CN110309408B - Automatic searching method - Google Patents
Automatic searching method Download PDFInfo
- Publication number
- CN110309408B CN110309408B CN201810207878.0A CN201810207878A CN110309408B CN 110309408 B CN110309408 B CN 110309408B CN 201810207878 A CN201810207878 A CN 201810207878A CN 110309408 B CN110309408 B CN 110309408B
- Authority
- CN
- China
- Prior art keywords
- user
- search
- text
- scene
- abstract
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
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
The invention discloses an automatic searching method, which comprises the following steps: the first step: creating a user search intent computing system, the user search intent computing system comprising: creating a personal scene event map of the user; one or more combinations of copy, cut, paste events of user operations prior to searching; before searching, a specific event that a user contacts on the device; before searching, based on the information, voice, video and email communication content of the user; based on the user's questions; based on recent hot news events; new knowledge points; contact-based; photographing and video based on a user; searching a record library based on the personal scene of the user; guessing based on deep learning; and a second step of: acquiring a search intention of a user; and a third step of: generating search text based on the intent; fourth step: the automatic popup frame displays at least one search text, and clicks and opens a search webpage; or/and a search page that directly displays the search text.
Description
Technical Field
The invention relates to the technical field of searching methods, in particular to an automatic searching method.
Background
Search engines may be used to locate particular documents in a collection of documents. In addition, search engines may be used to locate specific keywords or phrases in documents. The search engine may use one or more indexes to locate particular documents, keywords, or phrases. In addition, search engines may perform boolean and other operations in the course of locating particular information. It is necessary to devise an automated search method that makes searches faster, more accurate, and more targeted.
Disclosure of Invention
The invention aims to provide an automatic searching method for solving the problems in the background art.
In order to achieve the above purpose, the invention provides the following technical scheme: a method of automated searching, comprising the steps of:
the first step: creating a user search intent computing system, the user search intent computing system comprising:
(1) Creating a personal scene event map of the user;
(2) Before searching, one or more of copying, cutting and pasting events operated by a user are combined;
(3) Before searching, a specific event that a user contacts on the device;
(4) Before searching, based on the information, voice, video and email communication content of the user;
(5) Based on the user's questions;
(6) Based on recent hot news events;
(7) Based on the new knowledge points;
(8) Based on the contact, contact information and address book auxiliary information; the address book auxiliary information comprises: unit, address and remark information;
(9) Photographing and video based on a user;
(10) Searching a record library based on the personal scene of the user;
(11) Probability prediction based on deep learning;
(12) Based on the content of the document edited by the user.
And a second step of: acquiring a search intention of a user;
and a third step of: generating search text based on the intent; the generating search text includes:
(1) Directly adopting the existing text;
(2) Optimizing and generating a search text on the basis of the existing text; firstly, analyzing the existing text through an entry initial URL, then, storing other entry URLs contained in the page into an entry URL library by a page link analyzer based on a defined HTML rule, then, sending a grabbing request to the analyzed URLs, sending acquired data to a knowledge data analyzer, and finally, carrying out triple storage collection processing on the filtered knowledge data by the knowledge data analyzer to achieve optimization generation of search text, wherein the knowledge data analyzer is responsible for acquiring required knowledge data in each entry page;
(3) Intelligently generating search text;
the intelligent generation of search text includes:
(1) converting pictures, audio, video and animation into texts;
(2) intelligently generating word number threshold values which accord with preset search keywords through text summarization technology on long sentences and text chapters; the automatic digest process includes three basic steps:
s1, a text analysis process: analyzing and processing long sentences and text of chapters, and identifying redundant information;
s2, selecting and generalizing text contents: identifying important information from a document, compressing a text by a method of extracting the abstract and understanding the abstract to form abstract representation; the abstract type abstract consists of fragments extracted from the original text, and the understanding type abstract is formed by reorganizing the content of the original text;
s3, conversion and generation processes of the abstract: the recombination of the original text content is realized or the abstract is generated according to semantic representation, and the continuity of the abstract is ensured;
fourth step: the automatic popup frame displays at least one search text, and clicks and opens a search webpage; or/and directly display a search page based on the search text.
Preferably, the step of creating a personal scene event map of the user includes:
step 1: presetting a multi-level scene tag library, and presetting corresponding scene logic relations, association probability calculation rules among scenes, time sequence relations among scenes and chain series expression relations;
step 2: acquiring and calculating a multi-level scene tag value of a user through a base station and a satellite positioning system used by a user mobile phone, correspondingly storing the multi-level scene tag value into a preset multi-level scene tag library, and generating a user personal scene according to a preset scene logic relationship;
step 3: based on a plurality of independent user personal scenes, calculating the association probability among the plurality of independent user personal scenes in a preset time period according to the association probability calculation rule among the scenes, and generating a user personal scene map according to the time sequence relationship and the chained serial expression relationship among the preset scenes.
Preferably, the personal scene search record library is stored as a blockchain cloud server, and comprises private blockchain equipment, public blockchain equipment, a node server, a central server, a hardware firewall, a blockchain storage module, a sending module and a receiving module; the private blockchain device and the public blockchain device store public and private event acquisition data of individuals through the blockchain storage module, the public and private event acquisition data are distributed and transmitted to the node server in real time through the transmission module, the event data acquired in real time are received through the receiving module and uploaded to the central server through the node server, and the cloud server further comprises a hardware firewall; the data in the block chain storage module is divided into a block chain backup module, a block chain isolation module and a block chain link point module, the existing block chain data is isolated into independent modules according to functions, the block chain link point module is arranged in a node server and supports cross-network access, and when the modules have running errors, the block chain backup module and the block chain isolation module update and repair the single modules.
Preferably, the deep learning based guess includes: the method is applied to a personal scene search record library, filtering and artifact removal preprocessing operations are carried out on generated data, then characteristic parameters of people, time, weather, places, sexes, professions, ages, real-time and the like of the data are extracted, PCA correlation principal component analysis is carried out, principal component characteristic parameters are extracted, then a recognition model is trained, and a support vector machine with a radial basis function as a kernel function is selected as a emotion recognition model.
Compared with the prior art, the invention has the beneficial effects that: generating a personal scene of a user by constructing a preset multi-level scene library; the number of topics is rich in the process of researching the obtained user model, so that the sparsity of the original data is solved; the personal scene data is created, so that the function recommendation is more targeted, the effective rate of the function recommendation is improved, and a large amount of large user data is quickly obtained through the function recommendation by supplementing some multidimensional new feature scene information; the information sample is comprehensive through a topological network obtained by connecting multidimensional morphological related heterogeneous information together; the method and the system effectively help the user to quickly find the interested and high-quality information, and the semantic association provides quicker, more accurate and stronger purposeful background knowledge search and has the advanced and accurate prediction effect. A particular document in the collection of documents may be precisely located. In addition, the search engine may be further used to locate particular keywords or phrases in documents and to locate particular documents, keywords or phrases using one or more indexes. Boolean and other operations are performed quickly in locating specific information. The generated data is subjected to preprocessing operations such as filtering and artifact removal, then characteristic parameters such as person, time, weather, place, sex, occupation, age and real time of the data are extracted, principal component analysis of PCA (principal component analysis) is carried out, principal component characteristic parameters are extracted, then a recognition model is trained, and a support vector machine with a radial basis function as a kernel function is selected as an emotion recognition model, so that the recommendation effect and accuracy are improved, and convergence of the data is facilitated through characteristic preprocessing of the front PCA principal component analysis.
Drawings
FIG. 1 is a flow diagram of a method of automated searching;
FIG. 2 is a schematic diagram of a process flow for optimizing and generating search text for an automated search method.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1:
referring to fig. 1 and 2, the invention provides a technical scheme: a method of automated searching, comprising the steps of:
the first step: creating a user search intent computing system, the user search intent computing system comprising:
(1) Creating a personal scene event map of the user;
(2) Before searching, one or more of copying, cutting and pasting events operated by a user are combined;
(3) Before searching, a specific event that a user contacts on the device;
(4) Before searching, based on the information, voice, video and email communication content of the user;
(5) Based on the user's questions;
(6) Based on recent hot news events;
(7) Based on the new knowledge points;
(8) Based on the contact, contact information and address book auxiliary information; the address book auxiliary information comprises: unit, address and remark information;
(9) Photographing and video based on a user;
(10) Searching a record library based on the personal scene of the user;
(11) Probability prediction based on deep learning;
(12) Based on the content of the document edited by the user.
And a second step of: acquiring a search intention of a user;
and a third step of: generating search text based on the intent; the generating search text includes:
(1) Directly adopting the existing text;
(2) Optimizing and generating a search text on the basis of the existing text; firstly, analyzing the existing text through an entry initial URL, then, storing other entry URLs contained in the page into an entry URL library by a page link analyzer based on a defined HTML rule, then, sending a grabbing request to the analyzed URLs, sending acquired data to a knowledge data analyzer, and finally, carrying out triple storage collection processing on the filtered knowledge data by the knowledge data analyzer to achieve optimization generation of search text, wherein the knowledge data analyzer is responsible for acquiring required knowledge data in each entry page;
(3) Intelligently generating search text;
the intelligent generation of search text includes:
(1) converting pictures, audio, video and animation into texts;
(2) intelligently generating word number threshold values which accord with preset search keywords through text summarization technology on long sentences and text chapters; the automatic digest process includes three basic steps:
s1, a text analysis process: analyzing and processing long sentences and text of chapters, and identifying redundant information;
s2, selecting and generalizing text contents: identifying important information from a document, compressing a text by a method of extracting the abstract and understanding the abstract to form abstract representation; the abstract type abstract consists of fragments extracted from the original text, and the understanding type abstract is formed by reorganizing the content of the original text;
s3, conversion and generation processes of the abstract: the recombination of the original text content is realized or the abstract is generated according to semantic representation, and the continuity of the abstract is ensured;
fourth step: the automatic popup frame displays at least one search text, and clicks and opens a search webpage; or/and directly display a search page based on the search text.
Preferably, the step of creating a personal scene event map of the user includes:
step 1: presetting a multi-level scene tag library, and presetting corresponding scene logic relations, association probability calculation rules among scenes, time sequence relations among scenes and chain series expression relations;
step 2: acquiring and calculating a multi-level scene tag value of a user through a base station and a satellite positioning system used by a user mobile phone, correspondingly storing the multi-level scene tag value into a preset multi-level scene tag library, and generating a user personal scene according to a preset scene logic relationship;
step 3: based on a plurality of independent user personal scenes, calculating the association probability among the plurality of independent user personal scenes in a preset time period according to the association probability calculation rule among the scenes, and generating a user personal scene map according to the time sequence relationship and the chained serial expression relationship among the preset scenes.
Preferably, the personal scene search record library is stored as a blockchain cloud server, and comprises private blockchain equipment, public blockchain equipment, a node server, a central server, a hardware firewall, a blockchain storage module, a sending module and a receiving module; the private blockchain device and the public blockchain device store public and private event acquisition data of individuals through the blockchain storage module, the public and private event acquisition data are distributed and transmitted to the node server in real time through the transmission module, the event data acquired in real time are received through the receiving module and uploaded to the central server through the node server, and the cloud server further comprises a hardware firewall; the data in the block chain storage module is divided into a block chain backup module, a block chain isolation module and a block chain link point module, the existing block chain data is isolated into independent modules according to functions, the block chain link point module is arranged in a node server and supports cross-network access, and when the modules have running errors, the block chain backup module and the block chain isolation module update and repair the single modules.
Example 2:
a method of creating a multi-level event and scene graph feature comprising:
step 1: presetting a multi-level event tag library and a multi-level scene tag library, and presetting corresponding scene event logic relations, scene logic relations, association probability calculation rules among scenes, association probability calculation rules among events, a scene map establishing method and an event map establishing method; the multi-level event labels comprise a first-level event label and a multi-level subordinate event label; the primary event tag includes: a subject event and a guest event; the main event refers to an event occurring at a computer terminal; the object event refers to an event which is detected and identified by the computer terminal through a visual auditory identification system and occurs outside;
specifically, in this embodiment, the following design is made for the preset multi-level event tag library:
the main event refers to an event occurring at the computer terminal, and the main event is a primary event tag, which comprises a secondary event tag: chat, phone call, use APP, browse; use of APP in a secondary event tag comprises a tertiary event tag: weChat is used.
The object event is an externally-generated event detected and identified by the computer terminal through the visual and auditory identification system, and is a primary event tag; it contains a secondary event tag: meeting, shopping, watching a movie, singing; the watch movie in the secondary event tag contains a tertiary event tag: viewing science fiction pieces; the science fiction pieces in the three-level event tag comprise four-level event tags: see avermectin.
Step 2: the computer terminal acquires and/or calculates a multi-level scene tag value of a user through visual auditory recognition, correspondingly stores the multi-level scene tag value into a preset multi-level scene tag library, generates a personal scene of the user according to a preset scene logic relation, calculates association probabilities among a plurality of independent user personal scenes in a preset time period according to an association probability calculation rule among the scenes, and generates a user personal scene map according to a preset scene map establishment method; the multi-level scene tag comprises a first-level scene tag and a multi-level subordinate scene tag, wherein the first-level scene tag comprises four tags of people, time, places and weather; the scene graph establishing method is characterized in that a plurality of independent user personal scenes within a certain time are established, a chained serial expression graph is established based on a forward time sequence relation and a reverse time sequence relation, and the probability of association occurrence is marked on the edges between adjacent user personal scenes; the preset time period can be customized by a user;
the timing relationship includes: before: a scene occurs before another scene; after: a scene occurs after another scene; includes: one scene contains another scene; is_incorporated: one scene is contained by another scene; during: a scene maintains a state for a period of time; simulte-ous: simultaneously; iafter: the scenes occur immediately next to one another and they do not overlap, uninterrupted; the Ibe-form scene occurs before another scene and they do not overlap, uninterrupted; lIaaentity: representing the same scene; begins: one scene start causes another scene to start; ends: one scene end results in the other scene end; beginby: one scene starts with the start of the other scene, as opposed to Begins; end_by: one scene Ends with the end of the other scene, as opposed to Ends;
the calculation method of the time sequence relation comprises the following steps: when a scene is acquired, the starting time point and the ending time point of the scene are acquired, and the time sequence relation is determined through time calculation; the method for determining the forward and reverse time sequence relationship comprises the following steps: representing the adjacent occurrence relationship of the scene on the time axis according to 13 time sequence relationships obtained by the above calculation method; the expression of the chain tandem refers to: based on the forward and reverse time sequence relation, establishing a chained serial expression diagram among a plurality of independent user personal scenes, and labeling probability information on adjacent edges;
example 3
A method of time-series creation of a multi-dimensional feature map of a personal scene, comprising:
step 1: presetting a multi-level scene tag library, presetting corresponding scene logic relations, association probability calculation rules among scenes, timing sequence relations among scenes and chain series expression relations, wherein the multi-level scene tags comprise primary scene tags and multi-level subordinate scene tags, and the primary scene tags comprise four tags of people, time, places and weather;
specifically, in this embodiment, the following design is made for the preset multi-level scene tag library:
1. the person is a first-level scene tag, and corresponding tag items are preset; it contains a secondary scene tag: gender/occupation/marital status/age/health status/mood, presetting corresponding tag entries; the doctor in the secondary scene tag contains a tertiary scene tag: the job position/position is preset, corresponding label entries are preset, and all preset label entries are arranged according to the logical affiliation to build a library.
2. Time is a primary scene tag; it contains a secondary scene tag: season/annual status/birthday/workday; the workday in the secondary scene tag contains a tertiary scene tag: a morning shift-in period/a afternoon shift-in period; the morning shift period in the three-level scene tag contains four-level scene tags: presetting corresponding tag entries in the period of just going to work/the period of fast going to work, and arranging and building libraries according to the logic affiliation of all preset tag entries.
3. The location is a primary scene tag; it contains a secondary scene tag: home/company/restaurant/movie theater; restaurants in the secondary scene tags contain the tertiary scene tags: western restaurant/middle restaurant; the Chinese hall in the three-level scene tag contains four-level scene tags: and presetting corresponding tag entries in the Xiang-ban/Chuan-ban, and arranging and building all preset tag entries according to the logical affiliation.
4. Weather is a primary scene tag; it contains a secondary scene tag: good weather/bad weather/weather disasters; bad weather in the secondary scene tag comprises a tertiary scene tag: weather/temperature/wind power, presetting corresponding label entries, and arranging and building libraries according to logic affiliations of all preset label entries.
Step 2, acquiring and/or calculating a multi-level scene tag value of a user through a base station or/and a satellite positioning system used by a user mobile phone, correspondingly storing the multi-level scene tag value into a preset multi-level scene tag library, generating a user personal scene according to a preset scene logic relation, acquiring the multi-level scene tag value of the user through a preset user portrait database or/and real-time calculation, automatically acquiring the multi-level scene tag value of the time through a system, dynamically acquiring the multi-level scene tag value of the place through the base station or/and the satellite positioning system used by the user mobile phone, and automatically acquiring the multi-level scene tag value of weather through the system, wherein the scene logic relation is a rule for composing a personal scene describing the user after the scene tag is defined among all scene tags of the multi-level scene tag library to acquire the corresponding scene tag value;
example 4
Preferably, the deep learning based guess includes: the method is applied to a personal scene search record library, filtering and artifact removal preprocessing operations are carried out on generated data, then characteristic parameters of people, time, weather, places, sexes, professions, ages, real-time and the like of the data are extracted, PCA correlation principal component analysis is carried out, principal component characteristic parameters are extracted, then a recognition model is trained, and a support vector machine with a radial basis function as a kernel function is selected as a emotion recognition model.
Working principle: generating a personal scene of a user by constructing a preset multi-level scene library; the number of topics is rich in the process of researching the obtained user model, so that the sparsity of the original data is solved; the personal scene data is created, so that the function recommendation is more targeted, the effective rate of the function recommendation is improved, and a large amount of large user data is quickly obtained through the function recommendation by supplementing some multidimensional new feature scene information; the information sample is comprehensive through a topological network obtained by connecting multidimensional morphological related heterogeneous information together; the method and the system effectively help the user to quickly find the interested and high-quality information, and the semantic association provides quicker, more accurate and stronger purposeful background knowledge search and has the advanced and accurate prediction effect. A particular document in the collection of documents may be precisely located. In addition, the search engine may be further used to locate particular keywords or phrases in documents and to locate particular documents, keywords or phrases using one or more indexes. Boolean and other operations are performed quickly in locating specific information. The generated data is subjected to preprocessing operations such as filtering and artifact removal, then characteristic parameters such as person, time, weather, place, sex, occupation, age and real time of the data are extracted, principal component analysis of PCA (principal component analysis) is carried out, principal component characteristic parameters are extracted, then a recognition model is trained, and a support vector machine with a radial basis function as a kernel function is selected as an emotion recognition model, so that the recommendation effect and accuracy are improved, and convergence of the data is facilitated through characteristic preprocessing of the front PCA principal component analysis.
The automatic searching method is faster, more accurate and stronger in purpose, and has the advanced and accurate prediction effect.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (2)
1. A method of automated searching, characterized by: the method comprises the following steps:
the first step: creating a user search intent computing system, the user search intent computing system comprising:
(1) Creating a personal scene event map of the user;
(2) Before searching, one or more of copying, cutting and pasting events operated by a user are combined;
(3) Before searching, a specific event that a user contacts on the device;
(4) Before searching, based on the information, voice, video and email communication content of the user;
(5) Based on the user's questions;
(6) Based on recent hot news events;
(7) Based on the new knowledge points;
(8) Based on the contact, contact information and address book auxiliary information; the address book auxiliary information comprises: unit, address and remark information;
(9) Photographing and video based on a user;
(10) Searching a record library based on the personal scene of the user;
(11) Probability prediction based on deep learning;
(12) Based on the document content edited by the user;
and a second step of: acquiring a search intention of a user;
and a third step of: generating search text based on the intent; the generating search text includes:
(1) Directly adopting the existing text;
(2) Optimizing and generating a search text on the basis of the existing text; firstly, analyzing the existing text through an entry initial URL, then, storing other entry URLs contained in the page into an entry URL library by a page link analyzer based on a defined HTML rule, then, sending a grabbing request to the analyzed URLs, sending acquired data to a knowledge data analyzer, and finally, carrying out triple storage collection processing on the filtered knowledge data by the knowledge data analyzer to achieve optimization generation of search text, wherein the knowledge data analyzer is responsible for acquiring required knowledge data in each entry page;
(3) Intelligently generating search text;
the intelligent generation of search text includes:
(1) converting pictures, audio, video and animation into texts;
(2) intelligently generating word number threshold values which accord with preset search keywords through text summarization technology on long sentences and text chapters; the automatic digest process includes three basic steps:
s1, a text analysis process: analyzing and processing long sentences and text of chapters, and identifying redundant information;
s2, selecting and generalizing text contents: identifying important information from a document, compressing a text by a method of extracting the abstract and understanding the abstract to form abstract representation; the abstract type abstract consists of fragments extracted from the original text, and the understanding type abstract is formed by reorganizing the content of the original text;
s3, conversion and generation processes of the abstract: the recombination of the original text content is realized or the abstract is generated according to semantic representation, and the continuity of the abstract is ensured;
fourth step: the automatic popup frame displays at least one search text, and clicks and opens a search webpage; or/and directly displaying a search page based on the search text;
the step of creating a personal scene event map of a user comprises:
step 1: presetting a multi-level scene tag library, and presetting corresponding scene logic relations, association probability calculation rules among scenes, time sequence relations among scenes and chain series expression relations;
step 2: acquiring and calculating a multi-level scene tag value of a user through a base station and a satellite positioning system used by a user mobile phone, correspondingly storing the multi-level scene tag value into a preset multi-level scene tag library, and generating a user personal scene according to a preset scene logic relationship;
step 3: based on a plurality of independent user personal scenes, calculating the association probability among the plurality of independent user personal scenes in a preset time period according to the association probability calculation rule among the scenes, and generating a user personal scene map according to the time sequence relationship and the chained serial expression relationship among the preset scenes;
the personal scene search record library is stored as a blockchain cloud server and comprises private blockchain equipment, public blockchain equipment, a node server, a central server, a hardware firewall, a blockchain storage module, a sending module and a receiving module; the private blockchain device and the public blockchain device store public and private event acquisition data of individuals through the blockchain storage module, the public and private event acquisition data are distributed and transmitted to the node server in real time through the transmission module, the event data acquired in real time are received through the receiving module and uploaded to the central server through the node server, and the cloud server further comprises a hardware firewall; the data in the block chain storage module is divided into a block chain backup module, a block chain isolation module and a block chain link point module, the existing block chain data is isolated into independent modules according to functions, the block chain link point module is arranged in a node server and supports cross-network access, and when the modules have running errors, the block chain backup module and the block chain isolation module update and repair the single modules.
2. A method of automated searching as claimed in claim 1, wherein: the deep learning based guess includes: the method is applied to a personal scene search record library, filtering and artifact removal preprocessing operations are carried out on generated data, then real-time characteristic parameters of people, time, weather, places, sexes, professions and ages of the data are extracted, PCA correlation principal component analysis is carried out, principal component characteristic parameters are extracted, then a recognition model is trained, and a support vector machine with a radial basis function as a kernel function is selected as a emotion recognition model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810207878.0A CN110309408B (en) | 2018-03-09 | 2018-03-09 | Automatic searching method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810207878.0A CN110309408B (en) | 2018-03-09 | 2018-03-09 | Automatic searching method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110309408A CN110309408A (en) | 2019-10-08 |
| CN110309408B true CN110309408B (en) | 2023-07-14 |
Family
ID=68073305
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810207878.0A Active CN110309408B (en) | 2018-03-09 | 2018-03-09 | Automatic searching method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110309408B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102591992A (en) * | 2012-02-15 | 2012-07-18 | 苏州亚新丰信息技术有限公司 | Webpage classification identifying system and method based on vertical search and focused crawler technology |
| CN104809173A (en) * | 2015-04-10 | 2015-07-29 | 百度在线网络技术（北京）有限公司 | Search result processing method and device |
| CN105659234A (en) * | 2013-10-23 | 2016-06-08 | 微软技术许可有限责任公司 | Pervasive search architecture |
| CN106709072A (en) * | 2017-02-13 | 2017-05-24 | 长沙军鸽软件有限公司 | Method of obtaining intelligent conversation reply content based on shared corpora |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7512601B2 (en) * | 2005-01-18 | 2009-03-31 | Microsoft Corporation | Systems and methods that enable search engines to present relevant snippets |
| GB0608323D0 (en) * | 2006-04-27 | 2006-06-07 | Soft Image Systems Ltd | Codifying & reusing expertise in personal and organisation transformation |
-
2018
- 2018-03-09 CN CN201810207878.0A patent/CN110309408B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102591992A (en) * | 2012-02-15 | 2012-07-18 | 苏州亚新丰信息技术有限公司 | Webpage classification identifying system and method based on vertical search and focused crawler technology |
| CN105659234A (en) * | 2013-10-23 | 2016-06-08 | 微软技术许可有限责任公司 | Pervasive search architecture |
| CN104809173A (en) * | 2015-04-10 | 2015-07-29 | 百度在线网络技术（北京）有限公司 | Search result processing method and device |
| CN106709072A (en) * | 2017-02-13 | 2017-05-24 | 长沙军鸽软件有限公司 | Method of obtaining intelligent conversation reply content based on shared corpora |
Non-Patent Citations (2)
| Title |
|---|
| easy contextual intent prediction and slot detection;Bhargava等;《2013 IEEE international conference onacoustics speech and signal processing》;20131231;8337-8341 * |
| 人机对话系统综述;车万翔等;《人工智能》;20180210(第2期);76-82 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110309408A (en) | 2019-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3316586B1 (en) | Method and device for extracting keywords from bullet screen information and pushing related information including a web link | |
| CN100405371C (en) | Method and system for abstracting new word | |
| US8099406B2 (en) | Method for human editing of information in search results | |
| WO2016150083A1 (en) | Information input method and apparatus | |
| CN111694965B (en) | Image scene retrieval system and method based on multi-mode knowledge graph | |
| CN108984650B (en) | Computer-readable recording medium and computer device | |
| CN104850546B (en) | Display method and system of mobile media information | |
| WO2017036047A1 (en) | Information extraction method and information extraction device | |
| KR20160055930A (en) | Systems and methods for actively composing content for use in continuous social communication | |
| CN102737022B (en) | Method and device for acquiring and searching relevant knowledge information | |
| CN102456054B (en) | A kind of searching method and system | |
| CN101369281A (en) | Retrieval method based on video abstract metadata | |
| CN103399855B (en) | Behavior intention determining method and device based on multiple data sources | |
| CN108133058B (en) | Video retrieval method | |
| CN110633406A (en) | Event topic generation method and device, storage medium and terminal equipment | |
| CN111191133B (en) | Service search processing method, device and equipment | |
| CN111491209A (en) | Video cover determining method and device, electronic equipment and storage medium | |
| CN114881685A (en) | Advertisement delivery method, device, electronic device and storage medium | |
| CN109359240A (en) | A method of automation search | |
| US20190012384A1 (en) | Method And Apparatus For Pushing Information | |
| CN110309408B (en) | Automatic searching method | |
| US20220027419A1 (en) | Smart search and recommendation method for content, storage medium, and terminal | |
| CN111831885B (en) | Internet information retrieval system and method | |
| CN110895538A (en) | Data retrieval method, device, storage medium and processor | |
| CN104504070B (en) | A kind of method and apparatus of search |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |