CN116263782A - Intelligent winding method, system and storage medium based on question bank - Google Patents
Intelligent winding method, system and storage medium based on question bank Download PDFInfo
- Publication number
- CN116263782A CN116263782A CN202111502547.8A CN202111502547A CN116263782A CN 116263782 A CN116263782 A CN 116263782A CN 202111502547 A CN202111502547 A CN 202111502547A CN 116263782 A CN116263782 A CN 116263782A
- Authority
- CN
- China
- Prior art keywords
- question
- target
- questions
- test
- candidate
- 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
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000004804 winding Methods 0.000 title claims description 13
- 238000012360 testing method Methods 0.000 claims abstract description 214
- 230000006870 function Effects 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 4
- 230000035772 mutation Effects 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 description 8
- 238000004422 calculation algorithm Methods 0.000 description 5
- 230000002068 genetic effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000005192 partition Methods 0.000 description 3
- 238000012797 qualification Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 108700028369 Alleles Proteins 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000013100 final test Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/38—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/383—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/194—Calculation of difference between files
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Library & Information Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Databases & Information Systems (AREA)
- Physiology (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Genetics & Genomics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The invention relates to an intelligent scrolling method, a system and a storage medium based on a question bank, wherein the method comprises the following steps: adding parameter information for each test question; adding at least one knowledge point label for each test question; establishing learning sequence constraint among knowledge points; updating the difficulty scores of all the test questions; acquiring a question strategy; searching a plurality of preliminary test questions according to the target field, the target subjects and the target question types, and searching a plurality of candidate test questions according to the target knowledge point range; and completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type. Compared with the prior art, the method establishes the learning sequence constraint among the knowledge points, and can judge whether the learning sequence constraint exists between the non-fallen knowledge points and the fallen knowledge points when the knowledge points of the test question part fall into the range of the target knowledge points, thereby avoiding the occurrence of the test questions of the superclass on the test paper.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to an intelligent scrolling method, system and storage medium based on a question bank.
Background
With the continuous development of network technology, networks are rapidly penetrating into various aspects of social life, and various industries and fields are being transformed into automation and intellectualization so as to improve efficiency and liberate manpower. In the education industry, knowledge acquisition channels are numerous, such as schools, coaching institutions, the Internet, books and the like, people can learn according to needs, master new skills, and personal capacity is improved. However, the learning effect often needs an examination to verify, whether it is obligation education, professional class learning and functional training, the learning degree of the student on the knowledge point can be detected through the examination, and some qualification certificates often need to be obtained after the examination is qualified, so the examination is an important point of the education industry.
The traditional examination questions are often designed by teachers or specialists manually, manual examination papers are carried out according to chapters, knowledge points, questions and capabilities of examination objects to be examined, large manpower and material resources and time cost are required to be consumed, and related knowledge points can not be covered completely, so that a plurality of researchers try to automatically select examination papers from a question bank by using an information processing technology. However, there are still some disadvantages.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide an intelligent scrolling method, system and storage medium based on a question bank.
The aim of the invention can be achieved by the following technical scheme:
an intelligent winding method based on a question bank comprises the steps of constructing the question bank, obtaining a question strategy and winding;
the construction question bank specifically comprises the following steps:
adding parameter information for test questions in a question bank, wherein the parameter information comprises field information, subject information and question type information; adding at least one knowledge point label for each test question; establishing learning sequence constraint among knowledge points; updating the difficulty scores of all the test questions;
the question obtaining strategy specifically comprises the following steps:
acquiring the target field, the target subjects, the target knowledge point range, the target difficulty level, the target topic types and the target quantity of each topic type;
the group coil is specifically as follows:
searching a plurality of preliminary test questions in a question bank according to the target field, the target subjects and the target questions, and searching a plurality of candidate test questions in the preliminary test questions according to the target knowledge point range; and respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
Preferably, the searching of the plurality of candidate test questions in the preliminary test questions according to the target knowledge point range comprises the following specific steps:
respectively matching each preliminary test question with a target knowledge point range:
if all knowledge point labels of one preliminary test question fall into the range of the target knowledge point, the preliminary test question is used as a candidate test question, and matching of the next preliminary test question is carried out;
if all knowledge point labels of one preliminary test question do not fall into the range of the target knowledge point, matching the next preliminary test question directly;
if the partial knowledge point labels of one preliminary test question fall into the range of the target knowledge points, judging whether the knowledge point labels falling into the range of the target knowledge points and the knowledge point labels not falling into the range of the target knowledge points meet learning sequence constraint, if so, taking the preliminary test question as a candidate test question to match the next preliminary test question, and if not, directly matching the next preliminary test question.
Preferably, establishing a learning sequence constraint between knowledge points is specifically:
if a plurality of knowledge points with different learning orders are present, the knowledge points are divided into a cluster, and an order value is added to each knowledge point in the cluster, wherein the different order values represent different learning orders.
Preferably, updating the difficulty score of each test question specifically includes:
obtaining the answer accuracy after the test questions are imported into the question bank, increasing and decreasing the score according to the answer accuracy on the basis of the initial difficulty score, and obtaining the latest difficulty score, wherein the initial difficulty score is manually set when the test questions are imported into the question bank for the first time.
Preferably, completing the selection of the target question type comprises the following steps:
s1, generating a plurality of candidate solutions to form an initial population, and initializing a null global optimal solution; each candidate solution is a choice question of the objective title type, and comprises k examination questions, wherein k is greater than 0, k represents the target number of the objective title type, and any candidate solution is better than a global optimal solution initialized to be empty;
s2, calculating the fitness value of each candidate solution to obtain a current optimal solution and updating a global optimal solution;
s3, judging whether the population converges, if so, outputting a global optimal solution as a topic type topic, and if not, executing a step S4;
s4, performing selection operation, crossover operation and mutation operation to obtain a new population, and repeating the step S2.
Preferably, generating a candidate solution includes the steps of:
s11, acquiring a target difficulty level, determining the proportion of the questions with different difficulties based on the target difficulty level, and combining the target quantity of the topic types to obtain the set quantity of the questions with different difficulties in the topic types;
s12, finding out a title type from the candidate test questions, and classifying the difficulties of the candidate test questions according to the difficulty scores of the candidate test questions;
s13, selecting candidate test questions in each corresponding difficulty category according to the set number of the questions with different difficulties in the title type, and obtaining a candidate solution.
Preferably, in step S13, if the number of questions of a certain difficulty is q, that is, q candidate questions need to be selected in the corresponding difficulty classification, and q >1, selecting candidate questions in the corresponding difficulty classification includes the following steps:
s131, randomly selecting and storing a candidate test question;
s132, randomly selecting a candidate test question, calculating the similarity between the candidate test question and each stored candidate test question, executing the step S133 if the similarity is smaller than a preset similarity threshold, otherwise, repeating the step;
s133, storing the candidate test questions, if the number of the stored candidate test questions is equal to q, completing the selection of the classified candidate test questions, otherwise, repeating the step S132.
Preferably, the fitness value of each candidate solution is calculated using a fitness function whose input parameters include: scoring the difficulty of each examination question in the candidate solution, and using frequency and knowledge point coverage rate of each examination question in the candidate solution;
the using frequency is equal to N divided by N, N represents the number of times of being selected as the examination questions after the examination questions are introduced into the question bank, N represents the number of times of the associated group of the examination questions after the examination questions are introduced into the question bank, and if the question strategy of the two groups of the examination questions is the same, the two groups of the examination questions are associated with each other;
the coverage rate of the knowledge points is equal to M divided by M, M represents the number of the knowledge point labels of the examination questions falling into the range of the target knowledge points, and M represents the total number of the knowledge point labels of the examination questions.
An intelligent winding system based on a question bank, based on the intelligent winding method based on the question bank, comprises the following steps:
the database module is used for constructing a question bank, and specifically comprises the following steps: adding parameter information for test questions in a question bank, wherein the parameter information comprises field information, subject information and question type information; adding at least one knowledge point label for each test question; updating the difficulty scores of all the test questions;
the input/output module is used for acquiring a question strategy and outputting test paper, and the question strategy is specifically acquired by the following steps: acquiring the target field, the target subjects, the target knowledge point range, the target difficulty level, the target topic types and the target quantity of each topic type;
the group volume module is adjacent to the database module and is used for grouping volumes, and specifically comprises: searching a plurality of preliminary test questions in a question bank according to the target field, the target subjects and the target questions, and searching a plurality of candidate test questions in the preliminary test questions according to the target knowledge point range; and respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
A storage medium having stored thereon an executable computer program which when executed implements an intelligent method of question bank based coiling as described above.
Compared with the prior art, the invention has the following beneficial effects:
(1) The learning sequence constraint between the knowledge points is established, and when the knowledge points of the test question part fall into the range of the target knowledge points, whether the constraint on the learning sequence exists between the knowledge points which do not fall into the range of the target knowledge points or not can be judged, so that the test questions of the superclass on the test paper are avoided.
(2) For the selected questions of each target question type, a genetic algorithm is used, the most-selected questions of each target question type are obtained respectively, the most-selected questions of each target question type are integrated together, the final test paper is obtained, the distinction among the questions of different types is fully considered, the selected questions are carried out by the question types, the research granularity is smaller, and the method is more accurate.
(3) And obtaining the optimal questions of each objective title type by using a genetic algorithm, wherein the calculation of the fitness function considers the difficulty score, the use frequency and the knowledge point coverage rate of the questions, and the test paper meeting the question-setting strategy and having better quality is obtained.
Drawings
FIG. 1 is a flow diagram of an intelligent group roll method;
FIG. 2 is a schematic diagram of a test question storage structure in a question bank;
FIG. 3 is a schematic flow chart of a process for selecting questions using a genetic algorithm;
FIG. 4 is a schematic diagram of the architecture of an intelligent group roll system;
Detailed Description
The invention will now be described in detail with reference to the drawings and specific examples. The present embodiment is implemented on the premise of the technical scheme of the present invention, and a detailed implementation manner and a specific operation process are given, but the protection scope of the present invention is not limited to the following examples.
In the drawings, like structural elements are referred to by like reference numerals and components having similar structure or function are referred to by like reference numerals. The dimensions and thickness of each component shown in the drawings are arbitrarily shown, and the present invention is not limited to the dimensions and thickness of each component. Some of the elements in the drawings are exaggerated where appropriate for clarity of illustration.
Example 1:
an intelligent winding method based on a question bank is shown in figure 1 and comprises the steps of constructing the question bank, obtaining a question strategy and winding;
(1) The construction of the question bank is specifically as follows:
adding parameter information for test questions in a question bank, wherein the parameter information comprises field information, subject information and question type information, dividing the test questions in the question bank into a plurality of fields according to the field information, dividing the test questions in each field into a plurality of subjects according to the subject information, and dividing the test questions in each subject into a plurality of question types according to the question type information; adding at least one knowledge point label for each test question; establishing learning sequence constraint among knowledge points; updating the difficulty scores of all the test questions;
(2) The obtaining of the question strategy comprises the following steps:
acquiring the target field, the target subjects, the target knowledge point range, the target difficulty level, the target topic types and the target quantity of each topic type;
(3) The group coil is specifically as follows:
searching a plurality of preliminary test questions in a question bank according to the target field, the target subjects and the target questions, and searching a plurality of candidate test questions in the preliminary test questions according to the target knowledge point range; and respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
Each step is described below:
(1) Construction of question bank
(1) When the question bank is initially constructed, a plurality of questions can be imported in batches as the question bank, and the questions can be manually imported into the database through excel files and the like later, or the questions can be manually added one by one, or the questions can be modified, deformed and the like.
(2) In the database, as shown in fig. 2, the test questions can be stored in a partition mode according to fields, such as primary school, middle school, university, examination, certificate qualification and the like, each field is stored in a partition mode according to subjects, such as Chinese, mathematics, english, politics, history, geography, biology, physics, chemistry and the like of the middle school, such as the university is partitioned according to professional names, the university is partitioned into professional classes, public classes and the like under the professional directories, and the test questions are stored in a partition mode according to the types of the subjects, such as selection questions, blank filling questions, application questions and the like, and each subject can have own characteristic types, such as complete blank filling of English and the like. Considering that mathematics of primary school and mathematics of secondary school are greatly different, the field can be set to be a smaller field, such as first school, high school, etc. Therefore, the test questions in the question library are classified separately, and a plurality of test questions in the target field, the target subjects and the target question types can be found out quickly during subsequent paper assembly, so that the speed is higher.
(3) In the application, considering that some complicated questions can simultaneously examine a plurality of knowledge points, a single knowledge point label cannot completely represent the content of a test question, so that at least one knowledge point label is set for each test question, and theoretically, the more the knowledge point labels are, the more detailed the test question is in classification.
(4) Considering that the learning of the knowledge points has a sequence relation, a learning sequence constraint among the knowledge points is established, specifically: if a plurality of knowledge points with different learning orders are present, the knowledge points are divided into a cluster, and an order value is added to each knowledge point in the cluster, wherein the different order values represent different learning orders. If the multiplication distribution law can be learned only after the addition and multiplication are learned, a cluster is established, which comprises the addition, the multiplication and the multiplication distribution law, wherein the learning sequence of the multiplication distribution law is the latest. The ranking may be using arabic numerals, english letters, roman alphabet, etc.
Although the learning of some knowledge is not limited logically, there are often some artificial constraints, such as artificial rules that the learning of the Tang poetry is performed first and then the learning of the Song words, so that the constraint of the learning sequence between the Tang poetry and the Song words can be established. If the test questions stored in the question bank also record the chapter content of the matched teaching materials, learning sequence constraints can be generated according to the chapter sequence.
In theory, if the accuracy and the number of the knowledge point labels of each test question are enough, if the learning sequence constraint is established to be perfect enough, the phenomenon of superclass cannot occur after the target knowledge point range of the question is determined.
(5) The difficulty score of updating each test question is specifically:
obtaining the answer accuracy after the test questions are imported into the question bank, increasing and decreasing the score according to the answer accuracy on the basis of the initial difficulty score, and obtaining the latest difficulty score, wherein the initial difficulty score is manually set when the test questions are imported into the question bank for the first time. The percent system can be used, the ten system, the five-percent system and the like, and the higher the fraction, the more difficult the fraction.
When the question bank is imported for the first time, an initial difficulty score can be manually given, then the difficulty score can be updated periodically, for example, the update can be performed once every other week, the update can also be performed regularly, for example, the difficulty score update is performed once every 10 times of winding, or the difficulty score update is performed according to other rules.
The higher the answer accuracy of the test questions is, the lower the difficulty score of the test questions is, the more difficult the answer accuracy is, and the higher the difficulty score of the test questions is. Furthermore, if the answer level of the user can be obtained, the difficulty score can be updated in a weighted manner, if the answer level of the user is higher, a plurality of problems can be solved, when the answer of the user is wrong, the difficulty score can be increased by a larger weight, and if the answer level of the user is worse, when the answer of the user is correct, the difficulty score can be reduced by a larger weight.
(2) Obtaining a question strategy
The question strategy comprises a target field, a target subject, a target knowledge point range, a target difficulty level, a target question type and a target number of each subject title type,
the target field, the target subjects and the target subject are determined from a subject library, the target field and the target subjects are one, the target subjects are a plurality of, but if the comprehensive ability is required to be tested, the target fields and the target subjects can be selected, and similarly, the target subjects can be set as one for special exercises. The target knowledge point range can be set by referring to the knowledge point label provided by the question bank, and at least comprises one knowledge point, and further, if the test questions stored in the question bank also record the matched chapter content of the teaching materials, the chapter content and the like can be selected when the target knowledge point range is set. The total number of questions of the test paper and the number of the title types of each of the test paper are set according to the needs.
The target difficulty level may be a preset difficulty level, such as a level a difficulty, a level B difficulty, a level C difficulty, and the like. For example, under the level A difficulty, taking the difficulty rating of the percent as an example, the difficulty rating of 40% of test questions is required to be 80-100, the difficulty rating of 30% of test questions is required to be 40-80, the difficulty rating of 30% of test questions is required to be 0-40, and the like. The difficulty level can be manually input, and the quantity proportion of test questions with various difficulties in the test paper can be automatically set.
The difficulties of different types of questions are often different, if the application questions contain more knowledge points, the questions are generally difficult to select, and most of the questions selected in the conventional examination are test questions with lower difficulty, so in other embodiments, different difficulty ratios can be set according to different question types, for example, the ratio of the difficult, medium and easy test questions in the selection questions is 3:5:2, the ratio of difficult, medium and easy test questions in the application questions is 5:4:1, etc.
(3) Group rolls
(1) According to the target field, the target subjects and the target subject types, a plurality of preliminary test subjects are searched in the subject library, and because the test subjects in the subject library are sequentially stored according to the target field, the target subjects and the subject types in the embodiment, the plurality of preliminary test subjects can be quickly found, for example, the target field is primary school, the target subjects are mathematics, the target subject types comprise selection subjects, blank filling subjects and application subjects, and then the plurality of preliminary test subjects which belong to mathematic subjects and are the selection subjects, blank filling subjects and application subjects in the primary school field can be found;
(2) searching a plurality of candidate test questions in the preliminary test questions according to the range of the target knowledge points;
respectively matching each preliminary test question with a target knowledge point range:
a. if all knowledge point labels of one preliminary test question fall into the range of the target knowledge point, the preliminary test question is used as a candidate test question, and matching of the next preliminary test question is carried out; if the target knowledge point range is addition and subtraction and multiplication and division, the knowledge point label of one test question comprises addition and subtraction, and all the knowledge point labels fall into the target knowledge point range and can be used as candidate test questions;
b. if all knowledge point labels of one preliminary test question do not fall into the range of the target knowledge point, matching the next preliminary test question directly; if the target knowledge point range is the area of the triangle, the knowledge point label of one test question comprises multiplication and division, and all the knowledge points do not fall into the target knowledge point range, so that the knowledge point label cannot be used as a candidate test question;
c. if the partial knowledge point labels of one preliminary test question fall into the range of the target knowledge points, judging whether the knowledge point labels falling into the range of the target knowledge points and the knowledge point labels not falling into the range of the target knowledge points meet learning sequence constraint, if so, taking the preliminary test question as a candidate test question to match the next preliminary test question, and if not, directly matching the next preliminary test question.
If the target knowledge point range is square area calculation, the knowledge point label of a test question comprises square area calculation and circle area calculation, and a learning sequence constraint of learning circle area calculation is set manually, so that the test question is super-class for test objects without mastering a circle area calculation formula, and cannot be used as a preliminary test question, and if the target knowledge point range is replaced by circle area calculation, the test question can be used as a candidate test question.
(3) And respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
Wherein, genetic algorithm is used for completing the choice questions of a target question type, as shown in fig. 3, comprising the following steps:
s1, generating a plurality of candidate solutions to form an initial population, for example, an initial population with a scale of 50, and initializing a blank global optimal solution; each candidate solution is a choice question of the objective title type, and comprises k examination questions, wherein k is greater than 0, k represents the target number of the objective title type, and any candidate solution is better than a global optimal solution initialized to be empty;
wherein generating a candidate solution comprises the steps of:
s11, acquiring a target difficulty level, determining the proportion of the questions with different difficulties based on the target difficulty level, and combining the target quantity of the topic types to obtain the set quantity of the questions with different difficulties in the topic types; if the total number of questions of the test paper is 20, the target question type comprises a selection question, a gap filling question and an application question, the target number is 10, 5 and 5, the target difficulty level is A, the target difficulty level is distributed according to the difficulty proportion of the A level (40%, 30% and 30%), 4 questions with the difficulty score of 80-100 points are selected from the selection questions, 3 questions with the difficulty score of 40-80 points are selected from the 3 questions with the difficulty score of 0-40 points, and similarly, the gap filling question is provided with 2 questions with the difficulty score of 80-100 points, 2 questions with the difficulty score of 40-80 points and 1 question with the difficulty score of 0-40 points. Or 1 test question with the difficulty rating of 40-80 and 2 test questions with the difficulty rating of 0-40.
S12, finding out a title type from the candidate test questions, and classifying the difficulties of the candidate test questions according to the difficulty scores of the candidate test questions; and the difficulty scoring intervals of the A level are 0-40, 40-80 and 80-100, so that the problems are classified according to the three intervals, and if other target difficulty levels are selected, the problems are classified according to the score interval division of the target difficulty levels.
S13, selecting candidate test questions in each corresponding difficulty category according to the set number of the questions with different difficulties in the title type, and obtaining a candidate solution.
If the number of questions of a certain difficulty is q, that is q candidate questions need to be selected in the corresponding difficulty classification, and q is greater than 1, for example, 3 selection questions with the difficulty score of 0-40 points need to be selected, selecting candidate questions in the corresponding difficulty classification comprises the following steps:
s131, randomly selecting and storing a candidate test question;
s132, randomly selecting a candidate test question, calculating the similarity between the candidate test question and each stored candidate test question, executing the step S133 if the similarity is smaller than a preset similarity threshold, otherwise, repeating the step;
s133, storing the candidate test questions, if the number of the stored candidate test questions is equal to q, completing the selection of the classified candidate test questions, otherwise, repeating the step S132;
through the similarity judgment in the steps S131 to S133, the problem that similar questions appear in the same question type of the test paper can be avoided, and the quality of the test paper is improved. The similarity calculation may be performed according to conventional understanding, for example, using a trained neural network, or averaging all word embedments in the stem, and then calculating cosine similarity between two stem word embedments, and so on.
S2, calculating the fitness value of each candidate solution to obtain a current optimal solution and updating a global optimal solution;
calculating fitness values of each candidate solution using a fitness function, the input parameters of the fitness function comprising: scoring the difficulty of each examination question in the candidate solution, and using frequency and knowledge point coverage rate of each examination question in the candidate solution; the using frequency is equal to N divided by N, N represents the number of times of being selected as the examination questions after the examination questions are introduced into the question bank, N represents the number of times of being associated with the examination questions after the examination questions are introduced into the question bank, and if the question-issuing strategies of the two times of the examination questions are the same, the two times of the examination papers are associated with each other; the coverage rate of the knowledge points is equal to M divided by M, M represents the number of knowledge point labels of the examination questions falling into the range of the target knowledge points, and M represents the total number of knowledge point labels of the examination questions.
The fitness value is used for evaluating the merits of the solution, directly influences the convergence rate of the genetic algorithm, and needs to be as simple as possible in consideration of computing resources, in this embodiment, a specific function formula is not given, and relevant practitioners can design according to the needs, such as simply weighting and adding various factors.
In the method, the difficulty score, the use frequency and the knowledge point coverage rate are considered, the difficulty score can be used for indicating whether the candidate solution accords with the target difficulty level, the use frequency can indicate the use times of one test question, if the use frequency is higher, the test question is indicated to appear on the test paper once, if a set of brand-new test paper is required to appear, the test question can be punished, the test question is prevented from appearing on the test paper as the test question, but the use frequency is higher, the test question is indicated to be a high-frequency test point, and if the test paper is used for daily training, the probability of the test question appearing on the test paper can be increased.
The coverage rate of the knowledge points represents the correlation between the test question and the range of the target knowledge points, and the higher the coverage rate of the knowledge points is, the more relevant knowledge points of the test question are in the range of the target knowledge points, and the higher the coverage rate of the knowledge points is, the less the correlation is. If the knowledge point label of the test question includes T1, T2, T3 and T4, as long as T4 falls within the range of the target knowledge point, it is explained that only a small part of the test question may be used for examining the knowledge points within the range of the target knowledge point, and the test question with lower knowledge point coverage rate can be punished, and the knowledge point coverage rate of the test question in the obtained optimal solution should be higher.
S3, judging whether the population converges, if so, outputting a global optimal solution as a topic type topic, and if not, executing a step S4;
the population convergence condition may be that the iteration number reaches the maximum iteration number, for example, 500 times, or that the global optimal solution does not become better in the last kp iterations, kp may be 10 or 20, or that the fitness value of the global optimal solution reaches a preset fitness threshold, or the like, or the conditions may be integrated as the population convergence condition.
S4, performing selection operation, crossover operation and mutation operation to obtain a new population, and repeating the step S2.
Selecting, namely selecting better individuals in the population, reserving and sending the better individuals into a new population, and eliminating the rest individuals; the crossover operation, that is, two candidate solutions exchange one allele with each other, in this embodiment, the test questions with the exchange difficulty degree scored in one interval are obtained, and new individuals are sent into a new population; the mutation operation selects a new test question from the question bank to replace the test question in the candidate solution.
According to the method and the device, differences among different types of questions are considered, such as the number of knowledge points related to the questions, the difficulty of the questions and the like, the questions are selected according to the question types, the optimal question of each target question type is obtained, and then the optimal question is integrated together to obtain the test paper.
After the test paper is completed, the test paper can be delivered to teachers, specialists and the like for manual verification, and the quality of the test paper is further ensured.
The application also protects an intelligent winding system based on the question bank, and an intelligent winding method based on the question bank, as shown in fig. 4, comprises the following steps:
the database module is used for constructing a question bank, and specifically comprises the following steps: adding parameter information for test questions in a question bank, wherein the parameter information comprises field information, subject information and question type information; adding at least one knowledge point label for each test question; updating the difficulty scores of all the test questions;
the input/output module is used for acquiring a question strategy and outputting test paper, and the question strategy is specifically acquired by the following steps: acquiring the target field, the target subjects, the target knowledge point range, the target difficulty level, the target topic types and the target quantity of each topic type;
the group volume module is adjacent to the database module and is used for grouping volumes, and specifically comprises: searching a plurality of preliminary test questions in a question bank according to the target field, the target subjects and the target questions, and searching a plurality of candidate test questions in the preliminary test questions according to the target knowledge point range; and respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
A plurality of test paper templates can be preset, question types, question numbers and difficulty levels are defined on the test paper templates, and options such as whether to consider high-frequency questions or not can be defined.
The application also protects a storage medium, on which an executable computer program is stored, which when executed implements the intelligent method for composing volumes based on question banks.
The foregoing describes in detail preferred embodiments of the present invention. It should be understood that numerous modifications and variations can be made in accordance with the concepts of the invention by one of ordinary skill in the art without undue burden. Therefore, all technical solutions which can be obtained by logic analysis, reasoning or limited experiments based on the prior art by the person skilled in the art according to the inventive concept shall be within the scope of protection defined by the claims.
Claims (10)
1. The intelligent winding method based on the question bank is characterized by comprising the steps of constructing the question bank, obtaining a question strategy and winding;
the construction question bank specifically comprises the following steps:
adding parameter information for test questions in a question bank, wherein the parameter information comprises field information, subject information and question type information; adding at least one knowledge point label for each test question; establishing learning sequence constraint among knowledge points; updating the difficulty scores of all the test questions;
the question obtaining strategy specifically comprises the following steps:
acquiring the target field, the target subjects, the target knowledge point range, the target difficulty level, the target topic types and the target quantity of each topic type;
the group coil is specifically as follows:
searching a plurality of preliminary test questions in a question bank according to the target field, the target subjects and the target questions, and searching a plurality of candidate test questions in the preliminary test questions according to the target knowledge point range; and respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
2. The method for intelligent question bank-based paper assembly according to claim 1, wherein the searching of the plurality of candidate questions in the preliminary questions according to the target knowledge point range is specifically as follows:
respectively matching each preliminary test question with a target knowledge point range:
if all knowledge point labels of one preliminary test question fall into the range of the target knowledge point, the preliminary test question is used as a candidate test question, and matching of the next preliminary test question is carried out;
if all knowledge point labels of one preliminary test question do not fall into the range of the target knowledge point, matching the next preliminary test question directly;
if the partial knowledge point labels of one preliminary test question fall into the range of the target knowledge points, judging whether the knowledge point labels falling into the range of the target knowledge points and the knowledge point labels not falling into the range of the target knowledge points meet learning sequence constraint, if so, taking the preliminary test question as a candidate test question to match the next preliminary test question, and if not, directly matching the next preliminary test question.
3. The intelligent scrolling method based on the question bank according to claim 1, wherein the establishing of learning order constraints between knowledge points is specifically:
if a plurality of knowledge points with different learning orders are present, the knowledge points are divided into a cluster, and an order value is added to each knowledge point in the cluster, wherein the different order values represent different learning orders.
4. The intelligent grouping and scrolling method based on the question bank according to claim 1, wherein updating the difficulty score of each test question is specifically as follows:
obtaining the answer accuracy after the test questions are imported into the question bank, increasing and decreasing the score according to the answer accuracy on the basis of the initial difficulty score, and obtaining the latest difficulty score, wherein the initial difficulty score is manually set when the test questions are imported into the question bank for the first time.
5. The intelligent scrolling method of claim 1, wherein completing the selection of a target question type comprises the steps of:
s1, generating a plurality of candidate solutions to form an initial population, and initializing a null global optimal solution; each candidate solution is a choice question of the objective title type, and comprises k examination questions, wherein k is greater than 0, k represents the target number of the objective title type, and any candidate solution is better than a global optimal solution initialized to be empty;
s2, calculating the fitness value of each candidate solution to obtain a current optimal solution and updating a global optimal solution;
s3, judging whether the population converges, if so, outputting a global optimal solution as a topic type topic, and if not, executing a step S4;
s4, performing selection operation, crossover operation and mutation operation to obtain a new population, and repeating the step S2.
6. The intelligent question bank based scrolling method of claim 5, wherein generating a candidate solution comprises the steps of:
s11, acquiring a target difficulty level, determining the proportion of the questions with different difficulties based on the target difficulty level, and combining the target quantity of the topic types to obtain the set quantity of the questions with different difficulties in the topic types;
s12, finding out a title type from the candidate test questions, and classifying the difficulties of the candidate test questions according to the difficulty scores of the candidate test questions;
s13, selecting candidate test questions in each corresponding difficulty category according to the set number of the questions with different difficulties in the title type, and obtaining a candidate solution.
7. The method of claim 6, wherein in step S13, if the number of questions of a certain difficulty is q, i.e. q candidate questions need to be selected from the corresponding difficulty classifications, and q >1, then selecting candidate questions from the corresponding difficulty classifications comprises the following steps:
s131, randomly selecting and storing a candidate test question;
s132, randomly selecting a candidate test question, calculating the similarity between the candidate test question and each stored candidate test question, executing the step S133 if the similarity is smaller than a preset similarity threshold, otherwise, repeating the step;
s133, storing the candidate test questions, if the number of the stored candidate test questions is equal to q, completing the selection of the classified candidate test questions, otherwise, repeating the step S132.
8. The intelligent question bank based scrolling method of claim 5, wherein the fitness value of each candidate solution is calculated using a fitness function, the input parameters of the fitness function comprising: scoring the difficulty of each examination question in the candidate solution, and using frequency and knowledge point coverage rate of each examination question in the candidate solution;
the using frequency is equal to N divided by N, N represents the number of times of being selected as the examination questions after the examination questions are introduced into the question bank, N represents the number of times of the associated group of the examination questions after the examination questions are introduced into the question bank, and if the question strategy of the two groups of the examination questions is the same, the two groups of the examination questions are associated with each other;
the coverage rate of the knowledge points is equal to M divided by M, M represents the number of the knowledge point labels of the examination questions falling into the range of the target knowledge points, and M represents the total number of the knowledge point labels of the examination questions.
9. An intelligent question bank based scrolling system, characterized in that it is based on the intelligent question bank based scrolling method as claimed in any one of claims 1-8, comprising:
the database module is used for constructing a question bank, and specifically comprises the following steps: adding parameter information for test questions in a question bank, wherein the parameter information comprises field information, subject information and question type information; adding at least one knowledge point label for each test question; updating the difficulty scores of all the test questions;
the input/output module is used for acquiring a question strategy and outputting test paper, and the question strategy is specifically acquired by the following steps: acquiring the target field, the target subjects, the target knowledge point range, the target difficulty level, the target topic types and the target quantity of each topic type;
the group volume module is adjacent to the database module and is used for grouping volumes, and specifically comprises: searching a plurality of preliminary test questions in a question bank according to the target field, the target subjects and the target questions, and searching a plurality of candidate test questions in the preliminary test questions according to the target knowledge point range; and respectively completing the topic selection of each topic type based on the target difficulty level and the target number of each topic type to obtain a test paper.
10. A storage medium having stored thereon an executable computer program which when executed implements the subject library-based intelligent volume method of any of claims 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111502547.8A CN116263782A (en) | 2021-12-10 | 2021-12-10 | Intelligent winding method, system and storage medium based on question bank |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111502547.8A CN116263782A (en) | 2021-12-10 | 2021-12-10 | Intelligent winding method, system and storage medium based on question bank |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116263782A true CN116263782A (en) | 2023-06-16 |
Family
ID=86723238
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111502547.8A Pending CN116263782A (en) | 2021-12-10 | 2021-12-10 | Intelligent winding method, system and storage medium based on question bank |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116263782A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116542829A (en) * | 2023-07-03 | 2023-08-04 | 江西师范大学 | Personalized intelligent group scroll and on-line examination method |
CN117131104A (en) * | 2023-08-28 | 2023-11-28 | 河北望岳信息科技有限公司 | Intelligent question-drawing and winding method and device, electronic equipment and storage medium |
CN117216195A (en) * | 2023-11-08 | 2023-12-12 | 湖南强智科技发展有限公司 | Intelligent paper-making method, system, equipment and storage medium for course examination of universities |
CN117217209A (en) * | 2023-11-07 | 2023-12-12 | 湖南强智科技发展有限公司 | Intelligent college examination paper assembling method, system, equipment and storage medium |
CN117216081A (en) * | 2023-11-08 | 2023-12-12 | 联城科技(河北)股份有限公司 | Automatic question bank updating method and device, electronic equipment and storage medium |
CN117473973A (en) * | 2023-12-20 | 2024-01-30 | 河北金卷教育科技有限公司 | Intelligent test paper library management system and method |
CN117952797A (en) * | 2024-03-21 | 2024-04-30 | 深圳市华师兄弟教育科技有限公司 | Internet education supervision system and method based on artificial intelligence |
-
2021
- 2021-12-10 CN CN202111502547.8A patent/CN116263782A/en active Pending
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116542829A (en) * | 2023-07-03 | 2023-08-04 | 江西师范大学 | Personalized intelligent group scroll and on-line examination method |
CN117131104A (en) * | 2023-08-28 | 2023-11-28 | 河北望岳信息科技有限公司 | Intelligent question-drawing and winding method and device, electronic equipment and storage medium |
CN117131104B (en) * | 2023-08-28 | 2024-02-27 | 河北望岳信息科技有限公司 | Intelligent question-drawing and winding method and device, electronic equipment and storage medium |
CN117217209A (en) * | 2023-11-07 | 2023-12-12 | 湖南强智科技发展有限公司 | Intelligent college examination paper assembling method, system, equipment and storage medium |
CN117216195A (en) * | 2023-11-08 | 2023-12-12 | 湖南强智科技发展有限公司 | Intelligent paper-making method, system, equipment and storage medium for course examination of universities |
CN117216081A (en) * | 2023-11-08 | 2023-12-12 | 联城科技(河北)股份有限公司 | Automatic question bank updating method and device, electronic equipment and storage medium |
CN117216195B (en) * | 2023-11-08 | 2024-02-02 | 湖南强智科技发展有限公司 | Intelligent paper-making method, system, equipment and storage medium for course examination of universities |
CN117216081B (en) * | 2023-11-08 | 2024-02-06 | 联城科技(河北)股份有限公司 | Automatic question bank updating method and device, electronic equipment and storage medium |
CN117473973A (en) * | 2023-12-20 | 2024-01-30 | 河北金卷教育科技有限公司 | Intelligent test paper library management system and method |
CN117952797A (en) * | 2024-03-21 | 2024-04-30 | 深圳市华师兄弟教育科技有限公司 | Internet education supervision system and method based on artificial intelligence |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116263782A (en) | Intelligent winding method, system and storage medium based on question bank | |
CN107230174B (en) | Online interactive learning system and method based on network | |
Matzavela et al. | Decision tree learning through a predictive model for student academic performance in intelligent m-learning environments | |
CN111179675B (en) | Personalized exercise recommendation method and system, computer device and storage medium | |
CN107967254A (en) | Knowledge point prediction method and device, storage medium and electronic equipment | |
CN112784608B (en) | Test question recommending method and device, electronic equipment and storage medium | |
WO2021180249A1 (en) | Occupation recommendation method and apparatus, and device and medium | |
US20060099561A1 (en) | Automated assessment development and associated methods | |
CN112699283B (en) | Test paper generation method and device | |
CN111651677B (en) | Course content recommendation method, apparatus, computer device and storage medium | |
CN114201684A (en) | Knowledge graph-based adaptive learning resource recommendation method and system | |
CN111242816A (en) | Multimedia teaching plan making method and system based on artificial intelligence | |
CN114913729A (en) | Question selection method and device, computer equipment and storage medium | |
Hamim et al. | Student profile modeling using boosting algorithms | |
CN112419812A (en) | Exercise correction method and device | |
CN115438152B (en) | Simple answer scoring method and system based on multi-neural network and knowledge graph | |
Otero et al. | Finding informative code metrics under uncertainty for predicting the pass rate of online courses | |
CN112699933B (en) | Automatic identification method and system for processing capability of user teaching materials | |
CN112001536B (en) | High-precision discovery method for point defect minimum sample of mathematical ability of middle and primary schools based on machine learning | |
CN112015830B (en) | Question storage method suitable for adaptive learning | |
Prabaswara et al. | Classification of Secondary School Destination for Inclusive Students using Decision Tree Algorithm | |
Arafiyah et al. | Monitoring online learners’ performance based on learning progress prediction | |
Iwata et al. | Recommendation for English multiple-choice cloze questions based on expected test scores | |
CN116578695B (en) | Method and system for question group rolling based on DFS algorithm | |
Mustafa et al. | PROPER SAMPLE SIZES FOR ENGLISH LANGUAGE TESTING: A SIMPLE STATISTICAL ANALYSIS |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication |