CN104700833A - Big data speech classification method - Google Patents
Big data speech classification method Download PDFInfo
- Publication number
- CN104700833A CN104700833A CN201410844027.9A CN201410844027A CN104700833A CN 104700833 A CN104700833 A CN 104700833A CN 201410844027 A CN201410844027 A CN 201410844027A CN 104700833 A CN104700833 A CN 104700833A
- Authority
- CN
- China
- Prior art keywords
- data
- voice
- big data
- large data
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a big data speech classification method. The big data speech classification method comprises the following steps of 1 collecting speed samples as a training set, 2 looking for a big data speech classification optimized frequency spectrum matrix, 3 conducting spectral analysis on unmarked data, and 4 adopting frequency bands to classify big data speeches according to frequency spectrum data. By means of the big data speech classification method, feature information of different speech data can be effectively found under the big data situation, accordingly various relevant data are effectively classified, storage cost in the training process is effectively reduced, and the classification accuracy is higher than the accuracy in the prior art.
Description
Technical field
The present invention relates to a kind of large data-voice sorting technique.
Background technology
Along with developing rapidly of mobile Internet, more and more enter the life of people with digital camera smart mobile phone, panel computer, be easy to produce a large amount of individual voice messagings.Although utilizing time and catalogue to manage speech data is a kind of common method, lacks semantic level and voice are effectively managed.Therefore supervised learning method is utilized, by learning artificial labeled data, to obtain Classification of Speech model, then to not having the voice marked to carry out automatic speech classification.Because the common intrinsic dimensionality of voice is very high, therefore fourier transform method contributes to the raising of recognition performance.
The linear Fourier transformation method of traditional overall situation is mainly based on linear, and wherein linear discriminant analysis is widely used on pattern classification problem.Fisher face to make while between class distance in class sample separation from minimum mainly through maximizing, thus realize different classes of between separability.But it is huge that large data image classification is faced with classification number, the sample size of needs classification is huge waits difficulty.Linear discriminant analysis is for large data, and use cost is higher, and in order to obtain certain classification performance, it needs manually a large amount of mark samples.This makes Classification of Speech software development cost roll up, and needs manually a large amount of mark samples.
Therefore, find one to need to mark the automatic speech sorting technique that a small amount of sample can be met requirement and be very important.
Summary of the invention
Technical matters to be solved by this invention is to provide a kind of large data-voice sorting technique, reduces software development cost, rationally effectively classifies to a large amount of speech datas, distinguish and process.
The technical scheme that the present invention solves the problems of the technologies described above is as follows: a kind of large data-voice sorting technique, comprises the steps:
1) speech samples is collected as training set;
2) spectral matrix that the classification of large data-voice is optimum is found;
3) spectrum analysis is carried out to without labeled data;
4) frequency range is adopted to classify to large data-voice to frequency spectrum data.
Preferably, the spectral matrix that the classification of described searching large data-voice is optimum, comprises the following steps:
Step 1, set up local optimum objective function;
Step 2, set up global optimization objective function;
Step 3, utilize Fourier Transform Algorithm: by the question variation of new global optimization target for asking generalized eigenvalue problem, the optimum spectral matrix of large data-voice classification is by formula
x (n, k) corresponding to front m minimal eigenvalue.
The invention has the beneficial effects as follows: under can effectively finding out large data cases, the characteristic information of different phonetic data thus obtain effective classification of all kinds of related data, effectively reduce the carrying cost in training process, its classify accuracy is higher than prior art.
Accompanying drawing explanation
Fig. 1 is one-piece construction schematic diagram of the present invention;
Embodiment
Be described principle of the present invention and feature below in conjunction with accompanying drawing, example, only for explaining the present invention, is not intended to limit scope of the present invention.
As shown in Figure 1, a kind of large data-voice sorting technique, comprises the steps:
1) speech samples is collected as training set;
2) spectral matrix that the classification of large data-voice is optimum is found;
3) spectrum analysis is carried out to without labeled data;
4) frequency range is adopted to classify to large data-voice to frequency spectrum data.
Preferably, the spectral matrix that the classification of described searching large data-voice is optimum, comprises the following steps:
Step 1, set up local optimum objective function;
Step 2, set up global optimization objective function;
Step 3, utilize Fourier Transform Algorithm: by the question variation of new global optimization target for asking generalized eigenvalue problem, the optimum spectral matrix of large data-voice classification is by formula
x (n, k) corresponding to front m minimal eigenvalue.
Compared with prior art, the invention has the advantages that, under can effectively finding out large data cases, the characteristic information of different phonetic data thus obtain effective classification of all kinds of related data, effectively reduce the carrying cost in training process, its classify accuracy is higher than prior art.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (2)
1. a large data-voice sorting technique, is characterized in that, comprise the steps:
1) speech samples is collected as training set;
2) spectral matrix that the classification of large data-voice is optimum is found;
3) spectrum analysis is carried out to without labeled data;
4) frequency range is adopted to classify to large data-voice to frequency spectrum data.
2. large data-voice sorting technique according to claim 1, is characterized in that, the spectral matrix that the classification of described searching large data-voice is optimum, comprises the following steps:
Step 1, set up local optimum objective function;
Step 2, set up global optimization objective function;
Step 3, utilize Fourier Transform Algorithm: by the question variation of new global optimization target for asking generalized eigenvalue problem, the optimum spectral matrix of large data-voice classification is by formula
x (n, k) corresponding to front m minimal eigenvalue.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410844027.9A CN104700833A (en) | 2014-12-29 | 2014-12-29 | Big data speech classification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410844027.9A CN104700833A (en) | 2014-12-29 | 2014-12-29 | Big data speech classification method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104700833A true CN104700833A (en) | 2015-06-10 |
Family
ID=53347891
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410844027.9A Pending CN104700833A (en) | 2014-12-29 | 2014-12-29 | Big data speech classification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104700833A (en) |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH1020011A (en) * | 1996-07-02 | 1998-01-23 | Oki Electric Ind Co Ltd | Method and device for analyzing azimuth |
US20060085190A1 (en) * | 2004-10-15 | 2006-04-20 | Microsoft Corporation | Hidden conditional random field models for phonetic classification and speech recognition |
CN101196888A (en) * | 2006-12-05 | 2008-06-11 | 云义科技股份有限公司 | System and method for using digital audio characteristic set to specify audio frequency |
CN101196983A (en) * | 2006-12-08 | 2008-06-11 | 北京皎佼科技有限公司 | Image recognition method |
CN101847412A (en) * | 2009-03-27 | 2010-09-29 | 华为技术有限公司 | Method and device for classifying audio signals |
CN101067930B (en) * | 2007-06-07 | 2011-06-29 | 深圳先进技术研究院 | Intelligent audio frequency identifying system and identifying method |
CN102426835A (en) * | 2011-08-30 | 2012-04-25 | 华南理工大学 | Method for identifying local discharge signals of switchboard based on support vector machine model |
CN102968639A (en) * | 2012-09-28 | 2013-03-13 | 武汉科技大学 | Semi-supervised image clustering subspace learning algorithm based on local linear regression |
CN103440313A (en) * | 2013-08-27 | 2013-12-11 | 复旦大学 | Music retrieval system based on audio fingerprint features |
CN103488744A (en) * | 2013-09-22 | 2014-01-01 | 华南理工大学 | Big data image classification method |
CN103985381A (en) * | 2014-05-16 | 2014-08-13 | 清华大学 | Voice frequency indexing method based on parameter fusion optimized decision |
-
2014
- 2014-12-29 CN CN201410844027.9A patent/CN104700833A/en active Pending
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH1020011A (en) * | 1996-07-02 | 1998-01-23 | Oki Electric Ind Co Ltd | Method and device for analyzing azimuth |
US20060085190A1 (en) * | 2004-10-15 | 2006-04-20 | Microsoft Corporation | Hidden conditional random field models for phonetic classification and speech recognition |
CN101196888A (en) * | 2006-12-05 | 2008-06-11 | 云义科技股份有限公司 | System and method for using digital audio characteristic set to specify audio frequency |
CN101196983A (en) * | 2006-12-08 | 2008-06-11 | 北京皎佼科技有限公司 | Image recognition method |
CN101067930B (en) * | 2007-06-07 | 2011-06-29 | 深圳先进技术研究院 | Intelligent audio frequency identifying system and identifying method |
CN101847412A (en) * | 2009-03-27 | 2010-09-29 | 华为技术有限公司 | Method and device for classifying audio signals |
CN102426835A (en) * | 2011-08-30 | 2012-04-25 | 华南理工大学 | Method for identifying local discharge signals of switchboard based on support vector machine model |
CN102968639A (en) * | 2012-09-28 | 2013-03-13 | 武汉科技大学 | Semi-supervised image clustering subspace learning algorithm based on local linear regression |
CN103440313A (en) * | 2013-08-27 | 2013-12-11 | 复旦大学 | Music retrieval system based on audio fingerprint features |
CN103488744A (en) * | 2013-09-22 | 2014-01-01 | 华南理工大学 | Big data image classification method |
CN103985381A (en) * | 2014-05-16 | 2014-08-13 | 清华大学 | Voice frequency indexing method based on parameter fusion optimized decision |
Non-Patent Citations (4)
Title |
---|
余鹏飞等: "《基于乐纹特征和倒排索引的音乐检索***》", 《计算机应用与软件》 * |
朱冰莲等: "《认知无线电中基于QoS分级的频谱分配策略》", 《计算机工程》 * |
李鑫滨等: "《基于离散人工蜂群算法的认知无线电频谱分配》", 《***工程与电子技术》 * |
高洪元等: "《膜量子蜂群优化的多目标频谱分配》", 《物理学报》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109034159B (en) | Image information extraction method and device | |
CN108536870B (en) | Text emotion classification method fusing emotional features and semantic features | |
CN101247470B (en) | Method realized by computer for detecting scene boundaries in videos | |
CN107491435B (en) | Method and device for automatically identifying user emotion based on computer | |
CN111079406A (en) | Natural language processing model training method, task execution method, equipment and system | |
CN102722713B (en) | Handwritten numeral recognition method based on lie group structure data and system thereof | |
CN104166706A (en) | Multi-label classifier constructing method based on cost-sensitive active learning | |
CN103854645A (en) | Speech emotion recognition method based on punishment of speaker and independent of speaker | |
CN103839078A (en) | Hyperspectral image classifying method based on active learning | |
CN109492105A (en) | A kind of text sentiment classification method based on multiple features integrated study | |
Abir et al. | Bangla handwritten character recognition with multilayer convolutional neural network | |
CN111191033A (en) | Open set classification method based on classification utility | |
Royo-Letelier et al. | Disambiguating music artists at scale with audio metric learning | |
CN103488744A (en) | Big data image classification method | |
Nguyen et al. | Matching pursuit based robust acoustic event classification for surveillance systems | |
Abdelazim et al. | Automatic Dialect identification of Spoken Arabic Speech using Deep Neural Networks | |
CN112489689A (en) | Cross-database voice emotion recognition method and device based on multi-scale difference confrontation | |
CN110019556A (en) | A kind of topic news acquisition methods, device and its equipment | |
Li et al. | Review network for scene text recognition | |
CN103886097A (en) | Chinese microblog viewpoint sentence recognition feature extraction method based on self-adaption lifting algorithm | |
George et al. | Unsupervised query-by-example spoken term detection using segment-based bag of acoustic words | |
Pei et al. | Enhancing aphid detection framework based on ORB and convolutional neural networks | |
CN104700833A (en) | Big data speech classification method | |
CN109993381B (en) | Demand management application method, device, equipment and medium based on knowledge graph | |
CN113868389B (en) | Data query method and device based on natural language text and computer equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150610 |