Dissertation > Excellent graduate degree dissertation topics show
Research on Document Classification Algorithm Based on Semi-Supervised Learning
Author: QinFei
Tutor: YangYan
School: Southwest Jiaotong University
Course: Applied Computer Technology
Keywords: Text Classification Semi-supervised learning Nearest neighbor (NN) Similar samples
CLC: TP391.1
Type: Master's thesis
Year: 2010
Downloads: 208
Quote: 0
Read: Download Dissertation
Abstract
|
With the development of network information technology , people need to deal with in their daily work more and more text messages , text categorization as key technologies in this area increasing attention in recent years , traditional text classification requires a large amount of a known category text to help build a classifier usually the case , however , we can only get a small number of samples of known classes and a lot of unknown category sample , using only known class samples to build the classifier , not only the results obtained with certain limitations , the unknown class sample implicit information it is difficult to be effectively utilized , which caused a certain degree of waste of resources , and manual marking is unknown category sample also requires a lot of manpower and resources , semi-supervised learning came into being . Semi-supervised learning is a cross between supervised learning and unsupervised learning , a learning way , and only need part of the known classes of training samples , combined with the unknown category samples containing the knowledge to learn to build a classifier . On the basis of the existing semi-supervised classification algorithms proposed can effectively improve the classification performance of semi - supervised classification method based on a majority vote , and combined with text categorization , and propose a new method for an expanded sample set is known , the paper 's main work is as follows : 1. introduces the key technologies for text classification , including a representation of the text , the text pre-processing , feature selection , feature weight calculation , a common classification and classification performance assessment . 2 introduces the concept of semi-supervised learning , and combined with the existing semi-supervised classification algorithm , and the introduction of majority voting rules based on the nearest neighbor , the effectiveness of the method is proved by experiments . 3 the idea of the semi - supervised classification used for text classification , and a small sample of a join similar sample made ??according to the characteristics of the text semi-supervised learning method , the method first extracted by a known set of class samples representing each category Representative characteristics , according to these representatives characterized samples from unknown category set selected similar sample was added to the sample set of known classes , expand the size of the sample set of known categories , and then subsequent learning . The Chinese classification experiment using a standard data set to verify the effectiveness of the method .
|
Related Dissertations
- Research on Text Classification Based on Biomimetic Pattern Recongnition,TP391.1
- Tourism Comments on the Internet’s Semantic Analysis and Usefulness Research,TP391.1
- Based on Data Distribution Characteristics of Text Classification,TP391.1
- Semi-supervised Learning and Active Learning of Sentiment Classification Coupled with Domain Knowledge,TP181
- Semi-supervised hashing algorithm based Image Retrieval Methods,TP391.41
- Research and Implementation of Time Series Classification Based on Semi-supervised Learning,TP181
- Research on the Application of Data Mining Technology in the Regulating Matriculation for Postgraduate,TP311.13
- Comparison and Improvement of Two Methods Based on Semi-Supervised Learning,TP18
- Spam Filtering System Study Based on Non-negative Matrix Factorization,TP393.098
- Research on Intrusion Detection Based on Semi-Supervised SVM,TP393.08
- Sentiment Classification by Combining Lexicon-based and Machine Learning Methods,TP391.1
- The Research and Implementation of Text Classification Based on Meta-Information and Optimization,TP391.1
- The LVCSR system based on adaptive methods of semi-supervised learning,TN912.34
- Design and Implementation of Hydropower Fault Classifer Based on Support Vector Machine,TV738
- A Study on Some Problems of Semi-supervised Learning,TP181
- Cross-language Text Classification Research Based on Latent Semantic Dual Space,TP391.1
- Research of Spam Filtering System Based on Logistic Regression Model,TP393.098
- Researching on Chinese Text Classification Based on Naive Bayes and BP Neural Network,TP391.1
- Text Classification Technology and Its Application in Assistant Decision-Making,TP391.1
- Remote sensing image classification based on semi-supervised learning,TP751
- Semantic Analysis of Text Based on Emotional Classification,TP391.1
CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Text Processing
© 2012 www.DissertationTopic.Net Mobile
|