CN108920663A - A kind of method of automatic recommendation item content - Google Patents
A kind of method of automatic recommendation item content Download PDFInfo
- Publication number
- CN108920663A CN108920663A CN201810726202.2A CN201810726202A CN108920663A CN 108920663 A CN108920663 A CN 108920663A CN 201810726202 A CN201810726202 A CN 201810726202A CN 108920663 A CN108920663 A CN 108920663A
- Authority
- CN
- China
- Prior art keywords
- examination question
- knowledge point
- answer
- steps
- examination
- 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.)
- Granted
Links
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The present invention relates to machine learning techniques fields, disclose a kind of automatic method for recommending item content.It creates through the invention, it provides and a kind of is calculated using the information for being accurate to topic internal structure, the topic recommended method for being accurate to knowledge point level can be automatically performed, on the one hand interfere in the case where data-oriented completely without manpower, topic recommendation can be automatically finished, on the other hand, by introducing the ability information of the detailed information of step and user's details in topic, so that the item content recommended is more in line with user's needs.It further, since introducing enchancement factor in the method, can solve the problems, such as to repeat to recommend under same scene, can both ensure that the correlation of topic selection with adaptation degree according to adaptation degree random selection, and in turn avoid pushing away the determining awkward situation of topic under same scene.
Description
Technical field
The invention belongs to machine learning techniques fields, and in particular to a kind of automatic recommendation topic that can be applied to education sector
The method of content.
Background technique
In education sector, recommend most suitable exercise to student, can effectively improve student in determination and do topic quantity
In the case of learning effect.
Under traditional approach, this part is generally completed by explanation teacher.Teacher is explained based on oneself understanding to student
Suitable topic is selected to practice with the entitled student in oneself impression.Firstly, being topic quantity, teacher in view of student
Directly recommend to be difficult to accomplish that one problem of every completion recognizes one of new topic of recommendation with regard to real-time update later.Secondly, in view of religion
Teacher's memory is limited, is difficult to remember ability of each student in terms of each knowledge, it is also difficult to remember all topics, therefore
This way of recommendation can only guarantee the main demand of student and the main matching for wanting to recommend topic of teacher, it is difficult to ensure that for a long time
It is still matched under use.
Common proposed algorithm is generally based on " thing that user likes is similar " and " similar user likes class
As content " the two basis assume carry out, so that recommended user likes the approximate content of content and liking for approximated user
Content.However recommend the upper program and infeasible in topic.It is done in topic process in student, because valuable topic compares mostly
Difficulty, student tend not to like those topics for really facilitating to learn.Therefore topic can not be applied to based on the hypothesis liked
During mesh is recommended, the reliability of such proposed algorithm is not known where to begin naturally.
General automation topic recommendation is that topic type, classification and/or difficulty based on topic carry out.This mode can be with
Ensure that the topic recommended meets the requirement of student from topic general orientation, but actually recommends precision still very poor.It is first
Scheme based on classification is not fully reasonable, and topic often relates to the knowledge of many aspects, is referred to some class by force
Do not have lost recommendation accuracy inherently.Simultaneously as difficulty is often to be directed to topic mark, therefore be directed to student ability
Matching can only accomplish topic level, and influence each other for the difficulty between the various pieces of topic inside and often ignore.Due to
Upper two o'clock reason, the data accuracy of proposed algorithm can only reach topic level, recommend precision often worse.
Meanwhile traditional proposed algorithm mostly determines topic based entirely on index size, which results in phase same level
In the case where student be more easily seen identical topic, and the actual water of student can not effectively be promoted by repeating to do same problem
It is flat, therefore completely specified algorithm can not adapt to student and make the needs inscribed.
Summary of the invention
In order to solve the above problem in the presence of the prior art, it is an object of that present invention to provide one kind can high collocation degree
The automatic method for recommending item content.
The technical scheme adopted by the invention is as follows：
A kind of method of automatic recommendation item content, includes the following steps：
S100. the examination of knowledge point master data, user's current ability data and several examination questions is stored in advance in the database
Inscribe master data, wherein the knowledge point master data includes the topological order for expressing the successive customs examination system in all knowledge points, institute
Stating user's current ability data includes to be examination question set and user in the current power of each knowledge point, and the examination question is basic
Data include item content, the model answer containing at least two answer steps, the skill of solving a problem for corresponding at least one answer step
The knowledge point of skilful and corresponding each answer step and weight coefficient；
S101. object knowledge point set, target calculus technique set and target power value set by user are obtained；
S102. it is directed to each examination question, according to the object knowledge point set, the target calculus technique set, the mesh
Ability value, the knowledge point master data, the examination question master data of user's current ability data and the examination question are marked, is counted respectively
Calculate the following index of the examination question：The ratio F of target outer knowledge point quantity and knowledge point quantity in examination question in examination question1, know in examination question
Know the ratio F of knowledge point quantity in point topological order span and library2, user do not grasp in examination question knowledge in knowledge point quantity and examination question
The ratio F of point quantity3, user do not grasp the ratio F of knowledge point quantity in knowledge point topological order most posteriority position and library in examination question4、
The ratio F of answer step weight in step weight and examination question is answered in examination question outside target calculus technique5, non-targeted in examination question solve a problem
Skill answers the ratio F of answer step weight in step weight and examination question6, answered in step weight and examination question outside target in examination question
Answer the ratio F of step weight7And/or the probability ratio F of user's mistake knowledge point outside target8；
S103. it is directed to each examination question, all indexs obtained in step s 102 are spliced into a column vector；
S104. be directed to each examination question, by column vector with it is corresponding most adaptation examination question ideal column vector subtract each other, obtain error to
Amount, wherein the ideal column vector is 0 vector or is preset according to experience with students；
S105. it is directed to each examination question, adaptation coefficients are obtained according to error vector；
S106. the adaptation coefficients for summarizing all examination questions calculate the current select probability of each examination question according to following formula：
In formula, P (i) is the current select probability of i-th (i=1,2,3 ..., N) a examination question,For i-th (i=1,2,
3 ..., N) a examination question adaptation coefficients, N is examination question sum in database；
S107. based on the current select probability of each examination question, an examination question is randomly choosed, the item content of the examination question is made
To recommend topic.
Optimization, in the step S102, the outer knowledge point of target in the examination question of examination question is calculated in accordance with the following steps
The ratio F of knowledge point quantity in quantity and examination question1：
S2101. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2102. knowledge point quantity in knowledge point quantity and examination question is calculated in the examination question outside target according to following formula
Ratio F1：
In formula,Cover knowledge point set to belong to the examination question and is not belonging to the knowledge of the object knowledge point set
Point sum,To belong to the knowledge point sum that the examination question covers knowledge point set.
Optimization, in the step S102, knowledge point topological order in the examination question of examination question is calculated in accordance with the following steps
The ratio F of knowledge point quantity in span and library2：
S2201. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2202. it according to the knowledge point master data, determines that the examination question is covered in knowledge point set and is in topological order most
The most priori knowledge point of first position and most posteriority knowledge point in topological order rearmost position；
S2203. the ratio of knowledge point quantity in knowledge point topological order span and library in the examination question is calculated according to following formula
Value F2：
In formula, M is the knowledge point sum in database, XMaxLFor the topological serial number of the most posteriority knowledge point, XMinFFor institute
State the topological serial number of most priori knowledge point, XMaxL-XMinFFor topological order span in knowledge point in examination question.
Optimization, in the step S102, the user for calculating examination question in accordance with the following steps, which does not grasp in examination question, to be known
Know the ratio F of knowledge point quantity in point quantity and examination question3：
S2301. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2302. cover each knowledge point in knowledge point set for the examination question, from user's current ability data
In find corresponding current power, if the current power be lower than the target power value, using the knowledge point as use
Knowledge point in examination question is not grasped in family；
S2303. summarize to obtain user and do not grasp knowledge point set in examination question, and calculate the user not according to following formula
Grasp the ratio F of knowledge point quantity and knowledge point quantity in examination question in examination question3：
In formula,To belong to the knowledge point sum that the user does not grasp knowledge point set in examination question,To belong to
State the knowledge point sum that examination question covers knowledge point set.
Optimization, in the step S102, the user for calculating examination question in accordance with the following steps, which does not grasp in examination question, to be known
Know the ratio F of knowledge point quantity in point topological order most posteriority position and library4：
S2401. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2402. cover each knowledge point in knowledge point set for the examination question, from user's current ability data
In find corresponding current power, if the current power be lower than the target power value, using the knowledge point as use
Knowledge point in examination question is not grasped in family；
S2403. summarize to obtain user and do not grasp knowledge point set in examination question；
S2404. it according to the knowledge point master data, determines that the user does not grasp in examination question and is in knowledge point set
The most posteriority of topological order rearmost position does not grasp knowledge point；
S2405. the user is calculated according to following formula do not grasp knowledge point topological order most posteriority position and library in examination question
The ratio F of interior knowledge point quantity4：
In formula, M is the knowledge point sum in database,The topological serial number of knowledge point is not grasped for the most posteriority.
Optimization, in the step S102, target calculus technique in the examination question of examination question is calculated in accordance with the following steps
The ratio F of answer step weight in outer answer step weight and examination question5：
S2501. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2502. for each answer step in the examination question in answer set of steps, according to the examination question base of corresponding examination question
Notebook data judges whether the answer step corresponds to the target calculus technique in the target calculus technique set, if not corresponding to,
Step is answered outside using the answer step as target calculus technique in examination question；
S2503. summarize to obtain in examination question and answer set of steps outside target calculus technique, then calculate institute according to following formula
State the ratio F for answering answer step weight in step weight and examination question in examination question outside target calculus technique5：
In formula, wiTo be answered the in set of steps outside target calculus technique in the examination questionIt is a
Answer the weight coefficient of step, wjTo be answered the in set of steps in the examination questionA answer step
Weight coefficient,To belong in the examination question answer step sum for answering set of steps outside target calculus technique,For
Belong to the answer step sum of answer set of steps in the examination question；
And/or non-targeted calculus technique answer step weight and examination in the examination question of calculating examination question in accordance with the following steps
The ratio F of answer step weight in topic6：
S2601. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2602. it is directed to each answer step for being corresponding with calculus technique in the examination question in answer set of steps, according to
The target that the examination question master data of corresponding examination question judges whether the answer step corresponds in the target calculus technique set is solved a problem
Skill answers step for the answer step as calculus technique non-targeted in examination question if not corresponding to；
S2603. summarize to obtain non-targeted calculus technique answer set of steps in examination question, then calculate institute according to following formula
State the ratio F of non-targeted calculus technique answer step weight and the interior answer step weight of examination question in examination question6：
In formula, wkIt is answered the in set of steps for calculus technique non-targeted in the examination questionIt is a
Answer the weight coefficient of step, wjTo be answered the in set of steps in the examination questionA answer step
Weight coefficient,To belong to the answer step sum that non-targeted calculus technique in the examination question answers set of steps,For
Belong to the answer step sum of answer set of steps in the examination question.
Optimization, in the step S102, calculates answer step in the examination question of examination question outside target in accordance with the following steps
The ratio F of answer step weight in rapid weight and examination question7：
S2701. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2702. for each answer step in the examination question in answer set of steps, according to the examination question base of corresponding examination question
Notebook data judges whether the answer step corresponds to the knowledge point in the object knowledge point set, if not corresponding to, by the answer
Step answers step as target in examination question outside；
S2703. summarize to obtain in examination question and answer set of steps outside target, then calculated in the examination question according to following formula
The ratio F of answer step weight in step weight and examination question is answered outside target7：
In formula, wsTo be answered the in set of steps outside target in the examination questionA answer step
Weight coefficient, wjTo be answered the in set of steps in the examination questionThe weight system of a answer step
Number,To belong in the examination question answer step sum for answering set of steps outside target,It is solved to belong in the examination question
Answer the answer step sum of set of steps.
Optimization, further include following steps after the step S104：
S401. error vector is multiplied with preset weight row vector, completes the adjustment to the error vector.
Optimization, in the step S105, the method that adaptation coefficients are obtained according to error vector, including walk as follows
Suddenly：
S501. using two norms of error vector as adaptation coefficients.
Optimization, further include following steps before the step S106：
S600. it is directed to each examination question, examination question set has been done according in user's current ability data, has counted the user of the examination question
Topic number is done, adaptation coefficients correct upwards shown in following formula：
a′i=γtai
In formula, a 'iFor adaptation coefficients after the amendment of i-th (i=1,2,3 ..., N) a examination question, aiFor i-th (i=1,2,
3 ..., N) a examination question amendment before adaptation coefficients, γ is upward correction factor, and t is to do topic number in the user of the examination question.
Beneficial effects of the present invention are：
(1) the invention provide it is a kind of calculated using the information for being accurate to topic internal structure, can be automatic
Complete the topic recommended method for being accurate to knowledge point level, i.e., it is on the one hand dry completely without manpower in the case where data-oriented
It relates to, so that it may topic recommendation is automatically finished, on the other hand, by introducing the detailed information of step and user's details in topic
Ability information so that recommend item content be more in line with user's needs；
(2) by distinguishing target collection and highest priority, the recommendation to topic may make more to be bonded scene needs；
(3) by weighting adjustment demand to same group of index, in need while can avoid overlapping development to take into account institute in love
The problem of condition；
(4) this method can take into account new topic practice and review with old topic, can avoid the monotonicity that topic is recommended under same capabilities, i.e.,
Repetition is avoided by the amendment that topic number is done in addition to inscribe, but remains the possibility of recommendation, to realize review；
(5) this method introduces enchancement factor, can solve the problems, such as to repeat to recommend under same scene, can be according to adaptation journey
Random selection is spent, both ensure that the correlation of topic selection and adaptation degree, in turn avoids pushing away the embarrassed of topic determination under same scene
Border.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is the method flow schematic diagram of automatic recommendation item content provided by the invention.
Specific embodiment
With reference to the accompanying drawing and specific embodiment the present invention is further elaborated.It should be noted that for this
The explanation of a little way of example is used to help understand the present invention, but and does not constitute a limitation of the invention.
The terms "and/or", only a kind of incidence relation for describing affiliated partner, indicates that there may be three kinds of passes
System, for example, A and/or B, can indicate：Individualism A, individualism B exist simultaneously tri- kinds of situations of A and B, the terms
"/and " it is to describe another affiliated partner relationship, indicate may exist two kinds of relationships, for example, A/ and B, can indicate：Individually deposit
In A, two kinds of situations of individualism A and B, in addition, character "/" herein, typicallying represent forward-backward correlation object is a kind of "or" pass
System.
Embodiment one
As shown in Figure 1, the method for the automatic recommendation item content provided in this embodiment, includes the following steps.
S100. the examination of knowledge point master data, user's current ability data and several examination questions is stored in advance in the database
Inscribe master data, wherein the knowledge point master data includes the topological order for expressing the successive customs examination system in all knowledge points, institute
Stating user's current ability data includes to be examination question set and user in the current power of each knowledge point, and the examination question is basic
Data include item content, the model answer containing at least two answer steps, the skill of solving a problem for corresponding at least one answer step
The knowledge point of skilful and corresponding each answer step and weight coefficient.
In the step S100, the successive customs examination system in knowledge point refers to for knowledge point A and knowledge point B, if necessary
First learning knowledge point A, could learning knowledge point B again, then knowledge point A and knowledge point B has successive customs examination system：Knowledge point A is first
Knowledge point is tested, knowledge point B is aposterior knowledge point.So as to obtain having certain length and expression all knowledge point elder generations posteriority
The topological order of relationship.In addition, the examination question master data can also include the degree-of-difficulty factor and answer of corresponding each answer step
Successive customs examination system between step.
S101. object knowledge point set, target calculus technique set and target power value set by user are obtained.
In the step S101, the object knowledge point set, the target calculus technique set and the target energy
Force value is all from human-computer interaction interface, wherein can only have in the object knowledge point set and the target calculus technique set
One is empty set.
S102. it is directed to each examination question, according to the object knowledge point set, the target calculus technique set, the mesh
Ability value, the knowledge point master data, the examination question master data of user's current ability data and the examination question are marked, is counted respectively
Calculate the following index of the examination question：The ratio F of target outer knowledge point quantity and knowledge point quantity in examination question in examination question1, know in examination question
Know the ratio F of knowledge point quantity in point topological order span and library2, user do not grasp in examination question knowledge in knowledge point quantity and examination question
The ratio F of point quantity3, user do not grasp the ratio F of knowledge point quantity in knowledge point topological order most posteriority position and library in examination question4、
The ratio F of answer step weight in step weight and examination question is answered in examination question outside target calculus technique5, non-targeted in examination question solve a problem
Skill answers the ratio F of answer step weight in step weight and examination question6, answered in step weight and examination question outside target in examination question
Answer the ratio F of step weight7And/or the probability ratio F of user's mistake knowledge point outside target8。
In the step S102, it can be, but not limited to calculate in accordance with the following steps in the examination question of examination question and know outside target
Know the ratio F of knowledge point quantity in point quantity and examination question1：
S2101. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2102. knowledge point quantity in knowledge point quantity and examination question is calculated in the examination question outside target according to following formula
Ratio F1：
In formula,Cover knowledge point set to belong to the examination question and is not belonging to the knowledge of the object knowledge point set
Point sum,To belong to the knowledge point sum that the examination question covers knowledge point set.In the step S2101, due to described
It include item content, the model answer containing at least two answer steps and corresponding each answer step in examination question master data
Rapid knowledge point, therefore the examination question that can easily obtain the examination question covers knowledge point set.
In the step S102, it can be, but not limited to calculate knowledge point in the examination question of examination question in accordance with the following steps and open up
Flutter the ratio F of knowledge point quantity in sequence span and library2：
S2201. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2202. it according to the knowledge point master data, determines that the examination question is covered in knowledge point set and is in topological order most
The most priori knowledge point of first position and most posteriority knowledge point in topological order rearmost position；
S2203. the ratio of knowledge point quantity in knowledge point topological order span and library in the examination question is calculated according to following formula
Value F2：
In formula, M is the knowledge point sum in database, XMaxLFor the topological serial number of the most posteriority knowledge point, XMinFFor institute
State the topological serial number of most priori knowledge point, XMaxL-XMinFFor topological order span in knowledge point in examination question.Since the knowledge point is basic
Data include the topological order for expressing the successive customs examination system in all knowledge points, therefore can easily determine most priori knowledge
Point, most posteriority knowledge point and corresponding topological serial number.
In the step S102, the user that can be, but not limited to calculate examination question in accordance with the following steps does not grasp examination question
The ratio F of knowledge point quantity in interior knowledge point quantity and examination question3：
S2301. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2302. cover each knowledge point in knowledge point set for the examination question, from user's current ability data
In find corresponding current power, if the current power be lower than the target power value, using the knowledge point as use
Knowledge point in examination question is not grasped in family；
S2303. summarize to obtain user and do not grasp knowledge point set in examination question, and calculate the user not according to following formula
Grasp the ratio F of knowledge point quantity and knowledge point quantity in examination question in examination question3：
In formula,To belong to the knowledge point sum that the user does not grasp knowledge point set in examination question,To belong to
State the knowledge point sum that examination question covers knowledge point set.Since user's current ability data include user in each knowledge point
Current power, therefore be easy to judge that can some knowledge point be grasped by user.
In the step S102, the user that can be, but not limited to calculate examination question in accordance with the following steps does not grasp examination question
The ratio F of knowledge point quantity in interior knowledge point topological order most posteriority position and library4：
S2401. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2402. cover each knowledge point in knowledge point set for the examination question, from user's current ability data
In find corresponding current power, if the current power be lower than the target power value, using the knowledge point as use
Knowledge point in examination question is not grasped in family；
S2403. summarize to obtain user and do not grasp knowledge point set in examination question；
S2404. it according to the knowledge point master data, determines that the user does not grasp in examination question and is in knowledge point set
The most posteriority of topological order rearmost position does not grasp knowledge point；
S2405. the user is calculated according to following formula do not grasp knowledge point topological order most posteriority position and library in examination question
The ratio F of interior knowledge point quantity4：
In formula, M is the knowledge point sum in database,The topological serial number of knowledge point is not grasped for the most posteriority.
In the step S102, it can be, but not limited to calculate target in the examination question of examination question in accordance with the following steps and solve a problem
The ratio F of answer step weight in step weight and examination question is answered outside skill5：
S2501. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2502. for each answer step in the examination question in answer set of steps, according to the examination question base of corresponding examination question
Notebook data judges whether the answer step corresponds to the target calculus technique in the target calculus technique set, if not corresponding to,
Step is answered outside using the answer step as target calculus technique in examination question；
S2503. summarize to obtain in examination question and answer set of steps outside target calculus technique, then calculate institute according to following formula
State the ratio F for answering answer step weight in step weight and examination question in examination question outside target calculus technique5：
In formula, wiTo be answered the in set of steps outside target calculus technique in the examination questionIt is a
Answer the weight coefficient of step, wjTo be answered the in set of steps in the examination questionA answer step
Weight coefficient,To belong in the examination question answer step sum for answering set of steps outside target calculus technique,For
Belong to the answer step sum of answer set of steps in the examination question；
In the step S102, it can be, but not limited to calculate non-targeted solution in the examination question of examination question in accordance with the following steps
Inscribe the ratio F of answer step weight in skill answer step weight and examination question6：
S2601. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2602. it is directed to each answer step for being corresponding with calculus technique in the examination question in answer set of steps, according to
The target that the examination question master data of corresponding examination question judges whether the answer step corresponds in the target calculus technique set is solved a problem
Skill answers step for the answer step as calculus technique non-targeted in examination question if not corresponding to；
S2603. summarize to obtain non-targeted calculus technique answer set of steps in examination question, then calculate institute according to following formula
State the ratio F of non-targeted calculus technique answer step weight and the interior answer step weight of examination question in examination question6：
In formula, wkIt is answered the in set of steps for calculus technique non-targeted in the examination questionIt is a
Answer the weight coefficient of step, wjTo be answered the in set of steps in the examination questionA answer step
Weight coefficient,To belong to the answer step sum that non-targeted calculus technique in the examination question answers set of steps,For
Belong to the answer step sum of answer set of steps in the examination question.
In the step S102, it can be, but not limited to calculate target in the examination question of examination question in accordance with the following steps and solve outside
Answer the ratio F of answer step weight in step weight and examination question7：
S2701. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2702. for each answer step in the examination question in answer set of steps, according to the examination question base of corresponding examination question
Notebook data judges whether the answer step corresponds to the knowledge point in the object knowledge point set, if not corresponding to, by the answer
Step answers step as target in examination question outside；
S2703. summarize to obtain in examination question and answer set of steps outside target, then calculated in the examination question according to following formula
The ratio F of answer step weight in step weight and examination question is answered outside target7：
In formula, wsTo be answered the in set of steps outside target in the examination questionA answer step
Weight coefficient, wjTo be answered the in set of steps in the examination questionThe weight system of a answer step
Number,To belong in the examination question answer step sum for answering set of steps outside target,It is solved to belong in the examination question
Answer the answer step sum of set of steps.
In the step S102, the probability ratio F of user's mistake knowledge point outside target8Refer in examination question and
The answer step for not corresponding to knowledge point in object knowledge point set, according to user's current ability data, (it contains user and exists
The current power of each knowledge point) and the examination question master data (its degree-of-difficulty factor for containing corresponding each answer step),
Prediction calculates user's doing to probability in the answer step, and last join probability opinion can calculate user and do wrong at least one
Probability --- the i.e. probability of user's mistake knowledge point outside target of these answer steps.
S103. it is directed to each examination question, all indexs obtained in step s 102 are spliced into a column vector.
S104. be directed to each examination question, by column vector with it is corresponding most adaptation examination question ideal column vector subtract each other, obtain error to
Amount, wherein the ideal column vector is 0 vector or is preset according to experience with students.
After the step S104, the problem of taking into account all situations in need while in order to avoid overlapping development, also
Include the following steps：
S401. error vector is multiplied with preset weight row vector, completes the adjustment to the error vector.
S105. it is directed to each examination question, adaptation coefficients are obtained according to error vector.
In the step S105, the adaptation coefficients describe the examination question with the gap that is most adapted between examination question, i.e., it is suitable
Distribution coefficient is smaller, the corresponding more suitable present case of examination question.Optimization, it is described that the side of adaptation coefficients is obtained according to error vector
Method can be, but not limited to include the following steps：
S501. using two norms of error vector as adaptation coefficients.
S106. the adaptation coefficients for summarizing all examination questions calculate the current select probability of each examination question according to following formula：
In formula, P (i) is the current select probability of i-th (i=1,2,3 ..., N) a examination question,For i-th (i=1,2,
3 ..., N) a examination question adaptation coefficients, N is examination question sum in database.
Before the step S106, new topic practice is reviewed with old topic in order to balance, and topic under same capabilities is avoided to recommend
Monotonicity, further include following steps：
S600. it is directed to each examination question, examination question set has been done according in user's current ability data, has counted the user of the examination question
Topic number is done, adaptation coefficients correct upwards shown in following formula：
a′i=γtai
In formula, a 'iFor adaptation coefficients after the amendment of i-th (i=1,2,3 ..., N) a examination question, aiFor i-th (i=1,2,
3 ..., N) a examination question amendment before adaptation coefficients, γ is upward correction factor, and t is to do topic number in the user of the examination question.Thus
The amendment of topic number is done by being added, avoidable repetition is inscribed, but remains the possibility of recommendation, to realize review purpose.
S107. based on the current select probability of each examination question, an examination question is randomly choosed, the item content of the examination question is made
To recommend topic.
In the step S107, due to introducing enchancement factor, it can solve the problems, such as to repeat to recommend under same scene, i.e.,
It can both ensure that the correlation of topic selection with adaptation degree according to adaptation degree random selection, in turn avoid under same scene
Push away the determining awkward situation of topic.
To sum up, it using the method for recommending item content provided by the present embodiment automatically, has the following technical effect that：
(1) present embodiments provide it is a kind of calculated using the information for being accurate to topic internal structure, can be automatic complete
It is at the topic recommended method for being accurate to knowledge point level, i.e., on the one hand dry completely without manpower in the case where data-oriented
It relates to, so that it may topic recommendation is automatically finished, on the other hand, by introducing the detailed information of step and user's details in topic
Ability information so that recommend item content be more in line with user's needs；
(2) by distinguishing target collection and highest priority, the recommendation to topic may make more to be bonded scene needs；
(3) by weighting adjustment demand to same group of index, in need while can avoid overlapping development to take into account institute in love
The problem of condition；
(4) this method can take into account new topic practice and review with old topic, can avoid the monotonicity that topic is recommended under same capabilities, i.e.,
Repetition is avoided by the amendment that topic number is done in addition to inscribe, but remains the possibility of recommendation, to realize review；
(5) this method introduces enchancement factor, can solve the problems, such as to repeat to recommend under same scene, can be according to adaptation journey
Random selection is spent, both ensure that the correlation of topic selection and adaptation degree, in turn avoids pushing away the embarrassed of topic determination under same scene
Border.
The present invention is not limited to above-mentioned optional embodiment, anyone can show that other are each under the inspiration of the present invention
The product of kind form.Above-mentioned specific embodiment should not be understood the limitation of pairs of protection scope of the present invention, protection of the invention
Range should be subject to be defined in claims, and specification can be used for interpreting the claims.
Claims (10)
1. a kind of automatic method for recommending item content, which is characterized in that include the following steps：
S100. the examination question base of knowledge point master data, user's current ability data and several examination questions is stored in advance in the database
Notebook data, wherein the knowledge point master data includes the topological order for expressing the successive customs examination system in all knowledge points, the use
Family current ability data include to be examination question set and user in the current power of each knowledge point, the examination question master data
Calculus technique comprising item content, the model answer containing at least two answer steps, at least one corresponding answer step with
And knowledge point and the weight coefficient of corresponding each answer step；
S101. object knowledge point set, target calculus technique set and target power value set by user are obtained；
S102. it is directed to each examination question, according to the object knowledge point set, the target calculus technique set, the target energy
Force value, the knowledge point master data, the examination question master data of user's current ability data and the examination question, calculate separately this
The following index of examination question：The ratio F of target outer knowledge point quantity and knowledge point quantity in examination question in examination question1, knowledge point in examination question
The ratio F of knowledge point quantity in topological order span and library2, user do not grasp in examination question that knowledge is counted in knowledge point quantity and examination question
The ratio F of amount3, user do not grasp the ratio F of knowledge point quantity in knowledge point topological order most posteriority position and library in examination question4, examination question
The ratio F of answer step weight in step weight and examination question is answered outside interior target calculus technique5, non-targeted calculus technique in examination question
Answer the ratio F of answer step weight in step weight and examination question6, answer in step weight and examination question in examination question and answer outside target
The ratio F of step weight7And/or the probability ratio F of user's mistake knowledge point outside target8；
S103. it is directed to each examination question, all indexs obtained in step s 102 are spliced into a column vector；
S104. it is directed to each examination question, column vector is subtracted each other with the corresponding ideal column vector for being most adapted to examination question, obtains error vector,
Wherein, the ideal column vector is 0 vector or is preset according to experience with students；
S105. it is directed to each examination question, adaptation coefficients are obtained according to error vector；
S106. the adaptation coefficients for summarizing all examination questions calculate the current select probability of each examination question according to following formula：
In formula, P (i) is the current select probability of i-th (i=1,2,3 ..., N) a examination question,For i-th (i=1,2,3 ..., N) it is a
The adaptation coefficients of examination question, N are examination question sum in database；
S107. based on the current select probability of each examination question, an examination question is randomly choosed, using the item content of the examination question as pushing away
Recommend topic.
2. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S102
In, the ratio of knowledge point quantity in the outer knowledge point quantity of target in the examination question of examination question and examination question is calculated in accordance with the following steps
F1：
S2101. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2102. the ratio of knowledge point quantity and knowledge point quantity in examination question outside target in the examination question is calculated according to following formula
F1：
In formula,For belong to the examination question cover knowledge point set and be not belonging to the object knowledge point set knowledge point it is total
Number,To belong to the knowledge point sum that the examination question covers knowledge point set.
3. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S102
In, the ratio F of knowledge point topological order span and knowledge point quantity in library in the examination question of examination question is calculated in accordance with the following steps2：
S2201. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2202. it according to the knowledge point master data, determines that the examination question is covered and is in topological order position at first in knowledge point set
The most priori knowledge point set and the most posteriority knowledge point in topological order rearmost position；
S2203. the ratio F of knowledge point topological order span and knowledge point quantity in library in the examination question is calculated according to following formula2：
In formula, M is the knowledge point sum in database, XMaxLFor the topological serial number of the most posteriority knowledge point, XMinFFor it is described most
The topological serial number of priori knowledge point, XMaxL-XMinFFor topological order span in knowledge point in examination question.
4. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S102
In, the user for calculating examination question in accordance with the following steps does not grasp the ratio of knowledge point quantity in knowledge point quantity and examination question in examination question
Value F3：
S2301. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2302. cover each knowledge point in knowledge point set for the examination question, looked into from user's current ability data
Corresponding current power is found, if the current power is lower than the target power value, not using the knowledge point as user
Grasp knowledge point in examination question；
S2303. summarize to obtain user and do not grasp knowledge point set in examination question, and calculate the user according to following formula and do not grasp
The ratio F of knowledge point quantity and knowledge point quantity in examination question in examination question3：
In formula,To belong to the knowledge point sum that the user does not grasp knowledge point set in examination question,To belong to the examination
Topic covers the knowledge point sum of knowledge point set.
5. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S102
In, the user for calculating examination question in accordance with the following steps does not grasp knowledge point topological order most posteriority position and knowledge in library in examination question
The ratio F of point quantity4：
S2401. according to the examination question master data of corresponding examination question, summarize to obtain examination question and cover knowledge point set；
S2402. cover each knowledge point in knowledge point set for the examination question, looked into from user's current ability data
Corresponding current power is found, if the current power is lower than the target power value, not using the knowledge point as user
Grasp knowledge point in examination question；
S2403. summarize to obtain user and do not grasp knowledge point set in examination question；
S2404. it according to the knowledge point master data, determines that the user does not grasp and is in topology in examination question in knowledge point set
The most posteriority of sequence rearmost position does not grasp knowledge point；
S2405. it calculates the user according to following formula and does not grasp in examination question and know in knowledge point topological order most posteriority position and library
Know the ratio F of point quantity4：
In formula, M is the knowledge point sum in database,The topological serial number of knowledge point is not grasped for the most posteriority.
6. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S102
In, it calculates to answer outside target calculus technique in the examination question of examination question in accordance with the following steps and answers step in step weight and examination question
The ratio F of weight5：
S2501. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2502. for each answer step in the examination question in answer set of steps, according to the examination question basic number of corresponding examination question
It is judged that the target calculus technique whether the answer step corresponds in the target calculus technique set should if not corresponding to
Answer step answers step as target calculus technique in examination question outside；
S2503. summarize to obtain in examination question and answer set of steps outside target calculus technique, then calculate the examination according to following formula
The ratio F of answer step weight in step weight and examination question is answered in inscribing outside target calculus technique5：
In formula, wiTo be answered the in set of steps outside target calculus technique in the examination questionA answer
The weight coefficient of step, wjTo be answered the in set of steps in the examination questionThe power of a answer step
Weight coefficient,To belong in the examination question answer step sum for answering set of steps outside target calculus technique,To belong to
The answer step sum of answer set of steps in the examination question；
And/or in accordance with the following steps in the examination question of calculating examination question in non-targeted calculus technique answer step weight and examination question
Answer the ratio F of step weight6：
S2601. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2602. for each answer step for being corresponding with calculus technique in the examination question in answer set of steps, according to correspondence
The examination question master data of examination question judges whether the answer step corresponds to the target calculus technique in the target calculus technique set,
If not corresponding to, step is answered using the answer step as calculus technique non-targeted in examination question；
S2603. summarize to obtain non-targeted calculus technique answer set of steps in examination question, then calculate the examination according to following formula
The ratio F of non-targeted calculus technique answer step weight and the interior answer step weight of examination question in topic6：
In formula, wkIt is answered the in set of steps for calculus technique non-targeted in the examination questionA answer
The weight coefficient of step, wjTo be answered the in set of steps in the examination questionThe power of a answer step
Weight coefficient,To belong to the answer step sum that non-targeted calculus technique in the examination question answers set of steps,To belong to
The answer step sum of answer set of steps in the examination question.
7. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S102
In, the ratio for answering answer step weight in step weight and examination question in the examination question of examination question outside target is calculated in accordance with the following steps
Value F7：
S2701. according to the examination question master data of corresponding examination question, summarize to obtain and answer set of steps in examination question；
S2702. for each answer step in the examination question in answer set of steps, according to the examination question basic number of corresponding examination question
It is judged that whether the answer step corresponds to the knowledge point in the object knowledge point set, if not corresponding to, by the answer step
Step is answered outside as target in examination question；
S2703. summarize to obtain in examination question and answer set of steps outside target, then calculate target in the examination question according to following formula
The ratio F of answer step weight in outer answer step weight and examination question7：
In formula, wsTo be answered the in set of steps outside target in the examination questionThe power of a answer step
Weight coefficient, wjTo be answered the in set of steps in the examination questionThe weight coefficient of a answer step,To belong in the examination question answer step sum for answering set of steps outside target,Step is answered in the examination question to belong to
Suddenly the answer step sum gathered.
8. the as described in claim 1 a kind of automatic method for recommending item content, which is characterized in that the step S104 it
It afterwards, further include following steps：
S401. error vector is multiplied with preset weight row vector, completes the adjustment to the error vector.
9. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S105
In, the method for obtaining adaptation coefficients according to error vector includes the following steps：
S501. using two norms of error vector as adaptation coefficients.
10. a kind of automatic method for recommending item content as described in claim 1, which is characterized in that in the step S106
It before, further include following steps：
S600. it is directed to each examination question, has done examination question set according in user's current ability data, the user for counting the examination question inscribes
Number correct upwards shown in following formula to adaptation coefficients：
a′i=γtai
In formula, a 'iFor adaptation coefficients after the amendment of i-th (i=1,2,3 ..., N) a examination question, aiFor i-th (i=1,2,3 ..., N)
Adaptation coefficients before the amendment of a examination question, γ are upward correction factor, and t does topic number for the user in the examination question.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810726202.2A CN108920663B (en) | 2018-07-04 | 2018-07-04 | Method for automatically recommending title content |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810726202.2A CN108920663B (en) | 2018-07-04 | 2018-07-04 | Method for automatically recommending title content |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108920663A true CN108920663A (en) | 2018-11-30 |
| CN108920663B CN108920663B (en) | 2020-06-16 |
Family
ID=64425050
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810726202.2A Active CN108920663B (en) | 2018-07-04 | 2018-07-04 | Method for automatically recommending title content |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108920663B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109493666A (en) * | 2019-01-23 | 2019-03-19 | 广东小天才科技有限公司 | A kind of learning interaction method and facility for study |
| CN109785691A (en) * | 2019-01-18 | 2019-05-21 | 广东小天才科技有限公司 | A kind of method and system by terminal assisted learning |
| CN109783613A (en) * | 2019-01-23 | 2019-05-21 | 广东小天才科技有限公司 | One kind searching topic method and system |
| CN110175282A (en) * | 2019-04-18 | 2019-08-27 | 北京嗨学网教育科技股份有限公司 | The method and device of intelligent recommendation test question, storage medium |
| CN112084320A (en) * | 2019-06-14 | 2020-12-15 | 百度在线网络技术（北京）有限公司 | Test question recommendation method and device and intelligent equipment |
| CN116340657A (en) * | 2023-05-26 | 2023-06-27 | 深圳市菁优智慧教育股份有限公司 | Test question searching method and device based on user behaviors |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140188796A1 (en) * | 2012-12-28 | 2014-07-03 | Dropbox, Inc. | Application recommendation using stored files |
| CN105224665A (en) * | 2015-09-30 | 2016-01-06 | 广东小天才科技有限公司 | A kind of wrong topic management method and system |
| CN106126524A (en) * | 2016-06-07 | 2016-11-16 | 腾讯科技（深圳）有限公司 | Information-pushing method and device |
| CN106803123A (en) * | 2016-12-20 | 2017-06-06 | 广州中软信息技术有限公司 | A kind of automatic volume group method for online exam |
| CN107909520A (en) * | 2017-11-02 | 2018-04-13 | 浙江工商大学 | The method and apparatus that make the test based on examination question correlation |
-
2018
- 2018-07-04 CN CN201810726202.2A patent/CN108920663B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140188796A1 (en) * | 2012-12-28 | 2014-07-03 | Dropbox, Inc. | Application recommendation using stored files |
| CN105224665A (en) * | 2015-09-30 | 2016-01-06 | 广东小天才科技有限公司 | A kind of wrong topic management method and system |
| CN106126524A (en) * | 2016-06-07 | 2016-11-16 | 腾讯科技（深圳）有限公司 | Information-pushing method and device |
| CN106803123A (en) * | 2016-12-20 | 2017-06-06 | 广州中软信息技术有限公司 | A kind of automatic volume group method for online exam |
| CN107909520A (en) * | 2017-11-02 | 2018-04-13 | 浙江工商大学 | The method and apparatus that make the test based on examination question correlation |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109785691A (en) * | 2019-01-18 | 2019-05-21 | 广东小天才科技有限公司 | A kind of method and system by terminal assisted learning |
| CN109785691B (en) * | 2019-01-18 | 2021-09-24 | 广东小天才科技有限公司 | Method and system for assisting learning through terminal |
| CN109493666A (en) * | 2019-01-23 | 2019-03-19 | 广东小天才科技有限公司 | A kind of learning interaction method and facility for study |
| CN109783613A (en) * | 2019-01-23 | 2019-05-21 | 广东小天才科技有限公司 | One kind searching topic method and system |
| CN110175282A (en) * | 2019-04-18 | 2019-08-27 | 北京嗨学网教育科技股份有限公司 | The method and device of intelligent recommendation test question, storage medium |
| CN112084320A (en) * | 2019-06-14 | 2020-12-15 | 百度在线网络技术（北京）有限公司 | Test question recommendation method and device and intelligent equipment |
| CN112084320B (en) * | 2019-06-14 | 2023-09-15 | 百度在线网络技术（北京）有限公司 | Test question recommending method and device and intelligent equipment |
| CN116340657A (en) * | 2023-05-26 | 2023-06-27 | 深圳市菁优智慧教育股份有限公司 | Test question searching method and device based on user behaviors |
| CN116340657B (en) * | 2023-05-26 | 2023-08-29 | 深圳市菁优智慧教育股份有限公司 | Test question searching method and device based on user behaviors |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108920663B (en) | 2020-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108920663A (en) | A kind of method of automatic recommendation item content | |
| CN108875916B (en) | Advertisement click rate prediction method based on GRU neural network | |
| CN109829059A (en) | Recommend method, apparatus, equipment and the storage medium of knowledge point | |
| CN110032623B (en) | Method and device for matching question of user with title of knowledge point | |
| CN106844530A (en) | Training method and device of a kind of question and answer to disaggregated model | |
| CN107563500A (en) | A kind of video recommendation method and system based on user's head portrait | |
| CN112699283B (en) | Test paper generation method and device | |
| CN111723973B (en) | Learning effect optimization method based on user behavior causal relationship in MOOC log data | |
| CN106570080A (en) | Multilevel semantic matching method for cloud manufacturing resource services | |
| CN109299380A (en) | Exercise personalized recommendation method in online education platform based on multidimensional characteristic | |
| CN104199826A (en) | Heterogeneous media similarity calculation method and retrieval method based on correlation analysis | |
| CN112417127B (en) | Dialogue model training and dialogue generation methods, devices, equipment and media | |
| CN108563749B (en) | Online education system resource recommendation method based on multi-dimensional information and knowledge network | |
| CN108897879A (en) | A method of individualized teaching is realized by human-computer interaction | |
| Chen et al. | Identification of λ-fuzzy measures using sampling design and genetic algorithms | |
| CN106095812A (en) | Intelligent test paper generation method based on similarity measurement | |
| CN107203600A (en) | It is a kind of to utilize the evaluation method for portraying cause and effect dependence and sequential influencing mechanism enhancing answer quality-ordered | |
| CN110704510A (en) | User portrait combined question recommendation method and system | |
| CN109408726B (en) | Question answering person recommendation method in question and answer website | |
| CN109615009A (en) | A kind of learning Content recommended method and electronic equipment | |
| CN113239209A (en) | Knowledge graph personalized learning path recommendation method based on RankNet-transformer | |
| CN110019686A (en) | Determination method, apparatus, equipment and the storage medium of knowledge point master degree | |
| CN114429212A (en) | Intelligent learning knowledge ability tracking method, electronic device and storage medium | |
| CN109858024A (en) | A kind of source of houses term vector training method and device based on word2vec | |
| CN104008301B (en) | A kind of field concept hierarchical structure method for auto constructing |
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 | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 1 / F cb101-017, building 8, yard 1, Zhongguancun East Road, Haidian District, Beijing Patentee after: Beijing Sita Intelligent Technology Co.,Ltd. Address before: 1 / F cb101-017, building 8, yard 1, Zhongguancun East Road, Haidian District, Beijing Patentee before: BEIJING XIZI EDUCATION TECHNOLOGY Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |