CN107908654B - Knowledge base-based recommendation method, system and device - Google Patents
Knowledge base-based recommendation method, system and device Download PDFInfo
- Publication number
- CN107908654B CN107908654B CN201710947514.1A CN201710947514A CN107908654B CN 107908654 B CN107908654 B CN 107908654B CN 201710947514 A CN201710947514 A CN 201710947514A CN 107908654 B CN107908654 B CN 107908654B
- Authority
- CN
- China
- Prior art keywords
- word
- label
- information
- encyclopedic
- analyzed
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
Abstract
The invention discloses a recommendation method, a system and a device based on a knowledge base, which comprises the steps of reading information reading records of a user; obtaining a label of information in the information reading record according to the information reading record to obtain a label to be analyzed; and analyzing and generating a recommendation list according to the label to be analyzed. The invention obtains the relation chain between words by analyzing the information reading record of the user in real time and analyzing in the knowledge base and the encyclopedia platform search base, thereby better processing the problems of priority and relevance in the information recommendation process, effectively improving the recommendation accuracy rate and being widely applied to information recommendation application.
Description
Technical Field
The invention relates to the technical field of data mining, in particular to a recommendation method, a recommendation system and a recommendation device based on a knowledge base.
Background
In the recommendation theory, a theoretical method is called a recommendation method based on a knowledge base, which mainly solves the problem of strong and weak relationships between words, no open interface or system is available for providing the service at present, a set of knowledge base network needs to be built and maintained for using the recommendation method, but the relationship chain between words in the existing knowledge base network is weak, so that the recommendation method is not well suitable for information recommendation.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a recommendation method, system and apparatus based on a knowledge base according to vocabulary association.
The technical scheme adopted by the invention is as follows:
a recommendation method based on a knowledge base comprises the following steps:
reading the information reading record of the user;
obtaining a label of information in the information reading record according to the information reading record to obtain a label to be analyzed;
and analyzing and generating a recommendation list according to the label to be analyzed.
As a further improvement of the knowledge base-based recommendation method, the analyzing and generating a recommendation list according to the tag to be analyzed specifically includes:
obtaining a corresponding encyclopedia search associated word list according to the label to be analyzed;
performing intersection processing on the encyclopedic search associated word list of each label to be analyzed to obtain encyclopedic intersection associated words;
reading an encyclopedic search related word list corresponding to the encyclopedic intersection related word, judging whether a label to be analyzed exists in the encyclopedic search related word list corresponding to the encyclopedic intersection related word, and if so, superposing the word frequency of the existing label to be analyzed and the word frequency of the corresponding encyclopedic intersection related word to be used as the word frequency of the encyclopedic intersection related word; otherwise, eliminating an encyclopedic search related word list corresponding to the encyclopedic intersection related word, wherein the word frequency represents the word frequency in the encyclopedic search related word list;
and performing word frequency superposition on the labels in the information as weights, sequencing the labels according to the weights of the information from high to low, and adding the information into a temporary information association pool to obtain a recommendation list.
As a further improvement of the recommendation method based on the knowledge base, the step of analyzing and generating the recommendation list according to the tag to be analyzed further includes:
if the information quantity in the recommendation list is less than a preset information threshold value, acquiring a related word list and a related combination list corresponding to the label to be analyzed;
according to the associated word list and the associated combination list corresponding to the label to be analyzed, counting the occurrence ratio of associated words and the occurrence ratio of associated word combination;
and according to a preset rule, the obtained occurrence proportion and combination proportion of the related words are used as weights of the corresponding information to sort the information, and the information is added into a temporary information association pool to obtain a recommendation list.
As a further improvement of the knowledge base-based recommendation method, the step of obtaining a corresponding encyclopedic search related word list according to the tag to be analyzed specifically includes:
searching in various encyclopedia platform search libraries according to the labels to be analyzed to obtain search results of various encyclopedia platforms;
performing word segmentation on the obtained search results of all encyclopedia platforms to obtain words in the search results, and further obtaining words except the tags to be analyzed in the search results as search associated words;
carrying out Hash value calculation on the search associated words, and respectively counting the word frequency of the search associated words of each encyclopedic platform;
performing intersection processing on the types of the search associated words of all encyclopedic platforms to obtain search associated words after intersection processing is performed;
and according to the search associated words after intersection, performing average calculation on the word frequency of the search associated words, and further establishing and obtaining a corresponding encyclopedic search associated word list.
As a further improvement of the knowledge base-based recommendation method, the acquiring of the associated word list and the associated combination list corresponding to the to-be-analyzed tag specifically includes:
searching in a knowledge base to obtain a corresponding information article according to the label to be analyzed;
performing word segmentation processing on the obtained information article to generate a label of the obtained information article, and further obtaining the label except the label to be analyzed in the information article as a related word;
carrying out Hash value calculation on the label to be analyzed and the associated word, and counting the word frequency of the label to be analyzed and the associated word, so as to establish and obtain a corresponding associated word list;
and combining every two associated words, and counting the occurrence times of the combined associated words so as to establish and obtain a corresponding associated combination list.
As a further improvement of the knowledge base-based recommendation method, the formula for calculating the occurrence ratio of the associated word is as follows:
the appearance ratio of the relevant words is the word frequency of the relevant words/the word frequency of the label to be analyzed.
As a further improvement of the knowledge base-based recommendation method, the formula for calculating the occurrence ratio of the related word combination is as follows:
and the appearance proportion of the associated word combination is the appearance times of the associated word combination/the word frequency of the label to be analyzed.
The other technical scheme adopted by the invention is as follows:
a knowledge-base based recommendation system comprising:
a reading unit for reading the information reading record of the user;
the label obtaining unit is used for obtaining a label of the information in the information reading record according to the information reading record to obtain a label to be analyzed;
and the list generating unit is used for analyzing and generating a recommendation list according to the label to be analyzed.
The invention adopts another technical scheme that:
a knowledge-base-based recommendation apparatus comprising:
a memory for storing a program;
a processor for executing the program, the program causing the processor to perform the knowledge-base based recommendation method.
The invention has the beneficial effects that:
the recommendation method, the system and the device based on the knowledge base analyze the information reading records of the user in real time and analyze the information reading records in the knowledge base and the encyclopedic platform search base to obtain the relation chain between the words, so that the problems of priority and relevance can be better solved in the information recommendation process, and the recommendation accuracy is effectively improved.
Drawings
FIG. 1 is a flow chart of the steps of a knowledge-base based recommendation method of the present invention;
FIG. 2 is a block diagram of modules of a knowledge-base based recommendation system of the present invention.
Detailed Description
The following further describes embodiments of the present invention with reference to the accompanying drawings:
referring to fig. 1, the invention relates to a recommendation method based on a knowledge base, comprising the following steps:
reading the information reading record of the user;
obtaining a label of information in the information reading record according to the information reading record to obtain a label to be analyzed;
and analyzing and generating a recommendation list according to the label to be analyzed.
Further as a preferred embodiment, the analyzing and generating a recommendation list according to the tag to be analyzed specifically includes:
obtaining a corresponding encyclopedia search associated word list according to the label to be analyzed;
performing intersection processing on the encyclopedic search associated word list of each label to be analyzed to obtain encyclopedic intersection associated words;
reading an encyclopedic search related word list corresponding to the encyclopedic intersection related word, judging whether a label to be analyzed exists in the encyclopedic search related word list corresponding to the encyclopedic intersection related word, and if so, superposing the word frequency of the existing label to be analyzed and the word frequency of the corresponding encyclopedic intersection related word to be used as the word frequency of the encyclopedic intersection related word; otherwise, eliminating an encyclopedic search related word list corresponding to the encyclopedic intersection related word, wherein the word frequency represents the word frequency in the encyclopedic search related word list;
and performing word frequency superposition on the labels in the information as weights, sequencing the labels according to the weights of the information from high to low, and adding the information into a temporary information association pool to obtain a recommendation list.
Further as a preferred embodiment, the analyzing and generating a recommendation list according to the label to be analyzed further includes:
if the information quantity in the recommendation list is less than a preset information threshold value, acquiring a related word list and a related combination list corresponding to the label to be analyzed;
according to the associated word list and the associated combination list corresponding to the label to be analyzed, counting the occurrence ratio of associated words and the occurrence ratio of associated word combination;
and according to a preset rule, the obtained occurrence proportion and combination proportion of the related words are used as weights of the corresponding information to sort the information, and the information is added into a temporary information association pool to obtain a recommendation list.
Further as a preferred embodiment, the obtaining of the corresponding encyclopedic search related word list according to the tag to be analyzed specifically includes:
searching in various encyclopedia platform search libraries according to the labels to be analyzed to obtain search results of various encyclopedia platforms;
performing word segmentation on the obtained search results of all encyclopedia platforms to obtain words in the search results, and further obtaining words except the tags to be analyzed in the search results as search associated words;
carrying out Hash value calculation on the search associated words, and respectively counting the word frequency of the search associated words of each encyclopedic platform;
performing intersection processing on the types of the search associated words of all encyclopedic platforms to obtain search associated words after intersection processing is performed;
and according to the search associated words after intersection, performing average calculation on the word frequency of the search associated words, and further establishing and obtaining a corresponding encyclopedic search associated word list.
Further as a preferred embodiment, the acquiring a related word list and a related combination list corresponding to a tag to be analyzed specifically includes:
searching in a knowledge base to obtain a corresponding information article according to the label to be analyzed;
performing word segmentation processing on the obtained information article to generate a label of the obtained information article, and further obtaining the label except the label to be analyzed in the information article as a related word;
carrying out Hash value calculation on the label to be analyzed and the associated word, and counting the word frequency of the label to be analyzed and the associated word, so as to establish and obtain a corresponding associated word list;
and combining every two associated words, and counting the occurrence times of the combined associated words so as to establish and obtain a corresponding associated combination list.
Further preferably, the formula for calculating the occurrence ratio of the related word is as follows:
the appearance ratio of the relevant words is the word frequency of the relevant words/the word frequency of the label to be analyzed.
As a further improvement of the knowledge base-based recommendation method, the formula for calculating the occurrence ratio of the related word combination is as follows:
and the appearance proportion of the associated word combination is the appearance times of the associated word combination/the word frequency of the label to be analyzed.
In the embodiment of the invention, the information label is obtained by manually marking the label or performing word segmentation and word frequency statistics on the article content by a machine according to a word bank. One piece of information has tags not more than 6, and when the information is labeled, the result is sent to the knowledge base maintenance module, and the number corresponding to the information is stored in the information list corresponding to the tags. And the knowledge base module judges whether the label is in the mysql word stock, if not, the label is inserted, the hash value of the word is calculated, the value corresponding to the key of the corresponding counting times of the hash value is added with 1, the hash value of the word except the word is added to the association list, the times are added with 1, the word except the word is combined in pairs and inserted into the association combination list, and the times are added with 1. Example (c): the labeling results received were sports, basketball, nba, yaoming, assuming hash corresponds to a, b, c, d. For yaoming, there are: b, inserting a, b and c into the associated word list corresponding to d in the ssdb, and adding 1 to the corresponding number; the association combination list is inserted into a _ b, a _ c and b _ c, and the corresponding number is increased by 1.
In this embodiment, words in the word stock are searched one by timing each encyclopedia platform (wiki, google, Baidu, etc.), the search results are cut, intersections are screened out, the word frequency is averaged, the results are stored in an encyclopedia search related word list, and meanwhile, newly appeared words are inserted into the word stock of mysql. Example (c): the search yaoming (hash value is d), the search word-cutting result of platform 1 is: china 35, basketball 24, NBA 29 and CBA 12, the search word cutting effect of the platform 2 is as follows: basketball 20, NBA 21 and CBA 10, the screening result is as follows: the corresponding hash values are calculated for the basketballs 22((24+20)/2 ═ 22), NBA 25, CBA 11, and the hash and quantity are inserted.
In this embodiment, the information reading record of the user is read, the sent and read records corresponding to the user are obtained, and the first 20 pieces of the recently read information are taken to obtain the tags of the 20 pieces of information. The following operations are performed by following the 20 pieces of information:
events are strongly correlated: assume that the first tag attribute read recently is: NBA, CBA and Yaoming sequentially go to corresponding information lists, sent filtering and read filtering are carried out on the lists, and information hitting all labels is screened out. Inserting into corresponding temporary information correlation pool, and ending with 10 pieces.
External knowledge association: if the number of the information pieces obtained is less than 0, the information obtained from the filtering only hits on a label of the Yaomin, CBA or NBA, and the association list is needed. Firstly, reading an encyclopedia search related word list of yaoming, NBA and CBA, and taking the intersection of the three lists to obtain: basketball, etc., using the intersection of the two lists to obtain: rocket team, easy-to-build union, Madi, celebrity hall, basketball, etc.; then the top 10 of the encyclopedic search related word list corresponding to the intersection is read (sorted by word frequency from high to low), whether the Yaoming, the NBA or the CBA is in the encyclopedic search related word list or not is judged, the coincidence is reserved, the word frequency superposition is used as the sorting, and the non-coincidence is eliminated. Example (c): the top 10 encyclopedia search associated word lists of basketball may be eliminated without yaoming, NBA and CBA; the first 10 encyclopedia search associated word lists of the rocket team are reserved and can also comprise the yaoming, so the word frequency of the NBA is summed with the word frequency of the yaoming; the mdie tag may also be retained because the top 10 may all have yaoming, NBA, CBA, so the result of the word frequency superposition and use as a ranking may be: mandy > rocket team. Thus, a sequential associated word list is obtained, and then the tag information list of the information is sequentially circulated, and the news is sequenced by using the word frequency superposition and the weighting value. Such as: the information obtained from the information list of Yaoming, the weight of the information with labels of Yaoming, Maydi and rocket team is greater than the weight of the information with labels of Yaoming and Maydi, and the weight of the information with labels of Yaoming and Maydi is greater than the weight of the information with labels of Yaoming and rocket team. Inserting into corresponding temporary information correlation pool, and ending with 10 pieces.
Third, internal knowledge association: if the 10 pieces of related information are not obtained through the second step, screening is carried out according to the combined related word list and the related word list of the labels. Acquiring the times of the occurrence of the tags, the associated word list and the combined associated word list, and counting the occurrence proportion of the associated words and the occurrence proportion of the associated word combination, wherein the formula is as follows: the occurrence ratio of the relevant words is the word frequency of the relevant words/the word frequency of the label to be analyzed; and the appearance proportion of the associated word combination is the appearance times of the associated word combination/the word frequency of the label to be analyzed.
Such as: the occurrence frequency of the Yaoming is 40, the occurrence frequency of the NBA in a Yaoming related word list is 35, the famous hall is 8, and the proportion is NBA 35/40-0.875; celebrity hall 8/40 is 0.2. In the combined related word list of yaoming, if the number of NBA _ mdis combinations is 5, the ratio is 5/40-0.125.
Sorting the occurrence ratios from large to small, firstly considering the combined associated word list, screening out the information with the combinations in the information list, and sorting the information with the ratio as the weight. If the combined related words are not hit, the information of the hit related words is screened out by considering the related word list, and the information is sorted by taking the corresponding proportion as the weight. Inserting into corresponding temporary information correlation pool, and ending with 10 pieces.
If the associated information is still not obtained through the 3 steps, no further processing is performed.
In this embodiment, 20 pieces of recently read information may be circulated, the three steps of operations may be performed, and finally, each piece of information generates a related information list corresponding to each piece of information, the number of each list is less than or equal to 10, the information in the related information list corresponding to the number 1 is numbered 1, 21, 41, and.
The recommendation method based on the knowledge base comprises the following steps: the probability of an event (i.e., the probability of the event statistics co-occurring) and the information obtained from the outside are similar to what is called "experience" and "book knowledge" in life. The invention integrates the two, the encyclopedic related words have higher priority than the related words of internal information statistics, and the deviation caused by small information quantity or uneven proportion of the information base (the information base is formed after the system is built, the information base possibly only contains information in a period of time, and the quantity is limited, so that the correlation obtained according to the information statistics of the system is incomplete). For example: the information base has A, C labels with more information and no A, B labels with information, which cannot fully indicate that A, B is weaker than A, C, but it is possible that the relationship between A and B has stronger relationship in the field, and only the information collection speed, quantity and angle are not up to the standard. In this case, if there is an information item labeled with A, B, the priority recommendation associated with a is more desirable in the actual recommendation, and this problem can be solved based on the encyclopedia related word. Since the encyclopedia platform results in a summary of words, which may not include some information for a short time, or the encyclopedia platform requires a period of maintenance and update, so that some words cannot be associated, it also needs to consider some associated information obtained from the statistics of the internal database for recommendation.
Referring to fig. 2, the invention relates to a recommendation system based on knowledge base, comprising:
a reading unit for reading the information reading record of the user;
the label obtaining unit is used for obtaining a label of the information in the information reading record according to the information reading record to obtain a label to be analyzed;
and the list generating unit is used for analyzing and generating a recommendation list according to the label to be analyzed.
The invention relates to a recommendation device based on a knowledge base, which comprises:
a memory for storing a program;
a processor for executing the program, the program causing the processor to perform the knowledge-base based recommendation method.
From the above, the recommendation method, system and device based on the knowledge base, provided by the invention, can be used for analyzing the information reading records of the user in real time and analyzing the information reading records in the knowledge base and the encyclopedic platform search base to obtain the relationship chain between the words, so that the problems of priority and relevance can be better solved in the information recommendation process, and the recommendation accuracy is effectively improved.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (8)
1. A recommendation method based on a knowledge base is characterized by comprising the following steps:
reading the information reading record of the user;
obtaining a label of information in the information reading record according to the information reading record to obtain a label to be analyzed;
analyzing and generating a recommendation list according to the label to be analyzed;
wherein, the step of generating a recommendation list by analyzing according to the label to be analyzed specifically comprises:
obtaining a corresponding encyclopedia search associated word list according to the label to be analyzed;
performing intersection processing on the encyclopedic search associated word list of each label to be analyzed to obtain encyclopedic intersection associated words;
reading an encyclopedic search related word list corresponding to the encyclopedic intersection related word, judging whether a label to be analyzed exists in the encyclopedic search related word list corresponding to the encyclopedic intersection related word, and if so, superposing the word frequency of the existing label to be analyzed and the word frequency of the corresponding encyclopedic intersection related word to be used as the word frequency of the encyclopedic intersection related word; otherwise, eliminating an encyclopedic search related word list corresponding to the encyclopedic intersection related word, wherein the word frequency represents the word frequency in the encyclopedic search related word list;
stacking word frequencies of labels in the information to be used as weights, sequencing the labels according to the weights of the information from high to low, and adding the information into a temporary information association pool to obtain a recommendation list;
the encyclopedic search associated word list is a list obtained by searching encyclopedic or knowledge base according to associated words.
2. The knowledge-base-based recommendation method of claim 1, wherein: the step of generating a recommendation list by analyzing according to the label to be analyzed further comprises: if the information quantity in the recommendation list is less than a preset information threshold value, acquiring a related word list and a related combination list corresponding to the label to be analyzed;
according to the associated word list and the associated combination list corresponding to the label to be analyzed, counting the occurrence ratio of associated words and the occurrence ratio of associated word combination;
and according to a preset rule, the obtained occurrence proportion and combination proportion of the related words are used as weights of the corresponding information to sort the information, and the information is added into a temporary information association pool to obtain a recommendation list.
3. The knowledge-base-based recommendation method of claim 1, wherein: the step of obtaining a corresponding encyclopedic search associated word list according to the label to be analyzed specifically comprises the following steps:
searching in various encyclopedia platform search libraries according to the labels to be analyzed to obtain search results of various encyclopedia platforms;
performing word segmentation on the obtained search results of all encyclopedia platforms to obtain words in the search results, and further obtaining words except the tags to be analyzed in the search results as search associated words;
carrying out Hash value calculation on the search associated words, and respectively counting the word frequency of the search associated words of each encyclopedic platform;
performing intersection processing on the types of the search associated words of all encyclopedic platforms to obtain search associated words after intersection processing is performed;
and according to the search associated words after intersection, performing average calculation on the word frequency of the search associated words, and further establishing and obtaining a corresponding encyclopedic search associated word list.
4. The knowledge-base-based recommendation method of claim 2, wherein: the method for acquiring the associated word list and the associated combination list corresponding to the label to be analyzed specifically comprises the following steps:
searching in a knowledge base to obtain a corresponding information article according to the label to be analyzed;
performing word segmentation processing on the obtained information article to generate a label of the obtained information article, and further obtaining the label except the label to be analyzed in the information article as a related word;
carrying out Hash value calculation on the label to be analyzed and the associated word, and counting the word frequency of the label to be analyzed and the associated word, so as to establish and obtain a corresponding associated word list;
and combining every two associated words, and counting the occurrence times of the combined associated words so as to establish and obtain a corresponding associated combination list.
5. The knowledge-base-based recommendation method of claim 2, wherein: the calculation formula of the occurrence proportion of the associated words is as follows:
the appearance ratio of the relevant words is the word frequency of the relevant words/the word frequency of the label to be analyzed.
6. The knowledge-base-based recommendation method of claim 2, wherein: the calculation formula of the occurrence proportion of the relevant word combination is as follows:
and the appearance proportion of the associated word combination is the appearance times of the associated word combination/the word frequency of the label to be analyzed.
7. A knowledge-base-based recommendation system, comprising:
a reading unit for reading the information reading record of the user;
the label obtaining unit is used for obtaining a label of the information in the information reading record according to the information reading record to obtain a label to be analyzed;
the list generating unit is used for analyzing and generating a recommendation list according to the label to be analyzed;
wherein, the step of generating a recommendation list by analyzing according to the label to be analyzed specifically comprises:
obtaining a corresponding encyclopedia search associated word list according to the label to be analyzed;
performing intersection processing on the encyclopedic search associated word list of each label to be analyzed to obtain encyclopedic intersection associated words;
reading an encyclopedic search related word list corresponding to the encyclopedic intersection related word, judging whether a label to be analyzed exists in the encyclopedic search related word list corresponding to the encyclopedic intersection related word, and if so, superposing the word frequency of the existing label to be analyzed and the word frequency of the corresponding encyclopedic intersection related word to be used as the word frequency of the encyclopedic intersection related word; otherwise, eliminating an encyclopedic search related word list corresponding to the encyclopedic intersection related word, wherein the word frequency represents the word frequency in the encyclopedic search related word list;
stacking word frequencies of labels in the information to be used as weights, sequencing the labels according to the weights of the information from high to low, and adding the information into a temporary information association pool to obtain a recommendation list;
the encyclopedic search associated word list is a list obtained by searching encyclopedic or knowledge base according to associated words.
8. A knowledge-base-based recommendation apparatus, comprising:
a memory for storing a program;
a processor for executing the program, the program causing the processor to execute the knowledge base based recommendation method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710947514.1A CN107908654B (en) | 2017-10-12 | 2017-10-12 | Knowledge base-based recommendation method, system and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710947514.1A CN107908654B (en) | 2017-10-12 | 2017-10-12 | Knowledge base-based recommendation method, system and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107908654A CN107908654A (en) | 2018-04-13 |
| CN107908654B true CN107908654B (en) | 2021-12-07 |
Family
ID=61840505
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710947514.1A Active CN107908654B (en) | 2017-10-12 | 2017-10-12 | Knowledge base-based recommendation method, system and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107908654B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110188187A (en) * | 2019-06-04 | 2019-08-30 | 成都终身成长科技有限公司 | Article recommended method and device, storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102622451A (en) * | 2012-04-16 | 2012-08-01 | 上海交通大学 | System for automatically generating television program labels |
| CN103294778A (en) * | 2013-05-13 | 2013-09-11 | 百度在线网络技术（北京）有限公司 | Method and system for pushing messages |
| CN106446189A (en) * | 2016-09-29 | 2017-02-22 | 广州艾媒数聚信息咨询股份有限公司 | Message-recommending method and system |
| CN106513014A (en) * | 2016-09-22 | 2017-03-22 | 江苏大学 | Hierarchical-pore carbon-supported acidic solid catalyst and preparation method thereof |
| CN106649334A (en) * | 2015-10-29 | 2017-05-10 | 北京国双科技有限公司 | Conjunction word set processing method and device |
| CN107045533A (en) * | 2017-01-20 | 2017-08-15 | 广东技术师范学院天河学院 | Educational resource based on label recommends method and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150073931A1 (en) * | 2013-09-06 | 2015-03-12 | Microsoft Corporation | Feature selection for recommender systems |
-
2017
- 2017-10-12 CN CN201710947514.1A patent/CN107908654B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102622451A (en) * | 2012-04-16 | 2012-08-01 | 上海交通大学 | System for automatically generating television program labels |
| CN103294778A (en) * | 2013-05-13 | 2013-09-11 | 百度在线网络技术（北京）有限公司 | Method and system for pushing messages |
| CN106649334A (en) * | 2015-10-29 | 2017-05-10 | 北京国双科技有限公司 | Conjunction word set processing method and device |
| CN106513014A (en) * | 2016-09-22 | 2017-03-22 | 江苏大学 | Hierarchical-pore carbon-supported acidic solid catalyst and preparation method thereof |
| CN106446189A (en) * | 2016-09-29 | 2017-02-22 | 广州艾媒数聚信息咨询股份有限公司 | Message-recommending method and system |
| CN107045533A (en) * | 2017-01-20 | 2017-08-15 | 广东技术师范学院天河学院 | Educational resource based on label recommends method and system |
Non-Patent Citations (2)
| Title |
|---|
| 基于Android的推荐阅读系统设计;查英华等;《智能计算机与应用》;20161231;第6卷(第6期);第55-58页 * |
| 基于内容标签的消防知识推荐系统;项灵辉等;《工业控制计算机》;20130731;第26卷(第7期);第85-88页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107908654A (en) | 2018-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017092622A1 (en) | Legal provision search method and device | |
| US9342627B2 (en) | Determining word information entropies | |
| CN112395556B (en) | Abnormal user detection model training method, abnormal user auditing method and device | |
| CN102495892A (en) | Webpage information extraction method | |
| TW202014940A (en) | Training sample obtaining method, account prediction method, and corresponding devices | |
| CN105893478A (en) | Tag extraction method and equipment | |
| CN107437038A (en) | A kind of detection method and device of webpage tamper | |
| CN110471948A (en) | A kind of customs declaration commodity classifying intelligently method excavated based on historical data | |
| CN106033445A (en) | Method and device for obtaining article association degree data | |
| CN107545038B (en) | Text classification method and equipment | |
| CN107368526A (en) | A kind of data processing method and device | |
| CN103914538B (en) | theme capturing method based on anchor text context and link analysis | |
| CN110706015A (en) | Advertisement click rate prediction oriented feature selection method | |
| CN111382948A (en) | Method and device for quantitatively evaluating enterprise development potential | |
| CN114385775A (en) | Sensitive word recognition method based on big data | |
| CN107908654B (en) | Knowledge base-based recommendation method, system and device | |
| Mulhair et al. | Filtering artifactual signal increases support for Xenacoelomorpha and Ambulacraria sister relationship in the animal tree of life | |
| CN110287238A (en) | A kind of exception water quality detection method and system based on priori knowledge | |
| CN113536777A (en) | Extraction method, device and equipment of news keywords and storage medium | |
| CN113282686B (en) | Association rule determining method and device for unbalanced sample | |
| CN114021716A (en) | Model training method and system and electronic equipment | |
| CN113962218A (en) | Illegal application identification method, device and equipment and readable storage medium | |
| CN114637917A (en) | Information head bar recommendation method and device based on artificial intelligence | |
| CN111783452B (en) | Model training method, information processing method, device, equipment and storage medium | |
| CN113779243A (en) | Automatic commodity classification method and device and computer equipment |
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 |