹配的情况下(即此时用户输入的词条为包含敏感词),则将用户意图识别为与所述至少一个敏感词所对应的资源类型,即执行步骤S3,并可以显示排序后的资源。其中,当多个词分别与多个敏感词对应时(即两个以上的词在资源数据库中找到与其匹配的敏感词),处理模块20可以将用户意图识别为在资源数据库中第一个搜索到的匹配的敏感词所对应的资源类型,并排序搜索到的与所述资源类型对应的资源;
反之,在所述多个词与所述资源数据库中的所有敏感词均不匹配情况下(即此时用户输入的词条不包含敏感词),则需要执行步骤S4,即判断所述多个词中的至少一者与所述资源数据库中的多个命名实体中的至少一者是否匹配;
在所述多个词中的至少一者与所述资源数据库中的多个命名实体中的至少一者匹配的情况下(即此时用户输入的词条为包含命名实体),则将用户意图识别为与所述至少一个命名实体所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述多种资源类型对应的多个资源,即执行步骤S5,并可以显示排序后的资源。其中,当多个词分别与多个命名实体对应时(即两个以上的词在资源数据库中找到与其匹配的命名实体),处理模块20可以将计算这些命名实体的资源总权重值,并将用户意图识别为与具有最大资源总权重值的命名实体所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述多种资源类型对应的多个资源,其中一个命名实体的资源总权值为其对应的多种资源类型的权重值的总和;
反之,在所述多个词与所述资源数据库中的所有命名实体均不匹配的情况下(即此时用户输入的词条不包含命名实体),则需要执行步骤S6,即所述处理模块20需要判断所述多个词中的至少一者与所述资源数据库中的多
个词语中的至少一者是否匹配;
在所述多个词中的至少一者与所述资源数据库中的多个词语中的至少一个词语匹配的情况下,则将用户意图识别为与所述至少一个词语所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述资源类型对应的资源,即执行步骤S7,并可以显示排序后的资源。当多个词分别与多个词语对应时(即两个以上的词在资源数据库中找到与其匹配的词语),处理模块20可以对这些词语对应的多种资源类型的权重值进行加权,并按照加权后的权重值显示资源;
反之,在所述多个词与所述资源数据库中的所有词语均不匹配的情况下(即此时用户输入的词条也不包含任何词语),则执行步骤S8按照默认的顺序排序搜索到的与多个词有关的资源。
处理模块20可以执行根据用户的输入执行上述步骤S1-S8,并可以按照上述排序显示搜索到的资源,之后可以通过网页发布并推送给客户端设备(例如移动电话、计算机、平板电脑等),这样就可以在客户端设备上显示按照用户意图排序的搜索到的多种资源。
下面列举一些实际使用时可能出现的实施例来进一步阐明本发明的上述思想,但应该理解的是,这些实施例仅用于说明本发明的思想,本发明实施例并不局限于此,例如本领域技术人员可以在下面实施例的基础上设定各种词汇和用户输入:
实施例1
用户输入“**剧场版”,分词后得到的词中“剧场版”是敏感词,服务器100的处理模块20可以执行步骤S1-S3,即将用户意图识别为视频资源,因此,在搜索到资源后,处理模块20将与“**剧场版”相关的视频资源排列在搜索到的资源列表的最前端并推送给客户端设备。当用户输入多个敏感词时,例如用户输入的词条分词后得到“剧场版”和“游戏”,如果先匹配
的是敏感词剧场版,则处理模块20可以将用户意图识别为在资源数据库中剧场版所对应的资源类型,并排序搜索到的与所述资源类型对应的资源,反之亦然。
实施例2
用户输入“花千骨第五集”,分词后得到的词中“花千骨”是命名实体,服务器100的处理模块20可以执行步骤S1-S5,即将用户意图识别为首先是视频资源、其次是app、然后是音频,因此,在搜索到资源后,处理模块20将与“花千骨第五集”相关的资源按照视频、app、音频的顺序排列并推送给客户端设备。当用户输入多个命名实体时,例如用户输入的词条分词后得到“花千骨”和“武则天”,处理模块20可以计算这两个命名实体的资源总权值,即如上所述,花千骨在资源数据库中对应的资源类型的权重值分别为视频:5000,app:2000,音频:1500,资源总权值为8500;相似地,武则天在资源数据库中对应的资源类型的权重值分别为视频:6000,app:1500,音频:500,资源总权值为8000,由于花千骨的资源总权值最大,因此处理模块20将用户意图识别为与花千骨所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述多种资源类型对应的多个资源,反之亦然。
实施例3
用户输入“爸爸回家”,分词后得到“爸爸”为词语“爸爸”(此时分词得到的词“回家”并不是资源数据库10中的词语),即服务器100的处理模块20可以执行步骤S1-S7,即将用户意图识别为词语“爸爸”对应的权重值。即将用户意图识别为首先是视频资源、其次是app、然后是音频,因此,在搜索到资源后,处理模块20将与“爸爸回家”相关的资源按照视频、app、音频的顺序排列并推送给客户端设备。
实施例4
用户输入“爸爸麻将”,分词得到词语“爸爸”和“麻将”,即服务器100
的处理模块20可以执行步骤S1-S7,即将用户意图识别为词语“爸爸”和“麻将”的权重值的加权。即将用户意图识别为首先是视频资源、其次是app、然后是音频,因此,在搜索到资源后,处理模块20将与“爸爸麻将”相关的资源按照视频、app、音频的顺序排列并推送给客户端设备。
实施例5
用户输入“你在哪”,该词条分词得到的词既不是敏感词、命名实体、也不是词语,因此服务器100的处理模块20可以执行步骤S1-S8,即按照默认的顺序排序搜索到的资源并推送给客户端设备。
图3是根据本发明的一种实施方式的资源搜索方法的示例流程图,如图3所示,该方法可以包括以下步骤:
步骤1001,根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别;以及
步骤1002,按照用户意图识别的结果排序搜索到的资源。
一个例子是,所述根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别包括:将用户输入的所述词条根据分词算法将所述词条分为多个词;以及根据所述多个词进行资源搜索和用户意图识别。
一个例子是,所述资源数据库包括多个敏感词以及多个敏感词中的每一个敏感词所对应的资源类型,其中所述敏感词为能够标识资源类型的词语。
一个例子是,按照用户意图识别的结果排序搜索到的资源包括:所述多个词中的至少一者与所述资源数据库中的多个敏感词中的至少一者匹配的情况下,将用户意图识别为与所述至少一个敏感词所对应的资源类型,并排序搜索到的与所述资源类型对应的资源。
一个例子是,所述资源数据库包括多个命名实体以及多个命名实体中的每一个命名实体所对应的多种资源类型的权重值,其中所述命名实体为以名称为标识的实体。
一个例子是,按照用户意图识别的结果排序搜索到的资源包括:在所述多个词中的至少一者与所述资源数据库中的多个命名实体中的至少一者匹配的情况下,将用户意图识别为与所述至少一个命名实体所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述多种资源类型对应的多个资源。
一个例子是,所述资源数据库包括多个词语以及多个词语中的每一个词语所对应的多种资源类型的权重值。
一个例子是,按照用户意图识别的结果排序搜索到的资源包括:在所述多个词中的至少一者与所述资源数据库中的多个词语中的至少一者的情况下,将用户意图识别为与所述至少一个词语所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述资源类型对应的资源。
应当理解的是,上述资源搜索方法的各个具体实施方式,均已在示例服务器的实施方式中做了详细地说明(如上所述),在此不再赘述。并且,本领域技术人员可以根据本发明的公开选择上述各种实施方式中的任一者,或者选择上述各种实施方式的组合来配置服务器,并且其他的替换实施方式也落入本发明的保护范围。
采用本发明实施例提供的资源搜索方法以及服务器,能够实现将与用户输入的词条相关的所有类型的资源(例如包括视频、应用程序、音频等资源)按照用户的意图排序,例如,如果用户期望的是视频资源,则可以在大量搜索到的资源中,将视频资源排列在最前端,这样用户可以方便、快速地找到自己需要的资源,在节约了时间的同时,也大大提高了用户体验。
以上结合附图详细描述了本发明的优选实施方式,但是,本发明实施例并不限于上述实施方式中的具体细节,在本发明的技术构思范围内,可以对本发明的技术方案进行多种简单变型,这些简单变型均属于本发明的保护范围。
另外需要说明的是,在上述具体实施方式中所描述的各个具体技术特征,在不矛盾的情况下,可以通过任何合适的方式进行组合,为了避免不必要的重复,本发明实施例对各种可能的组合方式不再另行说明。
此外,本发明的各种不同的实施方式之间也可以进行任意组合,只要其不违背本发明的思想,其同样应当视为本发明所公开的内容。
Claims (16)
- 一种资源搜索方法,包括:根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别;以及按照用户意图识别的结果排序搜索到的资源。
- 根据权利要求1所述的方法,其中,所述根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别包括:将用户输入的所述词条根据分词算法分为多个词;以及根据所述多个词进行资源搜索和用户意图识别。
- 根据权利要求2所述的方法,其中,所述资源数据库包括多个敏感词以及多个敏感词中的每一个敏感词所对应的资源类型,其中所述敏感词为能够标识资源类型的词语。
- 根据权利要求3所述的方法,其中,按照用户意图识别的结果排序搜索到的资源包括:所述多个词中的至少一者与所述资源数据库中的多个敏感词中的至少一者匹配的情况下,将用户意图识别为与所述至少一个敏感词所对应的资源类型,并排序搜索到的与所述资源类型对应的资源。
- 根据权利要求4所述的方法,其中,所述资源数据库包括多个命名实体以及多个命名实体中的每一个命名实体所对应的多种资源类型的权重值,其中所述命名实体为以名称为标识的实体。
- 根据权利要求5所述的方法,其中,按照用户意图识别的结果排序 搜索到的资源包括:在所述多个词中的至少一者与所述资源数据库中的多个命名实体中的至少一者匹配的情况下,将用户意图识别为与所述至少一个命名实体所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述多种资源类型对应的多个资源。
- 根据权利要求6所述的方法,其中,所述资源数据库包括多个词语以及多个词语中的每一个词语所对应的多种资源类型的权重值。
- 根据权利要求7所述的方法,其中,按照用户意图识别的结果排序搜索到的资源包括:在所述多个词中的至少一者与所述资源数据库中的多个词语中的至少一者的情况下,将用户意图识别为与所述至少一个词语所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述资源类型对应的资源。
- 一种服务器,包括:资源数据库;以及处理模块,用于根据用户输入的词条在资源数据库中进行资源搜索和用户意图识别;以及用于按照用户意图识别的结果排序搜索到的资源。
- 根据权利要求9所述的服务器,所述处理模块具体用于:将用户输入的所述词条根据分词算法分为多个词;以及根据所述多个词进行资源搜索和用户意图识别。
- 根据权利要求10所述的服务器,所述资源数据库包括多个敏感词以及多个敏感词中的每一个敏感词所对应的资源类型,其中所述敏感词为能 够标识资源类型的词语。
- 根据权利要求11所述的服务器,所述处理模块具体用于:所述多个词中的至少一者与所述资源数据库中的多个敏感词中的至少一者匹配的情况下,将用户意图识别为与所述至少一个敏感词所对应的资源类型,并排序搜索到的与所述资源类型对应的资源。
- 根据权利要求12所述的服务器,所述资源数据库包括多个命名实体以及多个命名实体中的每一个命名实体所对应的多种资源类型的权重值,其中所述命名实体为以名称为标识的实体。
- 根据权利要求13所述的服务器,所述处理模块具体用于:在所述多个词中的至少一者与所述资源数据库中的多个命名实体中的至少一者匹配的情况下,将用户意图识别为与所述至少一个命名实体所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述多种资源类型对应的多个资源。
- 根据权利要求14所述的服务器,所述资源数据库包括多个词语以及多个词语中的每一个词语所对应的多种资源类型的权重值。
- 根据权利要求15所述的服务器,所述处理模块具体用于:在所述多个词中的至少一者与所述资源数据库中的多个词语中的至少一者的情况下,将用户意图识别为与所述至少一个词语所对应的多种资源类型的权重值,并按照权重值排序搜索到的与所述资源类型对应的资源。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US15/242,027 US20170161322A1 (en) | 2015-12-07 | 2016-08-19 | Method and electronic device for searching resource |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510887858.9 | 2015-12-07 | ||
| CN201510887858.9A CN105893427A (zh) | 2015-12-07 | 2015-12-07 | 资源搜索方法以及服务器 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US15/242,027 Continuation US20170161322A1 (en) | 2015-12-07 | 2016-08-19 | Method and electronic device for searching resource |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017096896A1 true WO2017096896A1 (zh) | 2017-06-15 |
Family
ID=57001938
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2016/089525 Ceased WO2017096896A1 (zh) | 2015-12-07 | 2016-07-10 | 资源搜索方法以及服务器 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105893427A (zh) |
| WO (1) | WO2017096896A1 (zh) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112380411A (zh) * | 2020-11-10 | 2021-02-19 | 北京达佳互联信息技术有限公司 | 敏感词处理方法、装置、电子设备、系统及存储介质 |
| CN113050806A (zh) * | 2019-12-27 | 2021-06-29 | 北京搜狗科技发展有限公司 | 一种信息推荐方法、装置和电子设备 |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543016A (zh) * | 2018-11-15 | 2019-03-29 | 北京搜狗科技发展有限公司 | 一种数据处理方法、装置和用于数据处理的装置 |
| CN111866609B (zh) * | 2019-04-08 | 2022-12-13 | 百度(美国)有限责任公司 | 用于生成视频的方法和装置 |
| CN110688572A (zh) * | 2019-09-24 | 2020-01-14 | 四川新网银行股份有限公司 | 冷启动状态下搜索意图的识别方法 |
| CN110674320B (zh) * | 2019-09-27 | 2022-03-18 | 百度在线网络技术(北京)有限公司 | 一种检索方法、装置和电子设备 |
| CN117056622A (zh) * | 2019-10-22 | 2023-11-14 | 海信视像科技股份有限公司 | 语音控制方法及显示设备 |
| CN111597433B (zh) * | 2020-04-10 | 2023-08-01 | 北京百度网讯科技有限公司 | 资源搜索方法、装置以及电子设备 |
| CN111797308B (zh) * | 2020-06-16 | 2023-11-28 | 北京达佳互联信息技术有限公司 | 一种资源推荐方法、装置、电子设备及介质 |
| CN114461843B (zh) * | 2022-01-04 | 2024-12-27 | 杭州网易云音乐科技有限公司 | 资源排序方法、装置、计算设备及存储介质 |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101833570A (zh) * | 2010-03-23 | 2010-09-15 | 深圳市五巨科技有限公司 | 一种移动终端页面推送优化的方法和装置 |
| CN102591977A (zh) * | 2012-01-04 | 2012-07-18 | 北京百度网讯科技有限公司 | 一种对搜索结果进行排序的方法以及系统 |
| CN102622417A (zh) * | 2012-02-20 | 2012-08-01 | 北京搜狗信息服务有限公司 | 对信息记录进行排序的方法和装置 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8027964B2 (en) * | 2007-07-13 | 2011-09-27 | Medio Systems, Inc. | Personalized query completion suggestion |
| CN102033952A (zh) * | 2010-12-24 | 2011-04-27 | 东莞宇龙通信科技有限公司 | 一种资源推荐方法及装置 |
-
2015
- 2015-12-07 CN CN201510887858.9A patent/CN105893427A/zh active Pending
-
2016
- 2016-07-10 WO PCT/CN2016/089525 patent/WO2017096896A1/zh not_active Ceased
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101833570A (zh) * | 2010-03-23 | 2010-09-15 | 深圳市五巨科技有限公司 | 一种移动终端页面推送优化的方法和装置 |
| CN102591977A (zh) * | 2012-01-04 | 2012-07-18 | 北京百度网讯科技有限公司 | 一种对搜索结果进行排序的方法以及系统 |
| CN102622417A (zh) * | 2012-02-20 | 2012-08-01 | 北京搜狗信息服务有限公司 | 对信息记录进行排序的方法和装置 |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113050806A (zh) * | 2019-12-27 | 2021-06-29 | 北京搜狗科技发展有限公司 | 一种信息推荐方法、装置和电子设备 |
| CN112380411A (zh) * | 2020-11-10 | 2021-02-19 | 北京达佳互联信息技术有限公司 | 敏感词处理方法、装置、电子设备、系统及存储介质 |
| CN112380411B (zh) * | 2020-11-10 | 2024-01-09 | 北京达佳互联信息技术有限公司 | 敏感词处理方法、装置、电子设备、系统及存储介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105893427A (zh) | 2016-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017096896A1 (zh) | 资源搜索方法以及服务器 | |
| US11657084B2 (en) | Correlating image annotations with foreground features | |
| US10140368B2 (en) | Method and apparatus for generating a recommendation page | |
| CN107683471B (zh) | 索引具有深度链接的网页 | |
| WO2015070673A1 (zh) | 浏览器侧进行网络搜索的方法与浏览器 | |
| US10452728B2 (en) | Method and apparatus for recommending keywords | |
| JP2019507417A (ja) | 多変数検索のためのユーザインターフェース | |
| TWI536188B (zh) | 應用程式的查找方法與系統 | |
| CN102902753A (zh) | 用于补全搜索词及建立个体兴趣模型的方法及装置 | |
| CN104462512B (zh) | 基于知识图谱的汉语信息搜索方法和装置 | |
| CN104462508A (zh) | 基于知识图谱的人物关系搜索方法和装置 | |
| US10762091B2 (en) | Interactive feedback for changes in search relevancy parameters | |
| CN103412881A (zh) | 提供搜索结果的方法及系统 | |
| US9798833B2 (en) | Accessing information content in a database platform using metadata | |
| 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 | |
| CN103020066A (zh) | 一种识别搜索需求的方法和装置 | |
| WO2016115943A1 (zh) | 一种音乐推荐方法及装置 | |
| CN110968800A (zh) | 一种信息推荐方法、装置、电子设备及可读存储介质 | |
| CN105574030A (zh) | 一种信息搜索方法及装置 | |
| WO2021135290A1 (zh) | 基于知识图谱的信息可视化方法、装置、设备及存储介质 | |
| CN104462504A (zh) | 基于搜索中提供推理过程数据的方法和装置 | |
| CN103927339A (zh) | 知识重组系统和知识重组方法 | |
| CN112825089A (zh) | 文章推荐方法、装置、设备及存储介质 | |
| CN102760127B (zh) | 基于扩展文本信息来确定资源类型的方法、装置及设备 | |
| US20170161322A1 (en) | Method and electronic device for searching resource |
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: 16872096 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: 16872096 Country of ref document: EP Kind code of ref document: A1 |