Dissertation > Excellent graduate degree dissertation topics show
The Speech Recognition Based on Error-Correcting Output Code and Support Vector Machine
Author: LiZuo
Tutor: ZhangXueYing
School: Taiyuan University of Technology
Course: Communication and Information System
Keywords: Speech recognition Support Vector Machine Error-correcting output coding algorithm Hadamard ECOC Hadamard ECOC sparse
CLC: TN912.34
Type: Master's thesis
Year: 2011
Downloads: 53
Quote: 0
Read: Download Dissertation
Abstract
|
With the increase in demand for human-computer interaction , speech recognition has become a hot topic of research , and broad development prospects . The support vector machine is used to solve the multi- class classification has many advantages , it is an important tool in the study of speech recognition . The support vector machine is the essence of a planning optimization problem , it can find the global optimum , to overcome the deficiencies of the traditional methods . Support vector machine , however , is mainly used to solve the problem of binary classification , and how it applies to multi - class classification , researchers have put forward a number of algorithms : one-on-one \acyclic graph method and ECOC algorithm . In this thesis, a combination of speech recognition system based on the the ECOC algorithm and support vector machine , the full text of the following major elements : (1) in-depth study of the error-correcting output coding algorithm . ECOC algorithm has some of the advantages of their own , can classification error can be corrected to a certain extent , thereby giving a more accurate result of the discrimination . (2) the ECOC algorithm and support vector machine combined with voice recognition systems , the decoder decoding based index lost constructed in this article . Application of a variety of encoding to perform voice recognition experiments . (3) This article uses the Hadamard ECOC , the algorithm is better than the performance of the identification of the other error-correcting output coding . However, when a large number of categories , the algorithm will be more time - consuming to training to identify network . To solve this problem , this paper Hadamard ECOC algorithm based on constructed sparse Hadamard ECOC algorithm , that is 1 / 2 sparse Hadamard ECOC algorithm and 1 /3 of the sparse Hadamard Error Correcting Output coding algorithm . The experimental results show that the more the number of categories of speech recognition , these two algorithms varying degrees to reduce training time while maintaining a high recognition rate .
|
Related Dissertations
- Research on Automatic Detection Algorithm for Substructure Distress of Highway Pavement Based on SVM,U418.6
- Multiple ANN/HMM Hybrid Used in Speech Recognition,TN912.34
- The Design of a DSP-Based Robot Speech Command Recgnition System,TN912.34
- Research on Transductive Support Vector Machine and Its Application in Image Retrieval,TP391.41
- Research for Infrared Image Target Identification and Tracking Technology,TP391.41
- Research of Diagnosing Cucumber Diseases Based on Hyperspectral Imaging,S436.421
- The New Privacy Preserving Algorithms for Linear Programming and SVMS,O221.1
- Prediction of Binding Affinity of Human Transporter Associated with Antigen Processing,R392.1
- Study on Applications of Soft Sensor in Chloromethane Recovery,TQ222.214
- Research on Classification and Matching for Hand Vein Image,TP391.41
- The Research and Realization of the Military Port Objects Classification Platform,TP751
- Prediction Research of the Surrounding Rock Deformation in Engineering by the Support Vector Machine,TV223.1
- Distortion effects on image quality evaluation and classification,TP391.41
- Based on self-learning social relation extraction research,TP391.1
- Based on Information Fusion Fault Diagnosis of Analog Circuits,TN710
- Classification based on gene expression profiles of tumor,R730.2
- Predicting Protein Protein Interactions and Its Active Sites Based on Data Mining Algorithm,TP311.13
- Research on Relation Extraction of Person Entity in News Webpage,TP393.092
- Research on Action Recognition Based on the Space-Time Interest Points,TP391.41
- Based on Support Vector Machine (SVM) Personal Credit Evaluation,TP18
- Algorithm Research on Video-based Vehicle Detection, Tracking and Recognition for Intelligent Transportation,TP391.41
CLC: > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Electro-acoustic technology and speech signal processing > Speech Signal Processing > Speech Recognition and equipment
© 2012 www.DissertationTopic.Net Mobile
|