CN109918561A - A kind of study recommended method of library's (studying space) - Google Patents
A kind of study recommended method of library's (studying space) Download PDFInfo
- Publication number
- CN109918561A CN109918561A CN201910021753.3A CN201910021753A CN109918561A CN 109918561 A CN109918561 A CN 109918561A CN 201910021753 A CN201910021753 A CN 201910021753A CN 109918561 A CN109918561 A CN 109918561A
- Authority
- CN
- China
- Prior art keywords
- user
- data
- knowledge point
- information
- study
- 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.)
- Pending
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention, which discloses one kind, can be applied to library, the study recommended method of studying space." the accurate sideslip " for effectively solving the problems, such as that existing recommendation class algorithm recommends study is recommended.When study class algorithm is applied in study recommendation, it is easy because the operation data of user is more, more accurately grasp the hobby of client, script will be learnt to the client of " uninteresting " professional knowledge, directly recommend some to be difficult firmly " to be lured ", for a user " more interesting ", " preferring ", " more amusement class " personalized recommendation video or book information in, to influence client's study.Recommended method effective solution problem of the invention, and propose what a kind of future was the theme with user, the development new direction of Smart library.
Description
Technical field
The technology is the interdisciplinary application of library science, computer application, is suitable in library, the construction of studying space
Or in upgrading.
Background technique
One, the exploration of the wisdom Model Establishment based on scholar
Various countries constantly apply to the prior art in the service and management of library's (studying space), and then improve the work of administrator
Make the service quality of efficiency.
With the variation of technology development and social environment, it is universal and scientific that library (studying space) becomes public education
The pusher of research, in recent years, the developmental research about library (studying space) are divided into two big Main Branches, and one is wisdom
Various technological means are applied to promotion management employee and made in efficiency and service by studying space, secondly for active studying space
Development exploration for the first time becomes service provider based on learner.
Existing technological means are mainly introduced library's (study by the concept of intelligent studying space (Smart library)
Space) construction, the arrival of the 4th scientific and technological revolution, and " International Library forum " conference (SILF 2018) hold,
Wherein forum has been attended in 269 representatives from global 23 countries and regions, and with " library allows social more wisdom more to wrap
The theme representatives of various countries of appearance " carry out current situation and exchange with the report of development trend.The U.S., Britain, China, Belgium, France,
The country such as Slovakia, Singapore relatively payes attention in spatial framework reconstruction and management service upgrading, and the U.S., Lithuania, China
The country such as Hong Kong, Canada is made that technology changes on remote online education and intelligent consulting for promoting collection building and increasing
Into.And China just mentions intelligent space (library) construction in national social science fund guide for application in 2019,
Active studying space be with Japan for pioneer concept introduce, be more precedent proposition by learner be construction bodies
Studying space, and terminate only with optimize library (studying space) service and working efficiency be dominate development pattern.
Clearly there is " active studying space " in Chinese society Projects Applying For The Science Fund guides in 2018, a few days ago by the end of the present patent application,
In China's latest academic research, the free thought of the only two ideal establishment models about active studying space.
A kind of method disclosed by the invention is to be total to the learning agent feature of active studying space and smart space feature
A kind of Novel learning space (library) deposited can be with a kind of example of new model of application development.
Two, " the accurate sideslip " for recommending class algorithm to apply on study is recommended
The development of proposed algorithm is several big electric business based on Alibaba and today's tops and news media as the beginning, various
Recommend class algorithm, for client for the personal preference of commodity, on preference is inferred, constantly amendment and improves, extend a variety of push away
Recommend algorithm.Various proposed algorithms are widely used in commercially valuable higher electric business commodity evaluation, film is recommended, tourism recommendation
On equal consumer products platform.Now, recommend class algorithm also gradually to start to tend to education sector, be used to Optimization Learning efficiency,
In include that field is applied to the personal preference books in library and borrows in recommendation, on internet teaching course preference resource recommendation.
Recommend class algorithm, is exactly to obtain accurately user's " hobby " and " preference " by excavating user data.
But just because of class algorithm volume feature is recommended, it is the personal preference according to user's operation and is recommended, so
Content-based recommendation, or the recommendation based on user or deep learning is no matter used to recommend, even knowledge mapping pushes away
It recommends, the recommendation for the aspect knowledge for " excessively " biasing toward user's " hobby " is all be easy to cause to show, since more and more users grasp
Make data, proposed algorithm can increasingly " accurate " the hobby and point of interest for recognizing user couple, recommended range can be more
Reduction or weighting are arrived, and can more be caused in the content of reader's interest or reader's preference, such as novel, story, film, caricature, this
It is a little to be easier to become the preference class data or books for causing user interest.
The algorithm advantage of class is recommended to be that the basic data read and user's operation data can obtain more accurately user
Hobby, the recommendation of preference, but in study, but becoming disadvantage.
Because it will more accurately grasp client hobby, by originally the client to be learnt, directly recommend it is some be difficult through
" more interesting " for a user that firmly " lures ", " preferring ", " more amusement class " film or book information in, formed and " run
Recommendation partially " influences learning efficiency to influence its normal study.
So being difficult to meet the real demand that study is recommended, on the contrary more users only with traditional proposed algorithm
Operation, more accurately grasps user information, more be easy to cause and user is attracted to see more there is " temptation ", is more biased towards user's " interest ", " love
It is good " data or resource on, to allow user more to deviate the original purpose of study, in order to avoid this kind of situation.
A kind of method disclosed by the invention or software or system are exactly solution to the problem thus.
Summary of the invention
The invention discloses a kind of embodiments, can solve following problems:
1, developing direction and development trend research of the academia (society) for library (studying space)
A kind of development pattern of novel (active) Smart library based on user is proposed, to library's (study
Space) development have application value, for same type research have innovate academic reference value.
2, the existing proposed algorithm of solution, which is applied in study recommendation, is easy " accurate sideslip " recommendation to influence learning effect
The problem of.
Existing proposed algorithm obtains user preferences with big data, and the characteristics of preference, the quotient with the sale of product class is widely applied
In product industry, the hobby and interest of user can be accurately caught, recommends user " preferring ", the quotient of " being difficult to resist temptation "
Product.But during Applied Learning is recommended, it is be easy to cause the hobby and interest for more accurately grasping user, it is easier to need to learn in user
User " preferring " is recommended when habit, film, caricature, the novel etc. of the inclined amusement and interest of " being difficult to resist temptation ", from
And influence the study heart of user.
A kind of method disclosed by the invention, by the methods of increase " requirements ", " count value " " preference value ", by a variety of calculations
The advantage and disadvantage of method, which are integrated, to be utilized, that is, saves the preference advantage of proposed algorithm, and avoid the excessive deviation of preference to well
The influence of habit.
3, " information explosion " epoch, the learning efficiency of learner
Vast books and electronic bits of data can not count, and a neck, a profession, an even subdiscipline just can have any more tens of thousands of
This book, tens parts of data have to let us consider that, how in such a epoch, accurately, clearly obtain oneself and want
Perhaps, the data or books looked for are easy to, but oneself " needs " study how is known in the life full of temptation and emotion
Data or books, are just difficult to.Those you like, the data wanted or books, you need to go to learn for those, need
The data or books understood but oneself do not recognized, are exactly the original intention of this algorithm, allow learner on the road of study, no
It is easy to wander away.
4, the directiveness reference of library (studying space) construction
The construction of library or studying space generally requires buying high-volume books and data, and fund is huge, but purchases back
The books and data come, but some eternal nobody's " browsing ", what also some people wanted does not have but.So using existing
There are data and data mining to obtain the directive property suggestion for needing to purchase books, and obtains this application mechanism using whole big datas
Data distribution and it is constructive know opinion, will all have important meaning for library, studying space, or using the development of mechanism
Justice.
The invention discloses a kind of method, one such embodiment is as follows:
1. one kind can be applied to the study recommended method in library's (studying space), main feature and realization step are as follows:
Step 1: user identity identification;
Independent database is established to each ID user, carries out storage and the later period ID user letter of user information and operation
The storage of breath.
Logging in system by user carries out the identity effect of login user, the user identity effect, including but not office first
It is limited to following effect mode: account log-on message, fingerprint log-on message, iris log-on message, palmmprint log-on message, face recognition
(recognition of face) log-on message, identity effect is compared with existing identity information, identical, then uses same independent ID and independent data
Library, it is different, then ID and independent information database are created for the user, the fingerprint login mode is needed from fingerprint identification device
Data are obtained, the iris login mode needs to obtain data from iris identification equipment, and the palmmprint login mode needs
Data are obtained from fingerprint identification device, the recognition of face login mode needs to obtain data from face identification device, wherein
User's independent information database, using User ID as the unique identification of the information database, described in independent information
Database is the database for storing operation information, personal information in space corresponding to the ID user.
Step 2: user information obtains;
It is comprehensive to obtain user related information, by society, school, profession (field), individual goal, obtain comprehensive, comprehensive, profession
User information and the knowledge that should learn.
User information acquisition is carried out, including obtains userspersonal information, user's operation information, wherein the individual subscriber is believed
Breath includes but is not limited to following information: individual subscriber situation information, personal place field or specialized information, personal short-term goal
Or learning objective information.
Organization information relevant information is obtained, wherein the mechanism relevant information includes but is not limited to following information: mechanism
Place field, character of institution, mechanism effect, the culture purpose of mechanism culture difference personnel, training method, cultural method, examination
Culture effect Assessment.
Acquisition field (industry or profession) information, wherein the field (industry or profession) information including but not limited to
Lower information: specific features knowledge point, the field that field (industry or profession) is related to specific subject, field (industry or profession) is related to
(industry or profession) development trend, field (industry or profession) current situation, field (industry or profession) developing history.
Step 3: establishing feature tag for user;
According to user related information, by society, school, profession (field), individual goal, related data is obtained, establish user
Related all knowledge points that should learn, establish knowledge base.The knowledge of knowledge base is classified, by higher level it is such into
Row classification, or classified using synonymous part of speech, or classified with knowledge mapping.But counting is established for each knowledge point
Value, the count value of (root node) wherein the most higher level that classifies classifies, using between all most higher levels (root node) as count value
Comparison, each independent subclassification (knowledge mapping or near synonym classification or higher level's classification) is interior to have oneself different grades of counting again
Value.
According to the school work target or short-term goal information of user, field (industry or profession) information, institute in conjunction with where user
In mechanism or unit information, user's operation information, the feature tag for carrying out all related informations is extracted, and establishes user characteristics label
Library, wherein the user characteristics tag library can be whole user characteristics label storages, or being will be the same as semantic (near synonym) feature mark
Label are sorted out, and similar (near synonym) knowledge mapping of feature tag is established, and are formed by multiple higher level's feature tag synonym knowledge
The user characteristics tag library of map composition;And study is established for all feature tags (higher level's feature tag synonym knowledge mapping)
Count, wherein it is described study be counted as user browse correlated knowledge point number, retrieval correlated knowledge point counting how many times it is every retrieval or
Browsing is primary, and response increases count value.
Step 4: user's learning demand data processing;
A requirements are established into each knowledge point, effect is prevented primarily to the timely hobby deviation for correcting proposed algorithm
The only personal preference of proposed algorithm walks out and needs the knowledge point range that learns, requirements can according to the demand of user target it is different
And different condition of increasing demand is different or the increased requirements numerical value of condition of the same race is different, requirements can be to count, can be with
For grade.
Target establishes " requirements " of different user feature tag, the demand of the user characteristics label according to demand
Value needs the corresponding feature knowledge point of the described user (the similar knowledge mapping of feature tag) learnt to need to learn for user
Urgent degree, establish quantification gradation;Its described requirement objective is following target but is not limited to following target: user's short-term goal
Data, school work target data, used studying space unit (mechanism) cultivate mesh number evidence.
Its described requirements, the condition of degree of increasing demand (counting or rate range), including but not limited to following value of increasing demand
Condition, and do not have sequence requirement: according to user's short-term goal or user's learning objective data, determining the knowledge that user needs to learn
The coverage area (field or professional industry) of point, is the knowledge point in the coverage area, increases level-one requirements,
The user of habit mikey (mechanism) training objective and plan needs the coverage area (field or professional industry) of knowledge point,
For the knowledge point of the coverage area, increase level-one requirements, is user's short-term goal (user's learning objective) and mikey
The user of (mechanism) training objective and plan needs the coverage area intersection of learning knowledge point, described in coverage area knowledge
Point increases level-one requirements, and specific course of cultivating is related to knowledge point, and the knowledge point increases level-one requirements, and industry is wanted
Ask knowledge point involved in standard or examination outline standard, according to and the correspondence of different knowledge point should learn degree, answer Grasping level,
It should understand that degree, correspondence increase corresponding level requirements value in various degree.
Step 5: user learns to recommend;
User's operation preference profiles label counting is established, the user's operation feature tag is counted as user in studying space
Operation behavior personal preference data, according to user to the preference of correlated knowledge point where different characteristic label, to spy
It levies label and carries out user " preference counting ".
" preference counting " described in it increases the condition that preference counts, including but not limited to the following conditions, and out-of-order is wanted
Ask: the knowledge point that user search information point is related to, the knowledge point increase level-one preference and count, and user browses information point and is related to
Knowledge point, the knowledge point increases level-one preference and counts, described in the user group that belongs to, preference counts higher is related to
Knowledge point, it is corresponding to increase corresponding preference counting in various degree.
The individual consumer's feature tag established by former steps has " study count value ", " requirements ", wherein described
Count value be characterized the browsing quantity of label related term, belonging to requirements be that the needs of this feature label related term are learned
The demand degree of habit, when recommend learning knowledge point pertinent texts data to user, " the study counting of same feature tag
Value " reversely influences the shared weight of " requirements " of same feature tag in user recommends to calculate, and requirements and shared pushes away
The product of the weight of algorithm is recommended, the basic data that same user's " preference counting " is recommended collectively as study is recommended, and is had similar
The user of feature tag group does user's classification, is formed similar " user group of feature tag ", a user can belong to more simultaneously
Category feature tagging user group.
2. according to claim 1, the active studying space application of this method, need with server (or server
Group) and client (or client group) as hardware rely on, the server (or server zone) be control server, progress
The total data of studying space stores or processing;Its described client (or client group) be linked to control server (or clothes
Be engaged in device group) electronic equipment, have the feature that
The main feature that active studying space has are as follows: there is interaction space, with books, data is borrowed and read function,
Wherein the interaction space has interaction space data, wherein the books, data, which are borrowed, has books, data with read function
Library data;Interaction space (Virtual Space that studying space in constructs) or a of the active interaction space between user described in it
The space (Virtual Space constructed in studying space) of people's release information, the books, data are that the space uses unit (machine
Structure) the copyrighted paper book of tool, data, e-book, data information bank.
Step 1: user identity identification;
With identity effect module, letter can be logged according to account log-on message, fingerprint log-on message, iris log-on message, palmmprint
Wherein any login mode carries out log-on message comparison for breath, face recognition (recognition of face) log-on message, identical, then using same
Independent ID and self contained data base, it is different, then ID and independent information database are created for the user, the fingerprint login mode needs
It arranges in pairs or groups fingerprint identification device, the iris login mode needs iris identification equipment of arranging in pairs or groups, the palmmprint login mode
Fingerprint identification device of arranging in pairs or groups is needed, the recognition of face login mode needs face identification device of arranging in pairs or groups, wherein the independence
Information database, using User ID as the mark of the information database, described in independent information database be store the ID
Operation information, personal information in space corresponding to user.
Step 2: user role is distinguished;
The other major function of role zone is established, is to reduce, the data volume of the Overall Acquisition of user data is excessive, leads to the later period
Calculating speed is excessively slow, the user of different identity, different for the demand of study, so the data that should be obtained are not laid particular stress on also just not
Together, when the role of user can distinguish, so that it may which different directions carries out the acquisition of user's every terms of information emphatically, to reduce data
Input quantity reduces the calculation amount of subsequent calculating.Make program that there is stronger practicability, reduces the hardware investment used.
It is different according to application unit (mechanism), it is equipped with different user Role Identity, different user role combines school work target
Or short-term goal, the acquisition data needed are different, recommend learning outcome also different.
Step 3: user information obtains;
With subscriber information module, including acquisition userspersonal information, user's operation information is obtained, wherein the individual subscriber is believed
Breath including but not limited to once information: individual subscriber situation information, user interaction information, it is personal where field or profession letter
Breath, user's operation information, personal short-term goal or learning objective information, wherein the specific item of the personal information obtained, according to user
Identity difference obtains different personal information.
Organization information relevant information is obtained, wherein the mechanism relevant information includes but is not limited to following information: mechanism
Place field, character of institution, mechanism effect, the culture purpose of mechanism culture difference personnel, training method, cultural method, examination
Culture effect Assessment obtains mechanism difference correlation according to user identity difference and is related to information;
Acquisition field (industry or profession) information, wherein the including but not limited to following letter of the field (industry or profession) information
Breath: field (industry or profession) is related to specific subject, the specific features knowledge point that field (industry or profession) is related to, field (row
Industry or profession) development trend, field (industry or profession) current situation, field (industry or profession) developing history, according to user
Identity difference obtains different field (industry or profession) information.
Step 4: establishing feature tag for user;
According to the school work target or short-term goal information of user, field (industry or profession) information, place machine in conjunction with where user
Structure or unit information, user's operation information, the feature tag for carrying out all related informations extract, and establish " user characteristics label
Library ", wherein the user characteristics tag library can be whole user characteristics label storages, or being will be similar, with semanteme (near synonym)
Feature tag is sorted out, and similar (near synonym) knowledge mapping of feature tag is established, and is formed synonymous by multiple higher level's feature tags
The user characteristics tag library of word knowledge mapping composition;And it is built for all feature tags (higher level's feature tag synonym knowledge mapping)
Vertical study counts, wherein the study is counted as browsing, retrieval counts.
Step 5: user's learning demand data processing:
Target establishes " requirements " of different user feature tag according to demand, and the requirements of the user characteristics label are
The urgent degree for the corresponding feature knowledge point of the described user (the similar knowledge mapping of feature tag) that user needs to learn, establish etc.
Grade quantization;Its described requirement objective is following target but is not limited to following target: in user's short-term goal data, interactive data
The knowledge point data of high frequency discussion, school work target data, used studying space unit (mechanism) cultivate mesh number evidence.
Its described requirements, the condition of degree of increasing demand (counting or rate range), including but not limited to following increase need
The condition of evaluation, and do not have sequence requirement: according to user's short-term goal or user's learning objective data, determine what user needed to learn
The coverage area (field or professional industry) of knowledge point, is the knowledge point in the coverage area, value of increasing demand,
The user of habit mikey (mechanism) training objective and plan needs the coverage area (field or professional industry) of knowledge point,
For the knowledge point of the coverage area, value of increasing demand is user's short-term goal (user's learning objective) and mikey (machine
Structure) user of training objective and plan needs the coverage area intersection of learning knowledge point, described in coverage area knowledge point, increase
Add requirements, the knowledge point of interactive data medium-high frequency discussion, corresponding knowledge point is increased demand value, and specific culture course is related to knowledge
Point, the knowledge point are increased demand value, knowledge point involved in industry requirement standard or examination outline standard, according to without
Correspondence with knowledge point should learn degree, answer Grasping level, should understand degree, corresponding to increase corresponding requirements in various degree.
Step 6: user learns to recommend;
User's operation preference profiles label counting is established, the user's operation feature tag is counted as user in studying space
Operation behavior personal preference data, according to user to the preference of correlated knowledge point where different characteristic label, to spy
It levies label and carries out user " preference counting ".
" preference counting " described in it increases the condition that preference counts, including but not limited to the following conditions, and out-of-order is wanted
Ask: the knowledge point that user search information point is related to, the knowledge point increase preference and count, and what user's browsing information point was related to knows
Know point, the knowledge point increases preference and counts, and user interaction data high frequency discusses knowledge point, and increase preference and counts, described in
In the user group belonged to, preference counting is higher to be related to knowledge point, and corresponding increase in various degree corresponds to preference counting.
The individual consumer's feature tag established by former steps has " study count value ", " requirements ", wherein described
Count value be characterized the browsing quantity of label related term, belonging to requirements be that the needs of this feature label related term are learned
The demand degree of habit, when recommend learning knowledge point pertinent texts data to user, " the study counting of same feature tag
Value " reversely influences the shared weight of " requirements " of same feature tag in user recommends to calculate, and requirements and shared pushes away
The product of the weight of algorithm is recommended, the basic data that same user's " preference counting " is recommended collectively as study is recommended, and is had similar
The user of feature tag group does user's classification, forms the user group of similar features label, and a user can belong to multiclass simultaneously
Feature tag user group.
3. feature also includes according to claim 1 or described in 2 any one
Its described user learns to recommend, and works as logging in system by user, and when the user does not have too many operation note, system will be according to existing
There are unit (mechanism) data, school work target or the short-term mesh of personal information data, place field (profession or industry), use space
Data are marked, by User-CF algorithm meter similar features tagging user group, same user can use in multiple feature tags simultaneously
Family group, and the recommending data knot according to collaborative filtering algorithm with not calculating acquisition feature tag user group
Fruit chooses wherein more each user group recommendation results in the top and is recommended;
Its described user learns to recommend, and works as logging in system by user, when which has more operation note, Item-CF can be used
The data that algorithm carries out user's operation preference calculate, and are obtaining the heap sort of similar features tagging user by the calculating of User-CF
Data recommendation as a result, by the result feedback data of history recommended user, after comprehensive three item datas, every item data selection row
The result of name earlier above is recommended, or according to deep learning algorithm, is carried out data training of the client before login, obtained under user
When primary login, the learning data that should recommend, user recommends when logging in;
The result feedback data of its user's history recommended user, refers to and works as logging in system by user, for system recommendation
Practise the operational feedback of recommending data, if operated and concrete operations content, the feedback data include but be not limited to
Following feedback data, out-of-order requirement: in recommended amount percentage, user's operation selection is pushed away for number of clicks and browsing time
Recommend the secondary preference data of content, the degree of association of user search information knowledge point and user's operation recommendation knowledge point, user
Interactive data high frequency discusses the degree of association of knowledge point and user's operation recommendation knowledge point, userspersonal information and user's operation
The degree of association of content knowledge point, user facility information (field or profession or industry) are associated with user's operation content knowledge point
Degree obtains the staining effect that study is recommended.
According to claim 1, or described in 2 any one, feature also includes
Its described user learns to recommend, and further includes that study plan is recommended, the study plan is recommended, with the user characteristics mark
Label or similar (near synonym) knowledge mapping of the feature tag, the retrieval entry knowledge point in user search information is the same as the user
The intersection of the correlated knowledge point of feature tag correlated knowledge point or similar (near synonym) knowledge mapping of the feature tag, gets
In the user characteristics label correlated knowledge point or the correlated knowledge point model of similar (near synonym) knowledge mapping of the feature tag
Enclose interior user search entry;
With the feature tag or similar (near synonym) the knowledge mapping data of the feature tag, the user characteristics label
User search entry in the relevant knowledge point range of correlated knowledge point or similar (near synonym) knowledge mapping of the feature tag
Data and user place field (profession or industry) related data, the related culture mesh of the used mechanism in space (unit)
Data, what the basic data that user's short-term goal or school work target data carry out study plan recommendation was recommended as study plan
Basic data carries out recommendation calculating;
Its described study plan is recommended, using the short-term purpose of user or school work purpose as learning objective, when according to actual log
Between and the institution where he works about user study plan or study arrange combine, carry out study plan design.
Its described study plan is recommended, by the feature tag of user or the similar knowledge mapping of the feature tag, with described
The study degree of urgency of " requirements " as feature tag correlated knowledge point obtains user described in same type using User-CF algorithm
The study correlated knowledge point of group obtains frequent of user group described in same type using FP Growth or PrefixSpan algorithm
Practise content (knowledge point) and obtain learning difficulty, and according to the knowledge point learn the high user of frequency operated by last time this know
In the data or content for knowing point study, the high Content Selection of study high frequency carries out recommending data in the knowledge point study plan
The preferential recommendation of high-frequency ranking;
Its described study plan is recommended, and is the short-term learning objective in place or school work objective plan according to user's actual access time
Different phase, the cultivation stage of unit one belongs to (mechanism), course completes the stage, recommended, and recommendations include, and answers
It practises, previews, the study plan for extending aspect is recommended, and is recommended according to the relevant knowledge of " requirements ", the relevant knowledge
The particular content that point is recommended, with the frequent learning Content of the user group, study frequency high user in the knowledge point is grasped
Study high frequency high data or books and user search letter in the data or content of the study of the last time of the work knowledge point
Know with the user characteristics label correlated knowledge point or the feature tag similar (near synonym) retrieval entry knowledge point in breath
Recommended the intersection knowledge point for knowing the correlated knowledge point of map;
Its described study plan is recommended, different according to user's actual access time, and replaces recommendation, or check that history pushes away
Recommend content.
5. a kind of study plan recommended program (plug-in unit) based on active studying space, it is characterized in that:
With user authentication module, user identity effect is carried out;
With User profile acquisition module, user information acquisition is carried out;
With feature tag module is established for user, the foundation of user characteristics label is carried out;
With user's learning demand data processing module, the data processing of user's learning demand is carried out;
Learn recommending module with user, carries out user and learn recommending data processing;
Its described module specific work steps is described in any one of claims 1 or 2.
According to claim 5, it is characterized in that:
Also with having books, acquisition of materials suggestion module, the purchasing recommendation module carries out " demand according to basic data
The higher knowledge point group of value " carries out high frequency study, the screening of retrieval knowledge point, and is learnt according to the high frequency, retrieval knowledge point
User learn frequency, establish sequence, study frequency is higher, but does not have copyrighted books or data to be built in studying space
View buying is recommended, the basic data, including but not limited to once categorical data: recommended according to user's learning knowledge point,
User's learning knowledge point feedback data and user search data;
7. according to claim 5, it is characterized in that:
Also there is visualization big data analysis to excavate module, by the continuous increasing of the applied unit in the space (mechanism) access user
More, user's learning data amount constantly increases, and can carry out Predictive Analytic by the big data in the space
The analysis of Capabilities(predictability) function, exploratory data analysis Exploratory Data is used to total data first
Analysis shows various data by way of " figure ", so that obtaining arriving for various data visualizes distribution map, when
Between sequence data and transformed variable, recycle hash matrix diagram to illustrate the relationship of variable between any two, and owned
Collect statistics amount, the collect statistics amount includes but is not limited to a under type: calculate mean value, maximum value, minimum value, on
Lower quartile and determine exceptional value, and respectively or summarize corresponding visualization " figure " is presented, described in " figure " include but not office
It is limited to following form: histogram, Stem-and-Leaf Plot, chart, cylinder.
A kind of study plan recommender system based on active studying space, needs with server (or server zone) and visitor
Family end (or client group) is relied on as hardware, and the server (or server zone) is control server, and it is empty to carry out study
Between total data storage or processing;Its described client (or client group) is to be linked to control server (or server zone)
Electronic equipment, have the feature that
With server;
With client, user's operation storage can be carried out;
With user authentication module, user identity effect is carried out;
With User profile acquisition module, user information acquisition is carried out;
With feature tag module is established for user, the foundation of user characteristics label is carried out;
With user's learning demand data processing module, the data processing of user's learning demand is carried out;
Learn recommending module with user, carries out user and learn recommending data processing;
Its described module specific work steps is the execution of any one of claims 1 or 2.
According to claim 8, it is characterized in that:
Also with having books, acquisition of materials suggestion module, the purchasing recommendation module carries out " demand according to basic data
The higher knowledge point group of value " carries out high frequency study, the screening of retrieval knowledge point, and is learnt according to the high frequency, retrieval knowledge point
User learn frequency, establish sequence, study frequency is higher, but does not have copyrighted books or data to be built in studying space
View buying is recommended, the basic data, including but not limited to once categorical data: recommended according to user's learning knowledge point,
User's learning knowledge point feedback data and user search data;
10. according to claim 8, it is characterized in that:
Also there is visualization big data analysis to excavate module, by the continuous increasing of the applied unit in the space (mechanism) access user
More, user's learning data amount constantly increases, and can carry out Predictive Analytic by the big data in the space
The analysis of Capabilities(predictability) function, exploratory data analysis Exploratory Data is used to total data first
Analysis shows various data by way of " figure ", so that obtaining arriving for various data visualizes distribution map, when
Between sequence data and transformed variable, recycle hash matrix diagram to illustrate the relationship of variable between any two, and owned
Collect statistics amount, the collect statistics amount includes but is not limited to following manner: calculate mean value, maximum value, minimum value, on
Lower quartile and determine exceptional value, and respectively or summarize corresponding visualization " figure " is presented, described in " figure " include but not office
It is limited to following form: histogram, Stem-and-Leaf Plot, chart, cylinder.
Detailed description of the invention:
Fig. 1 is a kind of flow chart of open case study on implementation
Fig. 2 is a kind of example of open case study on implementation
Specific embodiment:
1. working as a kind of method disclosed by the invention, when being applied in libraries of the universities, the mode implemented has following manner, but simultaneously
It is not limited to following manner:
The wherein undergraduate user of this class teaching of colleges and universities, after login system, system will verify its identity, for example newly
User will establish the virtual data base of the user, for storing the personal information and operation information of user.
And the user identity belonging by student's selection, the row as where student, teacher, profession, social personnel, social personnel
Industry etc. identity it is selected, as the student selects pupilage, place universities and colleges, and learn profession, user can not also select to use
Family identity.
In next step, school work target or short-term goal setting can be carried out, e.g., textual criticism, final examination, expands professional skill at graduation
Energy etc. option, and two or more can be selected according to demand, goal-setting can also be modified at any time, change the aim of learning,
It can not also select learning objective.
System carries out personal information acquisition according to user identity, the learning objective of selection, such as non-selected, then obtains whole
User information, or acquisition of information is carried out according to preset identity or purpose.
Module is obtained by data acquisition and takes userspersonal information, including student name, gender, student number wait users certainly
The data that row is filled in.
Such as, which can carry out data access, then can obtain Students ' Major, specific learned lesson, curriculum schedule
It arranges, so that professional each subject where obtaining student should gain knowledge a little, and gets the substantially study that student should gain knowledge a little
Time course (according to the hours of instruction arrangement of curriculum schedule).Such as, book lending system can be docked, user can be obtained in library
The record for readding books borrows content according to the time is borrowed, and carrying out correlated knowledge point extraction, (summary-type extracts, according to book classification
Obtain book attributes label).The system initial stage of establishing can carry out typing to the information of universities and colleges, such as class of languages universities and colleges, such as Engineering institute
School, such as higher education universities and colleges, vocational education universities and colleges.Such as, it can connect internet, can obtain, user's profession, the correlation of course
Knowledge point update, latest news, the newest explanation in the knowledge point of encyclopaedia entry.
The retrieval of user, browsing, the operation information of the operations acquisition user such as reading will also be passed through.
User characteristics tag library is established, by " student ", the limitation of " short-term goal ", the professional knowledge that needs are learnt
Point carries out whole acquisitions, while carrying out classification and extracting higher level, and classifying, and is stored, and mainly similar word is sorted out, and synonym is returned
It receives, close word is concluded, and carries out foundation association, is stored, and, by uppermost time, can also be made in the form of knowledge mapping
For root node, other similar parts of speech or close, similar word carries out the foundation of knowledge mapping as node, and by multiple knowledge graphs
Spectrum is stored.
To each knowledge point, count value is established, label is associated counting to each knowledge point, wherein using every time
Family retrieves some knowledge point, knowledge point, near synonym, while carrying out increase count value, and specific count value is according to the clear of retrieval
Look at duration, the duration for browsing the term is longer, and the count value being included in is bigger.
To each knowledge point, requirements are established, label is associated counting to each knowledge point, wherein thoroughly do away with and learn
Requirements are increased different value by the outline requirement for practising target, when short-term goal is textual criticism, is then obtained this by internet and are examined
Card, the examination outline in this year do not have this year such as, and automatic to search for last year examination outline, by examination outline, (picture category can pass through
Text region obtains), the Grasping level of answering of different knowledge points is obtained, corresponding to increase different demands value, requirements are higher, then represent
The knowledge point is more important, and it is higher to learn urgent degree, the requirements also according to user search when, the knowledge point of retrieval is big in examination
Knowledge point within the scope of guiding principle is identical, then the requirements of the knowledge point also increase.
To each knowledge point, preference counting is established, label counts each Knowledge Relation, wherein according to user
Browsing, retrieval, wait operation to preference count it is corresponding increase, small-scale knowledge point range is acted on less, but for compared with
Large-scale target, the technical value can increase in a wide range of, the institute based on the personal preference numerical value under conditions of demand of specialty knowledge
Accounting weight.
When user's first login is cold-started in order to prevent when without excessive operation data, so we can pass through user
The hobby of group carries out similar recommendation, and the specific method is as follows:
The recommendation of user calculates, and by the personal characteristics label of user, carries out the calculating of User-CF algorithm, it is high to obtain similarity
The high user of similarity is established user group by user, and the same user can pass through Jaccard public affairs in different user groups
Formula calculates, wherein | N (u) | it is the number of users for liking data u, | N (v) | it is the user for liking data v
Number, the high user of similarity establish to a user group.For the same user since the identity of setting is different, target is different, can be with
In different multiple user masses, according to collaborative filtering algorithm respectively by the hobby object of similar users group
Product carry out ranking, in user group, carry out ranking that more people like.More people like it is in the top recommend, can also
To be recommended by user's interest level algorithm, user a is to article e, and wherein similar users are b, and user collects S (a, k),
AndIt is the similarity of user A, B,User b to i like degree according to interest level recommendRecommending data is obtained as a result, being recommended.
When user has multiple login record, when having more operation note, Item-CF algorithm can be used and carry out user's operation
The data of preference calculate, and pass throughThe similarity calculation of browsing data, knowledge point is carried out, | N (u) | be
Like the number of users of data u, | N (v) | be the number of users for liking data v, | N (u) ∩ N (v) | be like simultaneously data type i and
The number of users of data type j obtains similar features tagging user group calculating by collaborative filtering algorithm
The data recommendation result of classification.Thoroughly do away with simultaneously that user's history is recommended as a result, clicking rate, the data of click trend classification, carry out
Recommendation results feedback, data carry out secondary recommendation and calculate.Three kinds of recommending datas, knot of three kinds of data by ranking earlier above are obtained altogether
Fruit is recommended.
The time is calculated to save, user can be subjected to initial user heap sort, and step in user after user's first login
Pre- recommendation results calculating is carried out before record.
The system also has user's study plan recommendation function, mainly reduces the shared weight of personalized recommendation, will be a
Property be entirely limited and needing in the knowledge that learns, the weight of requirements is expanded.
Its described study plan is recommended, by the feature tag of user or the similar knowledge mapping of the feature tag, with described
The study degree of urgency of " requirements " as feature tag correlated knowledge point obtains user described in same type using User-CF algorithm
The study correlated knowledge point of group obtains frequent of user group described in same type using FP Growth or PrefixSpan algorithm
Practise content (knowledge point) and obtain learning difficulty, and according to the knowledge point learn the high user of frequency operated by last time this know
In the data or content for knowing point study, the high Content Selection of study high frequency carries out recommending data in the knowledge point study plan
The preferential recommendation of high-frequency ranking.
With purchase of books recommendation function, all kinds of results and the height that uses of the space which mainly recommends according to user
Books in the copyright in school, data range be compared, demand degree is high, but there is no in copyright books and data it is specific
Number or range classification, carry out sub-category sequence.
It excavates with big data depth, since operation user is more, is used since the applied unit in the space (mechanism) accesses
Family is increasing, and user's learning data amount constantly increases, and can carry out Predictive by the big data in the space
The analysis of Analytic Capabilities(predictability) function, exploratory data analysis is used to total data first
Exploratory Data Analysis shows various data by way of " figure ", including by all numbers of users
Read classification according to, user, short-term goal classification, user grade, read seniority among brothers and sisters, read classification seniority among brothers and sisters, etc. data, carry out dotted
Figure is drawn, and forms the Visual Graph of a variety of data, and according to data relationship and demand as a result, counting respectively to different data
It calculates mean value, maximum value, minimum value, upper lower quartile and determines exceptional value, and respectively or to summarize presentation corresponding " figure ", thus
Some data results with predictability can be obtained.
2. working as this method, it is applied to library (country, city, regional rank etc.), the mode implemented has following manner,
But it is not limited to following manner:
User identity will increase more industries, specialized information, and more short-term goals, such as needle be arranged in the target of user
To textual criticism, professional skill is improved, is broadened one's knowledge etc. and system default form, and by retrievable user information, it is more logical
Cross selected user, field, profession, trade information are expanded by internet, obtain correlation more new on internet
The copyrighted data of tool in news, consulting, data and collection obtains the sector development trend by internet, and development is existing
Shape and extendible the scope of one's knowledge, correlated knowledge point, carry out establishing user characteristics label.Changing label has flexibility and changeability, according to
User logs in setting short-term goal every time and changes and change, according to the recent field industry of user's login, Career Information variation
Variation.
Specific proposed algorithm and specific data processing method, with colleges and universities using identical.
Claims (10)
1. a kind of study recommended method of library's (studying space), main feature and realization step are as follows:
Step 1: user identity identification;
Carry out the identity effect of login user, the user identity effect, including but not limited to following effect mode: account
Log-on message, fingerprint log-on message, iris log-on message, palmmprint log-on message, face recognition (recognition of face) log-on message, body
Part effect is compared with existing identity information, identical, then uses same independent ID and self contained data base, and difference is then new for the user
ID and independent information database are built, the fingerprint login mode needs to obtain data, the iris from fingerprint identification device
Login mode needs to obtain data from iris identification equipment, and the palmmprint login mode needs to obtain number from fingerprint identification device
According to the recognition of face login mode needs to obtain data from face identification device, wherein user's independent information data
Library, using User ID as the unique identification of the information database, described in independent information database be store the ID user
The database of operation information, personal information in corresponding space;
Step 2: user information obtains;
User information acquisition is carried out, including obtains userspersonal information, user's operation information, wherein the userspersonal information wraps
Include but be not limited to following information: individual subscriber situation information, personal place field or specialized information, personal short-term goal or
Target information is practised,
Organization information relevant information is obtained, wherein the mechanism relevant information includes but is not limited to following information: where mechanism
Field, character of institution, mechanism effect, the culture purpose of mechanism culture difference personnel, training method, cultural method, examination culture
The Assessment of effect,;
Acquisition field (industry or profession) information, wherein the including but not limited to following letter of the field (industry or profession) information
Breath: field (industry or profession) is related to specific subject, the specific features knowledge point that field (industry or profession) is related to, field (row
Industry or profession) development trend, field (industry or profession) current situation, field (industry or profession) developing history,;
Step 3: establishing feature tag for user;
According to the school work target or short-term goal information of user, field (industry or profession) information, place machine in conjunction with where user
Structure or unit information, user's operation information, the feature tag for carrying out all related informations extract, and establish user characteristics tag library,
Wherein the user characteristics tag library can be whole user characteristics labels storages, or for will with semantic (near synonym) feature tag into
Row is sorted out, and similar (near synonym) knowledge mapping of feature tag is established, and is formed by multiple higher level's feature tag synonym knowledge mappings
The user characteristics tag library of composition;And study meter is established for all feature tags (higher level's feature tag synonym knowledge mapping)
Number, wherein the study, which is counted as user, browses the every retrieval or clear of correlated knowledge point number, retrieval correlated knowledge point counting how many times
It lookes at once, response increases count value;
Step 4: user's learning demand data processing:
Target establishes " requirements " of different user feature tag according to demand, and the requirements of the user characteristics label are
The knowledge point (knowledge point of the similar knowledge mapping of feature tag) for the corresponding feature tag of the described user that user needs to learn
The urgent degree for needing to learn, establishes quantification gradation;Its described requirement objective is following target but is not limited to following target: user
Short-term goal data, school work target data, used studying space unit (mechanism) cultivate mesh number evidence,
Its described requirements, the condition of value of increasing demand (counting or rate range), including but not limited to following value of increasing demand
Condition, and do not have sequence requirement: according to user's short-term goal or user's learning objective data, determining the knowledge that user needs to learn
The coverage area (field or professional industry) of point, is the knowledge point in the coverage area, value of increasing demand, the study sky
Between the user of unit (mechanism) training objective and plan need the coverage area (field or professional industry) of knowledge point, for institute
The knowledge point for stating coverage area, value of increasing demand, is user's short-term goal (user's learning objective) and mikey (mechanism) is trained
The user for supporting target and plan needs the coverage area intersection of learning knowledge point, described in coverage area knowledge point, increasing needs
Evaluation, specific course of cultivating are related to knowledge point, and the knowledge point is increased demand value, industry requirement standard or examination outline mark
Knowledge point involved in standard, according to and the correspondence of different knowledge point should learn degree, answer Grasping level, should understand degree, it is corresponding different
Degree is increased demand value;
Step 5: user learns to recommend;
User's operation preference profiles label counting is established, the user's operation feature tag is counted as user in studying space
Operation behavior personal preference data, according to user to the preference of correlated knowledge point where different characteristic label, to spy
It levies label and carries out user " preference counting ",
" preference counting " described in it increases the condition that preference counts, including but not limited to the following conditions, and out-of-order requirement: using
The knowledge point that family retrieving information points are related to, the knowledge point increase preference and count, and user browses the knowledge point that information point is related to,
Its described knowledge point increases preference and counts, described in the user group that belongs to, preference count it is higher is related to knowledge point, it is corresponding not
Increase preference with degree to count;
The individual consumer's feature tag established by former steps has " study count value ", " requirements ", wherein the meter
Numerical value is characterized the browsing quantity of label related term, belonging to requirements be that the needs of this feature label related term are learnt
Demand degree, when recommend learning knowledge point pertinent texts data to user, " the study count value " of same feature tag, instead
To " requirements " for influencing same feature tag recommend the shared weight in calculating, requirements and shared proposed algorithm in user
Weight product, same to user's " preference counting " collectively as study recommend basic data recommend, and have similar characteristics mark
The user of label group does user's classification, is formed similar " user group of feature tag ", and a user can belong to multiple types spy simultaneously
Levy tagging user group.
2. according to claim 1, the active studying space application of this method, need with server (or server zone) and
Client (or client group) is relied on as hardware, and the server (or server zone) is control server, is learnt
The total data in space stores or processing;Its described client (or client group) is to be linked to control server (or server
Group) electronic equipment, have the feature that
The main feature that active studying space has are as follows: there is interaction space, with books, data is borrowed and read function,
Wherein the interaction space has interaction space data, wherein the books, data, which are borrowed, has books, data with read function
Library data;Interaction space (Virtual Space that studying space in constructs) or a of the active interaction space between user described in it
The space (Virtual Space constructed in studying space) of people's release information, the books, data are that the space uses unit (machine
Structure) the copyrighted paper book of tool, data, e-book, data information bank;
Step 1: user identity identification;
With identity effect module, letter can be logged according to account log-on message, fingerprint log-on message, iris log-on message, palmmprint
Wherein any login mode carries out log-on message comparison for breath, face recognition (recognition of face) log-on message, identical, then using same
Independent ID and self contained data base, it is different, then ID and independent information database are created for the user, the fingerprint login mode needs
It arranges in pairs or groups fingerprint identification device, the iris login mode needs iris identification equipment of arranging in pairs or groups, the palmmprint login mode
Fingerprint identification device of arranging in pairs or groups is needed, the recognition of face login mode needs face identification device of arranging in pairs or groups, wherein the independence
Information database, using User ID as the mark of the information database, described in independent information database be store the ID
Operation information, personal information in space corresponding to user;
Step 2: user role is distinguished,
It is different according to application unit (mechanism), it is equipped with different user Role Identity, different user role combines school work target or short
Phase target, the acquisition data needed are different, recommend learning outcome also different;
Step 3: user information obtains;
With subscriber information module, including acquisition userspersonal information, user's operation information is obtained, wherein the individual subscriber is believed
Breath including but not limited to once information: individual subscriber situation information, user interaction information, it is personal where field or profession letter
Breath, user's operation information, personal short-term goal or learning objective information, wherein the specific item of the personal information obtained, according to user
Identity difference obtains different personal information;
Organization information relevant information is obtained, wherein the mechanism relevant information includes but is not limited to following information: where mechanism
Field, character of institution, mechanism effect, the culture purpose of mechanism culture difference personnel, training method, cultural method, examination culture
Effect Assessment obtains mechanism difference correlation according to user identity difference and is related to information;
Acquisition field (industry or profession) information, wherein the including but not limited to following letter of the field (industry or profession) information
Breath: field (industry or profession) is related to specific subject, the specific features knowledge point that field (industry or profession) is related to, field (row
Industry or profession) development trend, field (industry or profession) current situation, field (industry or profession) developing history, according to user
Identity difference obtains different field (industry or profession) information;
Step 4: establishing feature tag for user;
According to the school work target or short-term goal information of user, field (industry or profession) information, place machine in conjunction with where user
Structure or unit information, user's operation information, the feature tag for carrying out all related informations extract, and establish " user characteristics label
Library ", wherein the user characteristics tag library can be whole user characteristics label storages, or being will be similar, with semanteme (near synonym)
Feature tag is sorted out, and similar (near synonym) knowledge mapping of feature tag is established, and is formed synonymous by multiple higher level's feature tags
The user characteristics tag library of word knowledge mapping composition;And it is built for all feature tags (higher level's feature tag synonym knowledge mapping)
Vertical study counts, wherein the study is counted as browsing, retrieval counts;
Step 5: user's learning demand data processing:
Target establishes " requirements " of different user feature tag according to demand, and the requirements of the user characteristics label are
The urgent degree for the corresponding feature knowledge point of the described user (the similar knowledge mapping of feature tag) that user needs to learn, establish etc.
Grade quantization;Its described requirement objective is following target but is not limited to following target: in user's short-term goal data, interactive data
The knowledge point data of high frequency discussion, school work target data, used studying space unit (mechanism) cultivate mesh number evidence,
Its described requirements, the condition of degree of increasing demand (counting or rate range), including but not limited to following value of increasing demand
Condition, and do not have sequence requirement: according to user's short-term goal or user's learning objective data, determining the knowledge that user needs to learn
The coverage area (field or professional industry) of point, is the knowledge point in the coverage area, value of increasing demand, the study sky
Between the user of unit (mechanism) training objective and plan need the coverage area (field or professional industry) of knowledge point, for institute
The knowledge point for stating coverage area, value of increasing demand, is user's short-term goal (user's learning objective) and mikey (mechanism) is trained
The user for supporting target and plan needs the coverage area intersection of learning knowledge point, described in coverage area knowledge point, increasing needs
Evaluation, the knowledge point of interactive data medium-high frequency discussion, corresponding knowledge point are increased demand value, and specific culture course is related to knowledge point,
Its described knowledge point is increased demand value, knowledge point involved in industry requirement standard or examination outline standard, according to and it is different
The correspondence of knowledge point should learn degree, answer Grasping level, should understand degree, corresponding to increase corresponding requirements in various degree;
Step 6: user learns to recommend;
User's operation preference profiles label counting is established, the user's operation feature tag is counted as user in studying space
Operation behavior personal preference data, according to user to the preference of correlated knowledge point where different characteristic label, to spy
It levies label and carries out user " preference counting ",
" preference counting " described in it increases the condition that preference counts, including but not limited to the following conditions, and out-of-order requirement: using
The knowledge point that family retrieving information points are related to, the knowledge point increase preference and count, and user browses the knowledge point that information point is related to,
Its described knowledge point increases preference and counts, and user interaction data high frequency discusses knowledge point, increase preference and count, described in belong to
In user group, preference counting is higher to be related to knowledge point, and corresponding increase in various degree corresponds to preference counting;
The individual consumer's feature tag established by former steps has " study count value ", " requirements ", wherein the meter
Numerical value is characterized the browsing quantity of label related term, belonging to requirements be that the needs of this feature label related term are learnt
Demand degree, when recommend learning knowledge point pertinent texts data to user, " the study count value " of same feature tag, instead
To " requirements " for influencing same feature tag recommend the shared weight in calculating, requirements and shared proposed algorithm in user
Weight product, same to user's " preference counting " collectively as study recommend basic data recommend, and have similar characteristics mark
The user of label group does user's classification, forms the user group of similar features label, and a user can belong to multiclass feature mark simultaneously
Sign user group.
3. feature also includes according to claim 1 or described in 2 any one
Its described user learns to recommend, and works as logging in system by user, and when the user does not have too many operation note, system will be according to existing
There are unit (mechanism) data, school work target or the short-term mesh of personal information data, place field (profession or industry), use space
Mark data, by but be not limited to User-CF algorithm meter similar features tagging user group, same user can be simultaneously in multiple
Feature tag user group, and according to but be not limited to collaborative filtering algorithm calculate separately obtain feature mark
It signs the recommending data of user group and is recommended as a result, choosing wherein more each user group recommendation results in the top;
Its described user learns to recommend, and works as logging in system by user, when which has more operation note, can be used but do not limit to
It is calculated in the data that Item-CF algorithm carries out user's operation preference, is passing through but be not limited to collaborative
Filtering algorithm calculates or obtains the data recommendation of similar features tagging user heap sort as a result, recommending to use by history
The result feedback data at family, after integrating all data, the result of every item data selection ranking earlier above is recommended;
Or according to deep learning algorithm, carrying out data training of the client before login should recommend acquisition user logs in next time when
Learning data, user log in when recommend;
The result feedback data of its user's history recommended user, refers to and works as logging in system by user, for system recommendation
Practise the operational feedback of recommending data, if operated and concrete operations content, the feedback data include but be not limited to
Following feedback data, out-of-order requirement: in recommended amount percentage, user's operation selection is pushed away for number of clicks and browsing time
Recommend the secondary preference data of content, the degree of association of user search information knowledge point and user's operation recommendation knowledge point, user
Interactive data high frequency discusses the degree of association of knowledge point and user's operation recommendation knowledge point, userspersonal information and user's operation
The degree of association of content knowledge point, user facility information (field or profession or industry) are associated with user's operation content knowledge point
Degree obtains the staining effect that study is recommended.
4. feature also includes according to claim 1 or described in 2 any one
Its described user learns to recommend, and further includes that study plan is recommended,
Its described study plan is recommended, with the user characteristics label or similar (near synonym) knowledge mapping of the feature tag,
Retrieval entry knowledge point in user search information is similar with the user characteristics label correlated knowledge point or the feature tag
The intersection of the correlated knowledge point of (near synonym) knowledge mapping is got in the user characteristics label correlated knowledge point or the spy
Levy the user search entry in the relevant knowledge point range of similar (near synonym) knowledge mapping of label;
With the feature tag or similar (near synonym) the knowledge mapping data of the feature tag, the user characteristics label
User search entry in the relevant knowledge point range of correlated knowledge point or similar (near synonym) knowledge mapping of the feature tag
Data and user place field (profession or industry) related data, the related culture mesh of the used mechanism in space (unit)
Data, what the basic data that user's short-term goal or school work target data carry out study plan recommendation was recommended as study plan
Basic data carries out recommendation calculating;
Its described study plan is recommended, using the short-term purpose of user or school work purpose as learning objective, when according to actual log
Between and the institution where he works about user study plan or study arrange combine, carry out study plan design;
Its described study plan is recommended, by the feature tag of user or the similar knowledge mapping of the feature tag, with " the demand
Study degree of urgency of the value " as feature tag correlated knowledge point, using but be not limited to described in User-CF algorithm acquisition same type
The study correlated knowledge point of user group, using but be not limited to FP Growth or PrefixSpan algorithm obtain same type described in
The frequent learning Content (knowledge point) of user group obtains learning difficulty, and according to operated by the high user of knowledge point study frequency
The study of the last time knowledge point data or content in, the high Content Selection of study high frequency carries out knowledge point study meter
The preferential recommendation of the high-frequency ranking of recommendation data in drawing;
Its described study plan is recommended, and is the short-term learning objective in place or school work objective plan according to user's actual access time
Different phase, the cultivation stage of unit one belongs to (mechanism), course completes the stage, recommended, and recommendations include, and answers
It practises, previews, the study plan for extending aspect is recommended, and is recommended according to the relevant knowledge of " requirements ", the relevant knowledge
The particular content that point is recommended, with the frequent learning Content of the user group, study frequency high user in the knowledge point is grasped
Study high frequency high data or books and user search letter in the data or content of the study of the last time of the work knowledge point
Know with the user characteristics label correlated knowledge point or the feature tag similar (near synonym) retrieval entry knowledge point in breath
Recommended the intersection knowledge point for knowing the correlated knowledge point of map;
Its described study plan is recommended, different according to user's actual access time, and replaces recommendation, or can check history
Recommendation.
5. a kind of study plan recommended program (plug-in unit) based on active studying space, it is characterized in that:
With user authentication module, user identity effect is carried out;
With User profile acquisition module, user information acquisition is carried out;
With feature tag module is established for user, the foundation of user characteristics label is carried out;
With user's learning demand data processing module, the data processing of user's learning demand is carried out;
Learn recommending module with user, carries out user and learn recommending data processing;
Its described module specific work steps is described in any one of claims 1 or 2.
6. according to claim 5, it is characterized in that:
Also with have books, acquisition of materials suggestion module,
Its described purchasing recommendation module carries out " requirements " higher knowledge point group according to basic data, carries out high frequency study, inspection
The screening of rope knowledge point, and learnt according to the high frequency, the user of retrieval knowledge point learns frequency, establishes sequence, learns frequency
It is higher, but do not have copyrighted books or data to carry out suggestion buying in studying space and recommend, the basic data, including but
It is not limited to categorical data: being recommended according to user's learning knowledge point, user's learning knowledge point feedback data and user search
Data.
7. according to claim 5, it is characterized in that:
Also there is visualization big data analysis to excavate module, by the continuous increasing of the applied unit in the space (mechanism) access user
More, user's learning data amount constantly increases, and can carry out Predictive Analytic by the big data in the space
The analysis of Capabilities(predictability) function, exploratory data analysis Exploratory Data is used to total data first
Analysis shows various data by way of " figure ", so that obtaining arriving for various data visualizes distribution map, when
Between sequence data and transformed variable, recycle hash matrix diagram to illustrate the relationship of variable between any two, and owned
Collect statistics amount, the collect statistics amount includes but is not limited to a under type: calculate mean value, maximum value, minimum value, on
Lower quartile and determine exceptional value, and respectively or summarize corresponding visualization " figure " is presented, described in " figure " include but not office
It is limited to following form: histogram, Stem-and-Leaf Plot, chart, cylinder.
8. a kind of study plan recommender system based on active studying space, needs with server (or server zone) and client
(or client group) is held to rely on as hardware, the server (or server zone) is control server, carries out studying space
Total data storage or processing;Its described client (or client group) is to be linked to control server (or server zone)
Electronic equipment has the feature that
With server, data storage is carried out, data calculate, data manipulation;
With client, user's operation storage can be carried out;
With user authentication module, user identity effect is carried out;
With User profile acquisition module, user information acquisition is carried out;
With feature tag module is established for user, the foundation of user characteristics label is carried out;
With user's learning demand data processing module, the data processing of user's learning demand is carried out;
Learn recommending module with user, carries out user and learn recommending data processing;
Its described module specific work steps is the execution of any one of claims 1 or 2.
9. according to claim 8, it is characterized in that:
Also with have books, acquisition of materials suggestion module,
Its described purchasing recommendation module carries out " requirements " higher knowledge point group according to basic data, carries out high frequency study, inspection
The screening of rope knowledge point, and learnt according to the high frequency, the user of retrieval knowledge point learns frequency, establishes sequence, learns frequency
It is higher, but do not have copyrighted books or data to carry out suggestion buying in studying space and recommend, the basic data, including but
It is not limited to categorical data: being recommended according to user's learning knowledge point, user's learning knowledge point feedback data and user search
Data.
10. according to claim 8, it is characterized in that:
Also there is visualization big data analysis to excavate module, by the continuous increasing of the applied unit in the space (mechanism) access user
More, user's learning data amount constantly increases, and can carry out Predictive Analytic by the big data in the space
The analysis of Capabilities(predictability) function, exploratory data analysis Exploratory Data is used to total data first
Analysis shows various data by way of " figure ", so that obtaining arriving for various data visualizes distribution map, when
Between sequence data and transformed variable, recycle hash matrix diagram to illustrate the relationship of variable between any two, and owned
Collect statistics amount, the collect statistics amount includes but is not limited to a under type: calculate mean value, maximum value, minimum value, on
Lower quartile and determine exceptional value, and respectively or summarize corresponding visualization " figure " is presented, described in " figure " include but not office
It is limited to following form: histogram, Stem-and-Leaf Plot, chart, cylinder.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910021753.3A CN109918561A (en) | 2019-01-10 | 2019-01-10 | A kind of study recommended method of library's (studying space) |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910021753.3A CN109918561A (en) | 2019-01-10 | 2019-01-10 | A kind of study recommended method of library's (studying space) |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109918561A true CN109918561A (en) | 2019-06-21 |
Family
ID=66960228
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910021753.3A Pending CN109918561A (en) | 2019-01-10 | 2019-01-10 | A kind of study recommended method of library's (studying space) |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109918561A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110489648A (en) * | 2019-08-15 | 2019-11-22 | 上海乂学教育科技有限公司 | Education resource dynamic pushing method and system |
CN110597960A (en) * | 2019-09-17 | 2019-12-20 | 香港教育大学 | Personalized online course and occupation bidirectional recommendation method and system |
CN110717048A (en) * | 2019-07-03 | 2020-01-21 | 王妃 | Learning correction method based on knowledge graph |
CN111191122A (en) * | 2019-12-20 | 2020-05-22 | 重庆邮电大学 | Learning resource recommendation system based on user portrait |
CN112733110A (en) * | 2020-11-23 | 2021-04-30 | 融钰四海(武汉)信息科技有限公司 | Enterprise client resource intelligent screening processing service platform |
TWI737046B (en) * | 2019-11-25 | 2021-08-21 | 石廷宇 | Method and system for training reading ability online |
TWI829372B (en) * | 2022-09-30 | 2024-01-11 | 華南商業銀行股份有限公司 | Funding demand forecasting system and forecasting method |
-
2019
- 2019-01-10 CN CN201910021753.3A patent/CN109918561A/en active Pending
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110717048A (en) * | 2019-07-03 | 2020-01-21 | 王妃 | Learning correction method based on knowledge graph |
CN110489648A (en) * | 2019-08-15 | 2019-11-22 | 上海乂学教育科技有限公司 | Education resource dynamic pushing method and system |
CN110597960A (en) * | 2019-09-17 | 2019-12-20 | 香港教育大学 | Personalized online course and occupation bidirectional recommendation method and system |
CN110597960B (en) * | 2019-09-17 | 2022-11-15 | 香港教育大学 | Personalized online course and occupation bidirectional recommendation method and system |
TWI737046B (en) * | 2019-11-25 | 2021-08-21 | 石廷宇 | Method and system for training reading ability online |
CN111191122A (en) * | 2019-12-20 | 2020-05-22 | 重庆邮电大学 | Learning resource recommendation system based on user portrait |
CN112733110A (en) * | 2020-11-23 | 2021-04-30 | 融钰四海(武汉)信息科技有限公司 | Enterprise client resource intelligent screening processing service platform |
TWI829372B (en) * | 2022-09-30 | 2024-01-11 | 華南商業銀行股份有限公司 | Funding demand forecasting system and forecasting method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109918561A (en) | A kind of study recommended method of library's (studying space) | |
Padhy et al. | The survey of data mining applications and feature scope | |
CN110533974B (en) | Intelligent volume assembling method and system and computer readable storage medium | |
CN109074642A (en) | machine learning device | |
ANDRADE et al. | Active Methodology, Educational Data Mining and Learning Analytics: A Systematic Mapping Study. | |
Zoric | Benefits of educational data mining | |
Patil et al. | Prediction system for student performance using data mining classification | |
Shao et al. | An optimized mining algorithm for analyzing students’ learning degree based on dynamic data | |
CN109992676A (en) | Across the media resource search method of one kind and searching system | |
CN112487290A (en) | Internet precision teaching method and system based on big data and artificial intelligence | |
Smith et al. | RANGEPACK: the philosophy underlying the development of a microcomputer-based decision support system for pastoral land management | |
Palshikar et al. | Automatic Shortlisting of Candidates in Recruitment. | |
Aitdaoud et al. | A New Pre-Processing Approach Based on Clustering Users Traces According to their Learning Styles in Moodle LMS. | |
US20170193620A1 (en) | Associate a learner and learning content | |
CN114662779A (en) | Student portrait-based programming score prediction method and system | |
Saleem et al. | Data mining course in information system department-case study of King Abdulaziz University | |
Cabo | Use of Machine Learning to Identify Predictors of Student Performance in Writing Viable Computer Programs with Repetition Loops and Methods | |
Kardan et al. | Learner clustering and association rule mining for content recommendation in self-regulated learning | |
Trivedi | Machine Learning Fundamental Concepts | |
Zeng | RETRACTED ARTICLE: A comparative study of the market demand for Chinese proficiency test preparation books: evidence from e-commerce data | |
Siddiqui et al. | Predicting the Student Performance via Machine Leaning Techniques | |
Christy et al. | The Review on Data Mining Techniques and its Applications | |
Wang et al. | Personalized Recommendation Method of Ideological and Political Education Resources Based on Data Mining | |
Dickelman | Learning and Performance Science for Digital Transformation | |
Liu | Design and implementation of teaching assessment system in higher vocational colleges based on association rule algorithm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190621 |