CN110472044A - Knowledge point classification method, device, readable storage medium storing program for executing and the server of mathematical problem - Google Patents
Knowledge point classification method, device, readable storage medium storing program for executing and the server of mathematical problem Download PDFInfo
- Publication number
- CN110472044A CN110472044A CN201910623328.1A CN201910623328A CN110472044A CN 110472044 A CN110472044 A CN 110472044A CN 201910623328 A CN201910623328 A CN 201910623328A CN 110472044 A CN110472044 A CN 110472044A
- Authority
- CN
- China
- Prior art keywords
- knowledge point
- target topic
- classification
- preset
- grade
- 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
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
Abstract
The invention belongs to field of computer technology more particularly to knowledge point classification method, device, computer readable storage medium and the servers of a kind of mathematical problem.The method is during carrying out mechanized classification, the characteristics of for mathematical problem, by the way of multi-stage combination classification, the automatic classification of knowledge points at different levels is embedded into an organic inter-related treatment process, it makes full use of the knowledge point of upper level to classify when carrying out next stage knowledge point and classifying to export, such as, it makes full use of the 1st grade of knowledge point to classify when carrying out the 2nd grade of knowledge point and classifying to export, it makes full use of the 2nd grade of knowledge point to classify when carrying out 3rd level knowledge point and classifying to export, and so on, until completing the knowledge point classification of afterbody.It can use the information of upper level knowledge point not only in this way to improve the classification accuracy of next stage knowledge point, but also knowledge point classification results at different levels can have been exported simultaneously in a procedure, and further improved classification effectiveness.
Description
Technical field
The invention belongs to field of computer technology more particularly to a kind of knowledge point classification method of mathematical problem, device, calculating
Machine readable storage medium storing program for executing and server.
Background technique
Each company is all the mode for taking manual sort to the classification of the knowledge point of mathematical problem in exam pool currently on the market.So
Regardless of be the mathematical problem in primary school, junior middle school or senior high school period, exam pool being all millions of, related knowledge
The manual sort that points also reach several hundred or even thousands of, traditional is not only with high costs, but also inefficiency, accuracy rate are low, can not
Carry out large scale data classification.
Summary of the invention
In view of this, the embodiment of the invention provides the knowledge point classification methods of a kind of mathematical problem, device, computer-readable
Storage medium and server, the with high costs, efficiency in a manner of solving the existing knowledge point to mathematical problem and carry out manual sort
Low and low accuracy rate problem.
The first aspect of the embodiment of the present invention provides a kind of knowledge point classification method of mathematical problem, may include:
Target topic, the mathematics of the entitled pending knowledge point classification of target are obtained from preset mathematics exam pool
Topic;
The characteristic information of the target topic is extracted, and the characteristic information of the target topic is input to the preset 1st
It is handled in grade classifier, obtains the 1st grade of knowledge point classification output of the target topic;
The classification of preceding i grades of knowledge points of the characteristic information of the target topic and the target topic is input to default
I+1 grade classifier in handled, obtain the target topic i+1 grade knowledge point classification output, i >=1;
I is increased into a digit, and judges whether i is less than preset knowledge point classification series;
If i is less than knowledge point classification series, it is described by the characteristic information of the target topic and institute to return to execution
The preceding i grades of knowledge points classification for stating target topic is input to the step of being handled in preset i+1 grade classifier;
If i is equal to knowledge point classification series, it regard the preceding i grades of knowledge points classification output of the target topic as institute
State the knowledge point classification results of target topic.
The second aspect of the embodiment of the present invention provides a kind of knowledge point sorter of mathematical problem, may include:
Topic obtains module, and for obtaining target topic from preset mathematics exam pool, the target is entitled pending
The mathematical problem of knowledge point classification;
Characteristic information extracting module, for extracting the characteristic information of the target topic;
First processing module, for by the characteristic information of the target topic be input in preset 1st grade of classifier into
Row processing obtains the 1st grade of knowledge point classification output of the target topic;
Second processing module, for by preceding i grades of knowledge points of the characteristic information of the target topic and the target topic
Classification, which is input in preset i+1 grade classifier, to be handled, and the i+1 grade knowledge point point of the target topic is obtained
Class output, i >=1;
Counting module, for i to be increased a digit;
Series judgment module, for judging whether i is less than preset knowledge point classification series;
As a result determining module knows preceding i grades of the target topic if being equal to knowledge point classification series for i
Know knowledge point classification results of the point classification output as the target topic.
The third aspect of the embodiment of the present invention provides a kind of computer readable storage medium, the computer-readable storage
Media storage has computer-readable instruction, and the computer-readable instruction realizes following steps when being executed by processor:
Target topic, the mathematics of the entitled pending knowledge point classification of target are obtained from preset mathematics exam pool
Topic;
The characteristic information of the target topic is extracted, and the characteristic information of the target topic is input to the preset 1st
It is handled in grade classifier, obtains the 1st grade of knowledge point classification output of the target topic;
The classification of preceding i grades of knowledge points of the characteristic information of the target topic and the target topic is input to default
I+1 grade classifier in handled, obtain the target topic i+1 grade knowledge point classification output, i >=1;
I is increased into a digit, and judges whether i is less than preset knowledge point classification series;
If i is less than knowledge point classification series, it is described by the characteristic information of the target topic and institute to return to execution
The preceding i grades of knowledge points classification for stating target topic is input to the step of being handled in preset i+1 grade classifier;
If i is equal to knowledge point classification series, it regard the preceding i grades of knowledge points classification output of the target topic as institute
State the knowledge point classification results of target topic.
The fourth aspect of the embodiment of the present invention provides a kind of server, including memory, processor and is stored in institute
The computer-readable instruction that can be run in memory and on the processor is stated, the processor executes described computer-readable
Following steps are realized when instruction:
Target topic, the mathematics of the entitled pending knowledge point classification of target are obtained from preset mathematics exam pool
Topic;
The characteristic information of the target topic is extracted, and the characteristic information of the target topic is input to the preset 1st
It is handled in grade classifier, obtains the 1st grade of knowledge point classification output of the target topic;
The classification of preceding i grades of knowledge points of the characteristic information of the target topic and the target topic is input to default
I+1 grade classifier in handled, obtain the target topic i+1 grade knowledge point classification output, i >=1;
I is increased into a digit, and judges whether i is less than preset knowledge point classification series;
If i is less than knowledge point classification series, it is described by the characteristic information of the target topic and institute to return to execution
The preceding i grades of knowledge points classification for stating target topic is input to the step of being handled in preset i+1 grade classifier;
If i is equal to knowledge point classification series, it regard the preceding i grades of knowledge points classification output of the target topic as institute
State the knowledge point classification results of target topic.
Existing beneficial effect is the embodiment of the present invention compared with prior art: the embodiment of the present invention is known using automation
Know point mode classification instead of traditional manual sort's mode, greatly reduce human cost, improve classification effectiveness, and into
During row mechanized classification, the characteristics of for mathematical problem, by the way of the multi-stage combination classification, by knowledge points at different levels from
Dynamic classification is embedded into an organic inter-related treatment process, make full use of when next stage knowledge point is classified
The knowledge point of level-one, which is classified, to be exported, for example, making full use of the 1st grade of knowledge point classification defeated when carrying out the 2nd grade of knowledge point and classifying
Out, it makes full use of the 2nd grade of knowledge point to classify when carrying out 3rd level knowledge point and classifying to export ..., and so on, until completing
Until the knowledge point classification of afterbody.The information of upper level knowledge point was can use both in this way to improve next stage knowledge point
Classification accuracy, and knowledge point classification results at different levels can be exported simultaneously in a procedure, further improve classification effect
Rate.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is only of the invention some
Embodiment for those of ordinary skill in the art without any creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is a kind of one embodiment flow chart of the knowledge point classification method of mathematical problem in the embodiment of the present invention;
Fig. 2 is the schematic flow diagram for extracting the characteristic information of target topic;
Fig. 3 is the schematic flow diagram that knowledge point classification is carried out to target topic;
Fig. 4 is a kind of one embodiment structure chart of the knowledge point sorter of mathematical problem in the embodiment of the present invention;
Fig. 5 is a kind of schematic block diagram of server in the embodiment of the present invention.
Specific embodiment
In order to make the invention's purpose, features and advantages of the invention more obvious and easy to understand, below in conjunction with the present invention
Attached drawing in embodiment, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that disclosed below
Embodiment be only a part of the embodiment of the present invention, and not all embodiment.Based on the embodiments of the present invention, this field
Those of ordinary skill's all other embodiment obtained without making creative work, belongs to protection of the present invention
Range.
Referring to Fig. 1, a kind of one embodiment of the knowledge point classification method of mathematical problem can wrap in the embodiment of the present invention
It includes:
Step S101, target topic is obtained from preset mathematics exam pool.
The mathematical problem of the entitled pending knowledge point classification of target.
In embodiments of the present invention, a variety of classification modes can be taken to carry out knowledge point classification, in different classification moulds
The mode for obtaining target topic under formula from mathematics exam pool is not also identical.
The server for executing knowledge point classification reads current pattern identification first, and according to the pattern identification determine into
Row knowledge point classification classification mode, the embodiment of the present invention one kind in the specific implementation, when pattern identification value be 1 when, then
Classification mode is preset first mode, and in such a mode, server is based on tensorflow serving and docker
Deployment services support classification in real time, and terminal device inputs topic mark, and server can take out the topic from mathematics exam pool,
And tensorflow serving service is called to carry out real-time grading to it.When pattern identification value is 2, then classification mode is
Preset second mode, when pattern identification value is 3, then classification mode is preset the third mode, under both modes,
Server is that python project is directly disposed on spark platform, and server can be inquired in mathematics exam pool in a triggered is
It is no to have new non-classified topic, if so, then to the progress batch classification of these topics, but unlike, under the second mode, clothes
Business device needs terminal device to carry out external trigger, and in a third mode, server can pass through timing automatic trigger.
Specifically, if the classification mode is preset first mode, user can be by terminal device to server
The first sort instructions are sent, first sort instructions are to carry out knowledge point classification to the specified mathematical problem in the mathematics exam pool
Instruction, the topic mark of the mathematical problem of topic mark namely the classification of pending knowledge point is carry in the instruction.Clothes
After business device receives the first sort instructions, the topic mark in first sort instructions is extracted, and from the mathematics exam pool
It is middle to choose mathematical problem corresponding with topic mark as the target topic.It should be noted that the topic of different mathematical problems
Target knowledge is different.
If the classification mode is preset second mode, user can send first to server by terminal device
Sort instructions, second sort instructions are to carry out knowledge point classification to all non-classified mathematical problems in the mathematics exam pool
Instruction.After server receives the second sort instructions, the state of each mathematical problem in the mathematics exam pool is obtained respectively
Mark, and choose each mathematical problem that status indicator is preset first value from the mathematics exam pool and inscribed as the target
Mesh.Wherein, only there are two types of possible values for the status indicator of each mathematical problem, wherein the first value (can be set to 0) represents
Non-classified state, the second value (can be set to 1) represent classified state.
If the classification mode is preset the third mode, server obtains current system time first, when described
When system time is preset triggering moment, the status indicator of each mathematical problem in the mathematics exam pool is obtained respectively, and from
It is each mathematical problem of first value as the target topic that status indicator is chosen in the mathematics exam pool.The triggering
Moment can be configured according to the actual situation, for example, equally spaced a triggering moment can be set every 1 hour, this
Sample will carry out classification processing to the non-classified topic increased newly in the mathematics exam pool automatically every 1 hour server.
Step S102, the characteristic information of the target topic is extracted.
Mathematical problem is a kind of special text different from plain text, has both included Chinese or English word in mathematical problem,
Some special mathematic signs are contained, for this feature of mathematical problem, the embodiment of the present invention is to general in the prior art
It is improved for the mode that word is analyzed, proposes and extract characteristic information in terms of the two from word and symbol respectively
New processing mode, detailed process is as shown in Figure 2:
Step S1021, the target topic is split as set of words and assemble of symbol.
It include each word for constituting the target topic in the set of words, the assemble of symbol summarizes including constituting
Each mathematic sign of the target topic.
In one kind of the invention implemented in the specific implementation, fractionation and subsequent processing only can be carried out to the stem of topic,
In the another kind implemented of the present invention in the specific implementation, can the contents such as stem, parsing and discussion to topic carry out splitting and
Subsequent processing, to obtain information more abundant, to improve the accuracy rate of final result.
Step S1022, the word of each word in the set of words is searched respectively in preset term vector database
Vector, and the term vector of each word is configured to the word matrix of the target topic.
The term vector database is the database for recording the corresponding relationship between word and term vector.The term vector can
To be the corresponding term vector according to obtained by word2vec model training word.This is indicated according to the contextual information of word
The probability that word occurs.Each vocabulary is first shown as a 0-1 vector still according to the thought of word2vec by the training of term vector
(one-hot) form, then word2vec model training is carried out with term vector, n-th of word, neural network are predicted with n-1 word
The pilot process obtained after model prediction is as term vector.Specifically, as " celebrations " one-hot vector assume be set to [1,0,
0,0 ... ..., 0], the one-hot vector of " conference " is [0,1,0,0 ... ..., 0], the one-hot vector of " smooth " for [0,0,
1,0 ... ..., 0], predict that the vector [0,0,0,1 ... ..., 0] of " closing ", model can generate the coefficient square of hidden layer by training
Battle array W, the product of the one-hot vector sum coefficient matrix of each word are the term vector of the word, and last form will be analogous to " celebrating
Wish [- 0.28,0.34, -0.02 ... ..., 0.92] " such a multi-C vector.
It, can be by the term vector structure of each word after the term vector for finding each word in the set of words
It builds as the word matrix of the target topic, wherein the corresponding term vector of every a line of the word matrix, i.e., first
The term vector of word forms the first row of the word matrix, and the term vector of second word forms the second of the word matrix
Row ... ..., the term vector of the WN word form the WN row of the word matrix, and WN is that the word in the set of words is total
Number.
Step S1023, each symbol in the assemble of symbol is searched respectively in preset symbolic vector database
Symbolic vector, and the symbolic vector of each symbol is configured to the sign matrix of the target topic.
The database of corresponding relationship of the symbolic vector database between record symbol and symbolic vector.Due to symbol
Limited amount, and there is no semantic dependency between each symbol, therefore the form of one-hot can be directly used as each
The symbolic vector of a symbol.
For example, if a total of 10 characters: a, b, c, d, e, f, g, h, i, j.
The then symbolic vector of character a are as follows: [1,0,0,0,0,0,0,0,0,0];
The symbolic vector of character b are as follows: [0,1,0,0,0,0,0,0,0,0];
The symbolic vector of character c are as follows: [0,0,1,0,0,0,0,0,0,0];
……
And so on, the symbolic vector of character j are as follows: [0,0,0,0,0,0,0,0,0,1].
After the symbolic vector for finding each symbol in the assemble of symbol, can by the symbol of each symbol to
Amount is configured to the sign matrix of the target topic, wherein every a line of the sign matrix corresponds to a symbolic vector, i.e.,
The symbolic vector of first symbol forms the first row of the sign matrix, and the symbolic vector of second symbol forms the symbol
Second row ... ... of matrix, the symbolic vector of the SN symbol form the SN row of the sign matrix, and SN is the glossary of symbols
Total number of symbols in conjunction.
Step S1024, using the word matrix and the sign matrix as the characteristic information of the target topic.
By process shown in Fig. 2, the feature of the target topic is extracted in terms of the two from word and symbol respectively
Information has sufficiently excavated information included in target topic, and the accuracy rate of final result can be greatly improved.
Step S103, the characteristic information of the target topic is input in preset 1st grade of classifier and is handled, obtained
Classify to the 1st grade of knowledge point of the target topic and exports.
The knowledge point classification of mathematical problem can be divided into CN in advance by embodiments of the present invention, according to mathematical problem the characteristics of
A classification series, wherein next stage knowledge point classification is the subdivision of upper level knowledge point classification.Such as the 1st grade of knowledge point classification
" property of figure ", the 2nd grade of knowledge point classification can be subdivided into " figure understanding is preliminary " and " construction with ruler and compasses ", and 3rd level is known
Knowing point classification can be subdivided into more again, and so on.
Below by taking the classification of three-level knowledge point as an example, in entire knowledge point classification system, the 1st grade of knowledge point classification, the 2nd grade
There is relation of interdependence between knowledge point classification and 3rd level knowledge point classification, the 2nd grade of knowledge point classification is the 1st grade of knowledge
The subclass of point classification, 3rd level knowledge point classification is the subclass of the 2nd grade of knowledge point classification, as shown in the table:
In the present embodiment, the classifier of CN rank is pre-set, wherein the 1st grade of classifier is used for mathematical problem
1st grade of knowledge point classification is identified that the 2nd grade of classifier is for identifying the 2nd grade of knowledge point classification of mathematical problem, with this
Analogize.
It is illustrated by taking the 1st grade of classifier as an example below, altogether includes preset three points in the 1st grade of classifier
Analyse model, respectively the first analysis model, the second analysis model and third analysis model, wherein first analysis model is
Neural network model for classifying to word, second analysis model are the nerve net for classifying to symbol
Network model, it is preferable that the two analysis models all use CNN to realize, this is because there is no too many in mathematical problem
Semantic information is also not present the long-distance dependence problem of information, is more key word information.And CNN in the text can be fine
Extraction keyword information, furthermore CNN can parallel processing, training speed is better than RNN.The third analysis model
Integrated treatment is carried out for the output to the first analysis model model and second analysis model, to obtain knowledge point
Classification output.
When the characteristic information of the target topic, which is input to, to be handled in the 1st grade of classifier, first using described the
One analysis model handles the word matrix in the characteristic information, obtains the first output vector, then using described the
Two analysis models handle the sign matrix in the characteristic information, obtain the second output vector, specific processing therein
Process is identical as the treatment process of CNN in the prior art, and for details, reference can be made to related contents in the prior art, no longer superfluous herein
It states.
It reuses the third analysis model to handle first output vector and second output vector, obtain
Classify to the 1st grade of knowledge point of the target topic and exports.
Specifically, the detailed process of the third analysis model may include:
Firstly, by first output vector and second output vector merge the mixing being as follows export to
Amount:
CombOutVec=(CbOutVal1,CbOutVal2,...,CbOutValod,,...,CbOutValODN)
Wherein, od is the serial number of each 1st grade of knowledge point classification, and 1≤od≤ODN, ODN are the 1st grade of knowledge point classification
Sum, CbOutValodFor the value of the mixing output vector in the od dimension, and CbOutValod=WdOutValod+
SbOutValod, WdOutValodFor value of first output vector in the od dimension, SbOutValodIt is described
Value of two output vectors in the od dimension, CombOutVec are the mixing output vector.
Then, the probability of the entitled each 1st grade of knowledge point classification of the target is calculated separately according to the following formula:
Wherein, Exp is natural exponential function, ProbodFor the general of entitled od the 1st grade of knowledge point classifications of the target
Rate.
Finally, the 1st grade of knowledge point classification for choosing maximum probability is defeated as the 1st grade of knowledge point classification of the target topic
Out.
Step S104, preceding i grades of knowledge points of the characteristic information of the target topic and the target topic are classified and is exported
It is input in preset i+1 grade classifier and is handled, obtain the i+1 grade knowledge point classification output of the target topic.
Wherein, i >=1.In the initial state, i=1 is set, i.e., is by the spy of the target topic in first time circulation
Preceding 1 grade of knowledge point classification of reference breath and the target topic, which is input in preset 2nd grade of classifier, to be handled, and is obtained
Classify to the 2nd grade of knowledge point of the target topic and export, and is to believe the feature of the target topic in second of circulation
The classification of first 2 grades (including the 1st grade and the 2nd grade) knowledge points of breath and the target topic is input to preset 3rd level classification
It is handled in device, obtains the 3rd level knowledge point classification output ... ... of the target topic, and so on, as shown in figure 3, its
In, the concrete processing procedure of each classifier is similar with the above-mentioned 1st grade for the treatment of process of classifier, and details are not described herein again.It needs
It should be noted that Fig. 3 is by taking the classification of three-level knowledge point as an example, the case where other series, is similar therewith.
Step S105, i is increased into a digit.
It executes: i=i+1.
Step S106, judge whether i is less than preset knowledge point classification series.
If i is less than knowledge point classification series (i.e. CN), S104 and its subsequent step are returned to step;If i etc.
In knowledge point classification series, S107 is thened follow the steps.
Step S107, the knowledge point by the preceding i grades of knowledge points classification output of the target topic as the target topic
Classification results.
It after model structure shown in Fig. 3 is set up, needs first to be trained it, in order to guarantee to train obtained mould
Type has wider applicability, needs to pre-establish the sample database of magnanimity, includes each year in the sample database
As training sample, these training samples can be by for mathematical problem in the associated materials such as textbook, workbook, reference of grade
It is divided into training set and verifying collection according to certain ratio, for example, training set accounts for 80%, verifying collection accounts for 20%, naturally it is also possible to root
According to actual conditions, other rations of division are set.Wherein, training set is used to training pattern, and verifying collection is used to verify model.When testing
When accuracy rate on card collection converges to a stationary value, then it is believed that model has trained completion, the trained mould can be used
Type carries out the knowledge point classification of mathematical problem.
In conclusion the embodiment of the present invention is using the knowledge point mode classification of automation instead of traditional manual sort side
Formula greatly reduces human cost, classification effectiveness is improved, and during carrying out mechanized classification, for mathematical problem
Feature, by the way of multi-stage combination classification, by the automatic classification of knowledge points at different levels be embedded into one it is organic inter-related
In treatment process, makes full use of the knowledge point of upper level to classify when carrying out next stage knowledge point and classifying and export, for example, carrying out
It makes full use of the 1st grade of knowledge point to classify when the 2nd grade of knowledge point classification to export, it is sufficiently sharp when carrying out 3rd level knowledge point and classifying
Classified with the 2nd grade of knowledge point and exported ..., and so on, until completing the knowledge point classification of afterbody.Both may be used in this way
The classification accuracy of next stage knowledge point is improved with the information using upper level knowledge point, and can in a procedure simultaneously
Knowledge point classification results at different levels are exported, classification effectiveness is further improved.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Corresponding to a kind of knowledge point classification method of mathematical problem described in foregoing embodiments, Fig. 4 shows implementation of the present invention
A kind of one embodiment structure chart of the knowledge point sorter for mathematical problem that example provides.
In the present embodiment, a kind of knowledge point sorter of mathematical problem may include:
Topic obtain module 401, for from preset mathematics exam pool obtain target topic, the target it is entitled into
The mathematical problem of row knowledge point classification;
Characteristic information extracting module 402, for extracting the characteristic information of the target topic;
First processing module 403, for the characteristic information of the target topic to be input in preset 1st grade of classifier
It is handled, obtains the 1st grade of knowledge point classification output of the target topic;
Second processing module 404, for by preceding i grades of knowledge of the characteristic information of the target topic and the target topic
Point classification, which is input in preset i+1 grade classifier, to be handled, and the i+1 grade knowledge point of the target topic is obtained
Classification output, i >=1;
Counting module 405, for i to be increased a digit;
Series judgment module 406, for judging whether i is less than preset knowledge point classification series;
As a result determining module 407, if being equal to knowledge point classification series for i, by preceding i grades of the target topic
Knowledge point classification results of the knowledge point classification output as the target topic.
Further, the characteristic information extracting module may include:
Topic split cells, for the target topic to be split as set of words and assemble of symbol, the set of words
In include each word for constituting the target topic, the assemble of symbol summarizes each number including constituting the target topic
Learn symbol;
Word matrix construction unit, it is each in the set of words for being searched respectively in preset term vector database
The term vector of a word, and the term vector of each word is configured to the word matrix of the target topic, the term vector number
It is the database for recording the corresponding relationship between word and term vector according to library;
Sign matrix construction unit, for being searched in the assemble of symbol respectively in preset symbolic vector database
The symbolic vector of each symbol, and the symbolic vector of each symbol is configured to the sign matrix of the target topic, the symbol
The database of number corresponding relationship of the vector data library between record symbol and symbolic vector;
Characteristic information construction unit, for using the word matrix and the sign matrix as the spy of the target topic
Reference breath.
Further, the first processing module may include:
First processing units, for using preset first analysis model to carry out the word matrix in the characteristic information
Processing, obtains the first output vector;
The second processing unit, for using preset second analysis model to carry out the sign matrix in the characteristic information
Processing, obtains the second output vector;
Third processing unit, for using preset third analysis model to first output vector and described second defeated
Outgoing vector is handled, and the 1st grade of knowledge point classification output of the target topic is obtained.
Further, the third processing unit may include:
Vector merges subelement, for first output vector and second output vector merging to be as follows
Mixing output vector:
CombOutVec=(CbOutVal1,CbOutVal2,...,CbOutValod,,...,CbOutValODN)
Wherein, od is the serial number of each 1st grade of knowledge point classification, and 1≤od≤ODN, ODN are the 1st grade of knowledge point classification
Sum, CbOutValodFor the value of the mixing output vector in the od dimension, and CbOutValod=WdOutValod+
SbOutValod, WdOutValodFor value of first output vector in the od dimension, SbOutValodIt is described
Value of two output vectors in the od dimension, CombOutVec are the mixing output vector;
Probability calculation subelement, for calculating separately the entitled each 1st grade of knowledge point classification of the target according to the following formula
Probability:
Wherein, Exp is natural exponential function, ProbodFor the general of entitled od the 1st grade of knowledge point classifications of the target
Rate;
Classification output subelement, for choosing the 1st grade of knowledge point classification of maximum probability the as the target topic the 1st
The classification output of grade knowledge point.
Further, the topic acquisition module may include:
Classification mode determination unit is known for reading current pattern identification, and according to pattern identification determination
Know the classification mode of point classification;
First mode processing unit, if being preset first mode, receiving terminal apparatus hair for the classification mode
The first sort instructions sent, first sort instructions are to carry out knowledge point classification to the specified mathematical problem in the mathematics exam pool
Instruction;The topic mark in first sort instructions is extracted, and chooses from the mathematics exam pool and is identified with the topic
Corresponding mathematical problem is as the target topic;
Second mode processing unit, if being preset second mode, receiving terminal apparatus hair for the classification mode
The second sort instructions sent, second sort instructions are to know all non-classified mathematical problems in the mathematics exam pool
Know the instruction of point classification;Obtain the status indicator of each mathematical problem in the mathematics exam pool respectively, and from the mathematics exam pool
Middle selection status indicator is each mathematical problem of preset first value as the target topic, and first value represents not
The state of classification;
The third mode processing unit obtains current system if being preset the third mode for the classification mode
Moment obtains the shape of each mathematical problem in the mathematics exam pool when the system time is preset triggering moment respectively
State mark, and choose each mathematical problem that status indicator is first value from the mathematics exam pool and inscribed as the target
Mesh.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description,
The specific work process of module and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, is not described in detail or remembers in some embodiment
The part of load may refer to the associated description of other embodiments.
The schematic block diagram that Fig. 5 shows a kind of server provided in an embodiment of the present invention illustrates only for ease of description
Part related to the embodiment of the present invention.
In the present embodiment, the server 5 may include: processor 50, memory 51 and be stored in the storage
In device 51 and the computer-readable instruction 52 that can run on the processor 50, such as execute the knowledge point of above-mentioned mathematical problem
The computer-readable instruction of classification method.The processor 50 realizes above-mentioned each number when executing the computer-readable instruction 52
Step in the knowledge point classification method embodiment of topic, such as step S101 to S107 shown in FIG. 1.Alternatively, the processing
Device 50 realizes the function of each module/unit in above-mentioned each Installation practice, such as Fig. 4 when executing the computer-readable instruction 52
The function of shown module 401 to 407.
Illustratively, the computer-readable instruction 52 can be divided into one or more module/units, one
Or multiple module/units are stored in the memory 51, and are executed by the processor 50, to complete the present invention.Institute
Stating one or more module/units can be the series of computation machine readable instruction section that can complete specific function, the instruction segment
For describing implementation procedure of the computer-readable instruction 52 in the server 5.
The processor 50 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), field programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processor
Deng.
The memory 51 can be the internal storage unit of the server 5, such as the hard disk or memory of server 5.
The memory 51 is also possible to the External memory equipment of the server 5, such as the plug-in type being equipped on the server 5 is hard
Disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card
(Flash Card) etc..Further, the memory 51 can also both include the internal storage unit of the server 5 or wrap
Include External memory equipment.The memory 51 is for storing needed for the computer-readable instruction and the server 5 it
Its instruction and data.The memory 51 can be also used for temporarily storing the data that has exported or will export.
The functional units in various embodiments of the present invention may be integrated into one processing unit, is also possible to each
Unit physically exists alone, and can also be integrated in one unit with two or more units.Above-mentioned integrated unit both may be used
To use formal implementation of hardware, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, technical solution of the present invention substantially or
Person says that all or part of the part that contributes to existing technology or the technical solution can body in the form of software products
Reveal and, which is stored in a storage medium, including several computer-readable instructions are used so that one
Platform computer equipment (can be personal computer, server or the network equipment etc.) executes described in each embodiment of the present invention
The all or part of the steps of method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-
Only Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can be with
Store the medium of computer-readable instruction.
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations;Although referring to aforementioned reality
Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each
Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features;And these are modified
Or replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of knowledge point classification method of mathematical problem characterized by comprising
Target topic, the mathematical problem of the entitled pending knowledge point classification of target are obtained from preset mathematics exam pool;
The characteristic information of the target topic is extracted, and the characteristic information of the target topic is input to preset 1st fraction
It is handled in class device, obtains the 1st grade of knowledge point classification output of the target topic;
The classification of preceding i grades of knowledge points of the characteristic information of the target topic and the target topic is input to preset the
It is handled in i+1 grades of classifiers, obtains the i+1 grade knowledge point classification output of the target topic, i >=1;
I is increased into a digit, and judges whether i is less than preset knowledge point classification series;
If i is less than knowledge point classification series, returns and execute the characteristic information and the mesh by the target topic
I grades of knowledge point classification are input to the step of being handled in preset i+1 grade classifier before title purpose;
If i is equal to knowledge point classification series, it regard the preceding i grades of knowledge points classification output of the target topic as the mesh
Title purpose knowledge point classification results.
2. the knowledge point classification method of mathematical problem according to claim 1, which is characterized in that described to extract the target topic
Purpose characteristic information includes:
The target topic is split as set of words and assemble of symbol, includes constituting the target topic in the set of words
Each word, the assemble of symbol summarizes each mathematic sign including constituting the target topic;
Search the term vector of each word in the set of words respectively in preset term vector database, and by each word
The term vector of language is configured to the word matrix of the target topic, and the term vector database is between record word and term vector
Corresponding relationship database;
Search the symbolic vector of each symbol in the assemble of symbol respectively in preset symbolic vector database, and will be each
The symbolic vector of a symbol is configured to the sign matrix of the target topic, and the symbolic vector database is record symbol and symbol
The database of corresponding relationship between number vector;
Using the word matrix and the sign matrix as the characteristic information of the target topic.
3. the knowledge point classification method of mathematical problem according to claim 2, which is characterized in that described by the target topic
Characteristic information be input in preset 1st grade of classifier and handled, obtain the 1st grade of knowledge point classification of the target topic
Output includes:
The word matrix in the characteristic information is handled using preset first analysis model, obtain the first output to
Amount;
The sign matrix in the characteristic information is handled using preset second analysis model, obtain the second output to
Amount;
First output vector and second output vector are handled using preset third analysis model, obtain institute
State the 1st grade of knowledge point classification output of target topic.
4. the knowledge point classification method of mathematical problem according to claim 3, which is characterized in that described to use preset third
Analysis model handles first output vector and second output vector, obtains the 1st grade of the target topic
Knowledge point classification, which exports, includes:
First output vector and second output vector are merged into the mixing output vector being as follows:
CombOutVec=(CbOutVal1,CbOutVal2,...,CbOutValod,,...,CbOutValODN)
Wherein, od is the serial number of each 1st grade of knowledge point classification, and 1≤od≤ODN, ODN are the sum of the 1st grade of knowledge point classification,
CbOutValodFor the value of the mixing output vector in the od dimension, and CbOutValod=WdOutValod+
SbOutValod, WdOutValodFor value of first output vector in the od dimension, SbOutValodIt is described
Value of two output vectors in the od dimension, CombOutVec are the mixing output vector;
The probability of the entitled each 1st grade of knowledge point classification of the target is calculated separately according to the following formula:
Wherein, Exp is natural exponential function, ProbodFor the probability of entitled od the 1st grade of knowledge point classifications of the target;
The 1st grade of knowledge point classification for choosing maximum probability is classified as the 1st grade of knowledge point of the target topic to be exported.
5. the knowledge point classification method of mathematical problem according to any one of claim 1 to 4, which is characterized in that it is described from
Target topic is obtained in preset mathematics exam pool includes:
Current pattern identification is read, and determines the classification mode for carrying out knowledge point classification according to the pattern identification;
If the classification mode be preset first mode, receiving terminal apparatus send the first sort instructions, described first
Sort instructions are that the instruction of knowledge point classification is carried out to the specified mathematical problem in the mathematics exam pool;
The topic mark in first sort instructions is extracted, and selection is corresponding with topic mark from the mathematics exam pool
Mathematical problem as the target topic;
If the classification mode be preset second mode, receiving terminal apparatus send the second sort instructions, described second
Sort instructions are that all non-classified mathematical problems in the mathematics exam pool are carried out with the instruction of knowledge point classification;
The status indicator of each mathematical problem in the mathematics exam pool is obtained respectively, and state mark is chosen from the mathematics exam pool
Each mathematical problem for preset first value is known as the target topic, and first value represents non-classified state;
If the classification mode is preset the third mode, current system time is obtained, when the system time is default
Triggering moment when, obtain the status indicator of each mathematical problem in the mathematics exam pool respectively, and from the mathematics exam pool
Choosing status indicator is each mathematical problem of first value as the target topic.
6. a kind of knowledge point sorter of mathematical problem characterized by comprising
Topic obtains module, for obtaining target topic, the entitled pending knowledge of target from preset mathematics exam pool
The mathematical problem of point classification;
Characteristic information extracting module, for extracting the characteristic information of the target topic;
First processing module, for the characteristic information of the target topic to be input in preset 1st grade of classifier
Reason obtains the 1st grade of knowledge point classification output of the target topic;
Second processing module, for preceding i grades of knowledge points of the characteristic information of the target topic and the target topic to be classified
It is input in preset i+1 grade classifier and is handled, the i+1 grade knowledge point classification for obtaining the target topic is defeated
Out, i >=1;
Counting module, for i to be increased a digit;
Series judgment module, for judging whether i is less than preset knowledge point classification series;
As a result determining module, if being equal to knowledge point classification series for i, by preceding i grades of knowledge points of the target topic
Knowledge point classification results of the classification output as the target topic.
7. the knowledge point sorter of mathematical problem according to claim 6, which is characterized in that the feature information extraction mould
Block includes:
Topic split cells wraps in the set of words for the target topic to be split as set of words and assemble of symbol
The each word for constituting the target topic is included, the assemble of symbol summarizes each mathematics symbol including constituting the target topic
Number;
Word matrix construction unit, for searching each word in the set of words respectively in preset term vector database
The term vector of language, and the term vector of each word is configured to the word matrix of the target topic, the term vector database
The database of corresponding relationship between record word and term vector;
Sign matrix construction unit, it is each in the assemble of symbol for being searched respectively in preset symbolic vector database
The symbolic vector of symbol, and the symbolic vector of each symbol is configured to the sign matrix of the target topic, the symbol to
Measure the database of corresponding relationship of the database between record symbol and symbolic vector;
Characteristic information construction unit, for believing using the word matrix and the sign matrix as the feature of the target topic
Breath.
8. the knowledge point sorter of mathematical problem according to claim 7, which is characterized in that the first processing module packet
It includes:
First processing units, for using preset first analysis model to the word matrix in the characteristic information at
Reason, obtains the first output vector;
The second processing unit, for using preset second analysis model to the sign matrix in the characteristic information at
Reason, obtains the second output vector;
Third processing unit, for using preset third analysis model to first output vector and described second export to
Amount is handled, and the 1st grade of knowledge point classification output of the target topic is obtained.
9. a kind of computer readable storage medium, the computer-readable recording medium storage has computer-readable instruction, special
Sign is, the mathematical problem as described in any one of claims 1 to 5 is realized when the computer-readable instruction is executed by processor
Knowledge point classification method the step of.
10. a kind of server, including memory, processor and storage can transport in the memory and on the processor
Capable computer-readable instruction, which is characterized in that realized when the processor executes the computer-readable instruction as right is wanted
The step of knowledge point classification method of mathematical problem described in asking any one of 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910623328.1A CN110472044A (en) | 2019-07-11 | 2019-07-11 | Knowledge point classification method, device, readable storage medium storing program for executing and the server of mathematical problem |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910623328.1A CN110472044A (en) | 2019-07-11 | 2019-07-11 | Knowledge point classification method, device, readable storage medium storing program for executing and the server of mathematical problem |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110472044A true CN110472044A (en) | 2019-11-19 |
Family
ID=68507259
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910623328.1A Pending CN110472044A (en) | 2019-07-11 | 2019-07-11 | Knowledge point classification method, device, readable storage medium storing program for executing and the server of mathematical problem |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110472044A (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3012807A1 (en) * | 2013-06-17 | 2016-04-27 | Digital Collaborations Corporation | Knowledge management device, knowledge management device terminal device and knowledge management device program |
CN105677747A (en) * | 2015-12-30 | 2016-06-15 | 华夏博雅(北京)教育科技发展有限公司 | Method and system for managing knowledge points in online teaching |
CN106599054A (en) * | 2016-11-16 | 2017-04-26 | 福建天泉教育科技有限公司 | Method and system for title classification and push |
CN106776724A (en) * | 2016-11-16 | 2017-05-31 | 福建天泉教育科技有限公司 | A kind of exercise question sorting technique and system |
CN107894999A (en) * | 2017-10-27 | 2018-04-10 | 成都准星云学科技有限公司 | Towards the topic type automatic classification method and system based on thinking of solving a problem of elementary mathematics |
CN108182177A (en) * | 2018-01-24 | 2018-06-19 | 谢德刚 | A kind of mathematics knowledge-ID automation mask method and device |
CN108920450A (en) * | 2018-06-06 | 2018-11-30 | 广东小天才科技有限公司 | A kind of knowledge point methods of review and electronic equipment based on electronic equipment |
CN109635100A (en) * | 2018-12-24 | 2019-04-16 | 上海仁静信息技术有限公司 | A kind of recommended method, device, electronic equipment and the storage medium of similar topic |
-
2019
- 2019-07-11 CN CN201910623328.1A patent/CN110472044A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3012807A1 (en) * | 2013-06-17 | 2016-04-27 | Digital Collaborations Corporation | Knowledge management device, knowledge management device terminal device and knowledge management device program |
CN105677747A (en) * | 2015-12-30 | 2016-06-15 | 华夏博雅(北京)教育科技发展有限公司 | Method and system for managing knowledge points in online teaching |
CN106599054A (en) * | 2016-11-16 | 2017-04-26 | 福建天泉教育科技有限公司 | Method and system for title classification and push |
CN106776724A (en) * | 2016-11-16 | 2017-05-31 | 福建天泉教育科技有限公司 | A kind of exercise question sorting technique and system |
CN107894999A (en) * | 2017-10-27 | 2018-04-10 | 成都准星云学科技有限公司 | Towards the topic type automatic classification method and system based on thinking of solving a problem of elementary mathematics |
CN108182177A (en) * | 2018-01-24 | 2018-06-19 | 谢德刚 | A kind of mathematics knowledge-ID automation mask method and device |
CN108920450A (en) * | 2018-06-06 | 2018-11-30 | 广东小天才科技有限公司 | A kind of knowledge point methods of review and electronic equipment based on electronic equipment |
CN109635100A (en) * | 2018-12-24 | 2019-04-16 | 上海仁静信息技术有限公司 | A kind of recommended method, device, electronic equipment and the storage medium of similar topic |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111222305B (en) | Information structuring method and device | |
CN108733644B (en) | A kind of text emotion analysis method, computer readable storage medium and terminal device | |
CN110019843A (en) | The processing method and processing device of knowledge mapping | |
CN108509413A (en) | Digest extraction method, device, computer equipment and storage medium | |
CN111783394B (en) | Training method of event extraction model, event extraction method, system and equipment | |
CN108491406B (en) | Information classification method and device, computer equipment and storage medium | |
CN111105209B (en) | Job resume matching method and device suitable for person post matching recommendation system | |
CN106202044A (en) | A kind of entity relation extraction method based on deep neural network | |
US10831993B2 (en) | Method and apparatus for constructing binary feature dictionary | |
CN109918560A (en) | A kind of answering method and device based on search engine | |
CN109902307A (en) | Name the training method and device of entity recognition method, Named Entity Extraction Model | |
CN106202032A (en) | A kind of sentiment analysis method towards microblogging short text and system thereof | |
CN108959474B (en) | Entity relation extraction method | |
CN111475615B (en) | Fine granularity emotion prediction method, device and system for emotion enhancement and storage medium | |
CN111666376B (en) | Answer generation method and device based on paragraph boundary scan prediction and word shift distance cluster matching | |
CN105975455A (en) | Information analysis system based on bidirectional recursive neural network | |
CN110134961A (en) | Processing method, device and the storage medium of text | |
CN109492093A (en) | File classification method and electronic device based on gauss hybrid models and EM algorithm | |
CN110232128A (en) | Topic file classification method and device | |
CN109472310A (en) | Determine the recognition methods and device that two parts of resumes are the identical talent | |
CN109726285A (en) | A kind of file classification method, device, storage medium and terminal device | |
CN109800309A (en) | Classroom Discourse genre classification methods and device | |
CN112100464B (en) | Question-answering community expert recommendation method and system combining dynamic interest and professional knowledge | |
CN113407644A (en) | Enterprise industry secondary industry multi-label classifier based on deep learning algorithm | |
CN104699819A (en) | Sememe classification method and device |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191119 |
|
RJ01 | Rejection of invention patent application after publication |