LSTM Based Short Message Service (SMS) Modeling for Spam Classification

目录

声明

1 Introduction

1.1What is Spam

1.2Spamming Motivations

1.3.1Research Background

1.3.2Research Significance

1.4Overseas and Domestic Research Progress

1.4.1Spam Detection in Short Message Service (SMS)

1.4.2Spam Detection in Email

1.5.1Main Contents

1.5.2The Structure of the Thesis

1.6 Summary

2Basic Theory and Related Work

2.1Basic Theory of Machine Learning

2.1.1Unsupervised Learning

2.1.2Supervised Learning

2.2Spam Filtering Techniques

2.2.1Machine Learning Approach to Spam Filtering

2.2.2Artificial Neural Network

2.2.3Deep Neural Network

2.3Spam Filtering Challenges for Machine Learning

2.3.1False Positive

2.3.2Concept Drift Handling in SMS

2.3.3E-mail Ranking or Prioritizing

2.4 Summary

3Experimental Model

3.1Proposed method LSTMs

3.2Word Embedding

3.3 Word2Vec

3.3.1Skip-gram Model

3.3.2Continuous Bag-of-Words (CBOW) Model

3.4Data Set

3.5Traditional Baseline Methods

3.5.1SVM (Support Vector Machine)

3.5.2Decision Tree

3.5.3KNN (K-Nearest Neighbors)

3.5.4Random Forest

3.5.5NB (Na(i)ve Bayes)

3.6 Summary

4Results and Discussions

4.1.1Detecting Strategy

4.1.2 Contributions

4.1.3Data Preprocessing

4.2.1Comparative Results

4.2.2Detecting Results

5Conclusion and Future Work

参考文献

Research Projects and Publications in Master Study

致谢

Dalian University of Technology Copyright Use Authorization of Master Degree Dissertation