CN110399455A - A kind of deep learning data digging method based on CNN and LSTM - Google Patents
A kind of deep learning data digging method based on CNN and LSTM Download PDFInfo
- Publication number
- CN110399455A CN110399455A CN201910488085.5A CN201910488085A CN110399455A CN 110399455 A CN110399455 A CN 110399455A CN 201910488085 A CN201910488085 A CN 201910488085A CN 110399455 A CN110399455 A CN 110399455A
- Authority
- CN
- China
- Prior art keywords
- cnn
- lstm
- data
- method based
- deep learning
- 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
- G06F16/33—Querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2216/00—Indexing scheme relating to additional aspects of information retrieval not explicitly covered by G06F16/00 and subgroups
- G06F2216/03—Data mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of deep learning data digging method based on CNN and LSTM, comprising the following steps: the feature vector for successively collecting each target object obtains raw data packets;To data prediction, character information library is established;Construct CNN-LSTM algorithm model;CNN-LSTM algorithm model is trained, and is tested;Data mining is carried out using CNN-LSTM algorithm model;Data feedback is obtained, is learnt.In the present invention, data mining is completed in study by building LSTM-CNN algorithm model, the efficiency of acquisition of information is substantially increased, solves the problems, such as that conventional information transfers difficulty, learner is helped to improve learning effect;Wherein CNN quickly identifies the corresponding character information of keyword, and LSTM quickly extends to the relevant information of keyword, supplemented using the form of replicated blocks chain, so as to transfer speed fast for information, information transfers accuracy height.
Description
Technical field
The present invention relates to learning areas more particularly to a kind of deep learning data digging methods based on CNN and LSTM.
Background technique
Currently, electronization is realized in the work of each department and each area substantially.Due to realizing that the time of electronization is not grown, believe
Breath amount is excessively huge, so that the accuracy of information excavating is low, speed is slow, has delayed constituent parts and personal study.
To solve the above problems, proposing a kind of deep learning data digging method based on CNN and LSTM in the application.
Summary of the invention
(1) goal of the invention
To solve technical problem present in background technique, the present invention proposes a kind of deep learning based on CNN and LSTM
Data digging method, the present invention complete data mining in study by building LSTM-CNN algorithm model, substantially increase letter
The efficiency obtained is ceased, solves the problems, such as that conventional information transfers difficulty, learner is helped to improve learning effect;Wherein CNN is to key
The corresponding character information of word is quickly identified that LSTM quickly prolongs the relevant information of keyword using the form of replicated blocks chain
Exhibition, supplement, so as to transfer speed fast for information, information transfers accuracy height.
(2) technical solution
To solve the above problems, the present invention provides a kind of deep learning data digging method based on CNN and LSTM, packet
Include following steps:
S1, the feature vector for successively collecting each target object, obtain raw data packets;
S2, to data prediction, establish character information library;
S3, building CNN-LSTM algorithm model;
S4, CNN-LSTM algorithm model is trained, and tested;
S5, data mining is carried out using CNN-LSTM algorithm model;
S6, data feedback is obtained, is learnt.
Preferably, in S1, feature vector is multiple attribute datas of target object, including party member's name, identity card letter
Breath, party membership, time of joining the party, the time limit of joining the party, post, work location.
Preferably, in S2, the mode of data prediction are as follows: with name, identity card and party member's information for main keyword,
Classify to the data in raw data packets, and be preset format by Data Format Transform, obtains character information library.
Preferably, in S3, CNN nerve network system and LSTM nerve network system are series connection modeling.
Preferably, in S4, the CNN model of construction includes input layer, convolutional layer, full articulamentum and output layer;It is wherein defeated
The length for entering the one-dimensional data sequence of layer input is 2k+1;Convolution kernel is one-dimentional structure, size 2k+1 in convolutional layer;Quan Lian
It connects layer and builds decline in CNN model;The downstream of full articulamentum is arranged in output layer.
Preferably, in S4, LSTM includes length memory layer and at least one full articulamentum.
Preferably, when carrying out data mining in S5, keyword is inputted first, CNN is to the corresponding character information of keyword
It is quickly identified, LSTM quickly extends to the relevant information of keyword, supplemented using the form of replicated blocks chain.
It preferably, further include correction system;In S6, the data that learner is fed back, further according to the accuracy of data
It is corrected, constantly improve model.
It preferably, further include record system;Record system records the data content of excavation, time, frequency, according to
Statistical conditions preferentially push learner, reduce the time of data mining.
Above-mentioned technical proposal of the invention has following beneficial technical effect:
One, in the present invention, data mining is completed in study by building LSTM-CNN algorithm model, substantially increases letter
The efficiency obtained is ceased, solves the problems, such as that conventional information transfers difficulty, learner is helped to improve learning effect;Wherein CNN is to key
The corresponding character information of word is quickly identified that LSTM quickly prolongs the relevant information of keyword using the form of replicated blocks chain
Exhibition, supplement, so as to transfer speed fast for information, information transfers accuracy height.
Two, in the present invention, the correction system and record system of setting improve the accuracy and convenience of data mining,
Further improving method system, promotes learning effect.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the deep learning data digging method based on CNN and LSTM proposed by the present invention.
Specific embodiment
In order to make the objectives, technical solutions and advantages of the present invention clearer, With reference to embodiment and join
According to attached drawing, the present invention is described in more detail.It should be understood that these descriptions are merely illustrative, and it is not intended to limit this hair
Bright range.In addition, in the following description, descriptions of well-known structures and technologies are omitted, to avoid this is unnecessarily obscured
The concept of invention.
As shown, a kind of deep learning data digging method based on CNN and LSTM proposed by the present invention, including it is following
Step:
S1, the feature vector for successively collecting each target object, obtain raw data packets;
S2, to data prediction, establish character information library;
S3, building CNN-LSTM algorithm model;
S4, CNN-LSTM algorithm model is trained, and tested;
S5, data mining is carried out using CNN-LSTM algorithm model;
S6, data feedback is obtained, is learnt.
In an alternative embodiment, in S1, feature vector is multiple attribute datas of target object, including party member
Name, ID card information, party membership, time of joining the party, the time limit of joining the party, post, work location.
In an alternative embodiment, in S2, the mode of data prediction are as follows: with name, identity card and party member letter
Breath is main keyword, is classified to the data in raw data packets, and is preset format by Data Format Transform, obtains word
Accord with information bank.
In an alternative embodiment, in S3, CNN nerve network system and LSTM nerve network system are that series connection is built
Mould.
In an alternative embodiment, in S4, the CNN model of construction include input layer, convolutional layer, full articulamentum and
Output layer;Wherein the length of the one-dimensional data sequence of input layer input is 2k+1;Convolution kernel is one-dimentional structure in convolutional layer, big
Small is 2k+1;Full articulamentum builds the decline in CNN model;The downstream of full articulamentum is arranged in output layer.
In an alternative embodiment, in S4, LSTM includes length memory layer and at least one full articulamentum.
In an alternative embodiment, when carrying out data mining in S5, keyword is inputted first, CNN is to keyword
Corresponding character information is quickly identified that LSTM quickly prolongs the relevant information of keyword using the form of replicated blocks chain
Exhibition, supplement.
It in an alternative embodiment, further include correction system;In S6, the data that learner is fed back, then root
It is corrected according to the accuracy of data, constantly improve model.
It in an alternative embodiment, further include record system;Record system is to the data content of excavation, time, frequency
Rate is recorded, and is preferentially pushed according to statistical conditions to learner, reduces the time of data mining.
In the present invention, data mining is completed in study by building LSTM-CNN algorithm model, substantially increases information
The efficiency of acquisition solves the problems, such as that conventional information transfers difficulty, learner is helped to improve learning effect;Wherein CNN is to keyword
Corresponding character information is quickly identified that LSTM quickly prolongs the relevant information of keyword using the form of replicated blocks chain
Exhibition, supplement, so as to transfer speed fast for information, information transfers accuracy height;The correction system and record system of setting, improve number
According to the accuracy and convenience of excavation, further improving method system promotes learning effect.
It should be understood that above-mentioned specific embodiment of the invention is used only for exemplary illustration or explains of the invention
Principle, but not to limit the present invention.Therefore, that is done without departing from the spirit and scope of the present invention is any
Modification, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.In addition, appended claims purport of the present invention
Covering the whole variations fallen into attached claim scope and boundary or this range and the equivalent form on boundary and is repairing
Change example.
Claims (9)
1. a kind of deep learning data digging method based on CNN and LSTM, which comprises the following steps:
S1, the feature vector for successively collecting each target object, obtain raw data packets;
S2, to data prediction, establish character information library;
S3, building CNN-LSTM algorithm model;
S4, CNN-LSTM algorithm model is trained, and tested;
S5, data mining is carried out using CNN-LSTM algorithm model;
S6, data feedback is obtained, is learnt.
2. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
In S1, feature vector is multiple attribute datas of target object, including party member's name, ID card information, party membership, when joining the party
Between, the time limit of joining the party, post, work location.
3. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
In S2, the mode of data prediction are as follows: with name, identity card and party member's information for main keyword, in raw data packets
Data classify, and by Data Format Transform be preset format, obtain character information library.
4. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
In S3, CNN nerve network system and LSTM nerve network system are series connection modeling.
5. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
In S4, the CNN model of construction includes input layer, convolutional layer, full articulamentum and output layer;A wherein dimension of input layer input
Length according to sequence is 2k+1;Convolution kernel is one-dimentional structure, size 2k+1 in convolutional layer;Full articulamentum is built in CNN mould
The decline of type;The downstream of full articulamentum is arranged in output layer.
6. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
In S4, LSTM includes length memory layer and at least one full articulamentum.
7. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
When carrying out data mining in S5, input keyword, CNN quickly identify the corresponding character information of keyword first,
LSTM quickly extends to the relevant information of keyword, is supplemented using the form of replicated blocks chain.
8. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
It further include correction system;In S6, the data that learner is fed back are corrected further according to the accuracy of data, constantly complete
Kind model.
9. a kind of deep learning data digging method based on CNN and LSTM according to claim 1, which is characterized in that
It further include record system;Record system records the data content of excavation, time, frequency, according to statistical conditions to study
Person is preferentially pushed, and the time of data mining is reduced.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910488085.5A CN110399455A (en) | 2019-06-05 | 2019-06-05 | A kind of deep learning data digging method based on CNN and LSTM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910488085.5A CN110399455A (en) | 2019-06-05 | 2019-06-05 | A kind of deep learning data digging method based on CNN and LSTM |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110399455A true CN110399455A (en) | 2019-11-01 |
Family
ID=68323046
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910488085.5A Pending CN110399455A (en) | 2019-06-05 | 2019-06-05 | A kind of deep learning data digging method based on CNN and LSTM |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110399455A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111177378A (en) * | 2019-12-20 | 2020-05-19 | 北京淇瑀信息科技有限公司 | Text mining method and device and electronic equipment |
CN111915218A (en) * | 2020-08-14 | 2020-11-10 | 中国工商银行股份有限公司 | Financial counterfeiting identification method and system based on LSTM-CNN |
CN112052853A (en) * | 2020-09-09 | 2020-12-08 | 国家气象信息中心 | Text positioning method of handwritten meteorological archive data based on deep learning |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107220506A (en) * | 2017-06-05 | 2017-09-29 | 东华大学 | Breast cancer risk assessment analysis system based on deep convolutional neural network |
CN107832289A (en) * | 2017-10-12 | 2018-03-23 | 北京知道未来信息技术有限公司 | A kind of name entity recognition method based on LSTM CNN |
CN108009674A (en) * | 2017-11-27 | 2018-05-08 | 上海师范大学 | Air PM2.5 concentration prediction methods based on CNN and LSTM fused neural networks |
CN109814523A (en) * | 2018-12-04 | 2019-05-28 | 合肥工业大学 | Method for diagnosing faults based on CNN-LSTM deep learning method and more attribute time series datas |
CN109840279A (en) * | 2019-01-10 | 2019-06-04 | 山东亿云信息技术有限公司 | File classification method based on convolution loop neural network |
-
2019
- 2019-06-05 CN CN201910488085.5A patent/CN110399455A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107220506A (en) * | 2017-06-05 | 2017-09-29 | 东华大学 | Breast cancer risk assessment analysis system based on deep convolutional neural network |
CN107832289A (en) * | 2017-10-12 | 2018-03-23 | 北京知道未来信息技术有限公司 | A kind of name entity recognition method based on LSTM CNN |
CN108009674A (en) * | 2017-11-27 | 2018-05-08 | 上海师范大学 | Air PM2.5 concentration prediction methods based on CNN and LSTM fused neural networks |
CN109814523A (en) * | 2018-12-04 | 2019-05-28 | 合肥工业大学 | Method for diagnosing faults based on CNN-LSTM deep learning method and more attribute time series datas |
CN109840279A (en) * | 2019-01-10 | 2019-06-04 | 山东亿云信息技术有限公司 | File classification method based on convolution loop neural network |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111177378A (en) * | 2019-12-20 | 2020-05-19 | 北京淇瑀信息科技有限公司 | Text mining method and device and electronic equipment |
CN111177378B (en) * | 2019-12-20 | 2023-09-26 | 北京淇瑀信息科技有限公司 | Text mining method and device and electronic equipment |
CN111915218A (en) * | 2020-08-14 | 2020-11-10 | 中国工商银行股份有限公司 | Financial counterfeiting identification method and system based on LSTM-CNN |
CN112052853A (en) * | 2020-09-09 | 2020-12-08 | 国家气象信息中心 | Text positioning method of handwritten meteorological archive data based on deep learning |
CN112052853B (en) * | 2020-09-09 | 2024-02-02 | 国家气象信息中心 | Text positioning method of handwriting meteorological archive data based on deep learning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110399455A (en) | A kind of deep learning data digging method based on CNN and LSTM | |
CN105512289B (en) | Image search method based on deep learning and Hash | |
CN108985317B (en) | Image classification method based on separable convolution and attention mechanism | |
CN103208123B (en) | Image partition method and system | |
CN105740773A (en) | Deep learning and multi-scale information based behavior identification method | |
CN110119673A (en) | Noninductive face Work attendance method, device, equipment and storage medium | |
CN104067314A (en) | Human-shaped image segmentation method | |
CN108228832B (en) | Time series data completion method based on distance matrix | |
CN109165337B (en) | Method and system for establishing bid and ask field association analysis based on knowledge graph | |
CN107292259A (en) | The integrated approach of depth characteristic and traditional characteristic based on AdaRank | |
CN106407208A (en) | Establishment method and system for city management ontology knowledge base | |
CN111062329B (en) | Unsupervised pedestrian re-identification method based on augmented network | |
CN106022363A (en) | Method for recognizing Chinese characters in natural scene | |
CN110008841B (en) | Expression recognition model construction method and system | |
KR102593835B1 (en) | Face recognition technology based on heuristic Gaussian cloud transformation | |
CN111080241A (en) | Internet platform-based data-based talent management analysis system | |
CN106294715A (en) | A kind of association rule mining method based on attribute reduction and device | |
CN106021442A (en) | Network news outline extraction method | |
CN105095857A (en) | Face data enhancement method based on key point disturbance technology | |
CN112418059A (en) | Emotion recognition method and device, computer equipment and storage medium | |
CN109948639A (en) | A kind of picture rubbish recognition methods based on deep learning | |
CN115374189B (en) | Block chain-based food safety tracing method, device and equipment | |
CN102542543A (en) | Block similarity-based interactive image segmenting method | |
CN108664993A (en) | A kind of convolutional neural networks image classification method of intensive weight connection | |
CN113935435A (en) | Multi-modal emotion recognition method based on space-time feature fusion |
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: 20191101 |
|
RJ01 | Rejection of invention patent application after publication |