A comparative study for content-based dynamic spam classification using four machine learning algorithms

YuBo,XuZong-ben

2008-01-01

Abstract:The growth of email users has resulted in the dramatic increasing of the spam emails during the past few years. In this paper, four machine learning algorithms, which are Naive Bayesian (NB), neura...

What problem does this paper attempt to address?

A Comparative Study for Content-Based Dynamic Spam Classification Using Four Machine Learning Algorithms

Bo Yu,Zong-ben Xu

DOI: https://doi.org/10.1016/j.knosys.2008.01.001

IF: 8.139

2008-01-01

Knowledge-Based Systems

Abstract:The growth of email users has resulted in the dramatic increasing of the spam emails during the past few years. In this paper, four machine learning algorithms, which are Naïve Bayesian (NB), neural network (NN), support vector machine (SVM) and relevance vector machine (RVM), are proposed for spam classification. An empirical evaluation for them on the benchmark spam filtering corpora is presented. The experiments are performed based on different training set size and extracted feature size. Experimental results show that NN classifier is unsuitable for using alone as a spam rejection tool. Generally, the performances of SVM and RVM classifiers are obviously superior to NB classifier. Compared with SVM, RVM is shown to provide the similar classification result with less relevance vectors and much faster testing time. Despite the slower learning procedure, RVM is more suitable than SVM for spam classification in terms of the applications that require low complexity.
Content-based Spam Email Detection Using N-gram Machine Learning Approach

Syed Md. Minhaz Hossain,Iqbal H. Sarker

DOI: https://doi.org/10.20944/preprints202109.0236.v1

2021-09-14

Abstract:Recently, spam emails have become a significant problem with the expanding usage of the Internet. It is to some extend obvious to filter emails. A spam filter is a system that detects undesired and malicious emails and blocks them from getting into the users' inboxes. Spam filters check emails for something "suspicious" in terms of text, email address, header, attachments, and language. However, we have used different features such as word2vec, word n-grams, character n-grams, and a combination of variable length n-grams for comparative analysis in our proposed approach. Different machine learning models such as support vector machine (SVM), decision tree (DT), logistic regression (LR), and multinomial naïve bayes (MNB) are applied to train the extracted features. We use different evaluation metrics such as precision, recall, f1-score, and accuracy to evaluate the experimental results. Among them, SVM provides 97.6 \% of accuracy, 98.8\% of precision, and 94.9\% of f1-score using a combination of n-gram features.
Precision in Classification: A Comparative Study of Logistic Regression, Naive Bayes, LSTM, and CNN for Spam Email Detection

Et al. Ratnam Dodda

DOI: https://doi.org/10.17762/ijritcc.v11i9.9233

2023-11-05

International Journal on Recent and Innovation Trends in Computing and Communication

Abstract:This research presents a comprehensive comparative study on the precision of classification algorithms applied to spam email detection. The study focuses on four distinct algorithms: Logistic Regression, Naive Bayes, Long Short-Term Memory (LSTM), and Convolutional Neural Network (CNN). The evaluation encompasses a common dataset to ensure a fair and rigorous assessment of each algorithm’s performance. We investigate and analyze precision metrics, considering the trade- offs between true positives and false positives. The comparative study provides insights into the strengths and limitations of each algorithm in effectively identifying spam emails. The results contribute valuable knowledge to the field of email security and classification, guiding practitioners and researchers towards informed algorithm selection based on precision considerations.
Email Classification Using Behavior and Time Features

Yeqin Shao,Quan Shi,Yanghua Xiao,Nik Bessis,Peter Norrington

DOI: https://doi.org/10.6138/jit.2017.18.3.20140403

2017-01-01

Abstract:The various forms and tremendous number of spam emails have brought great challenges to accurate email classification. In this paper, we present a behavior- and time-feature-based email classification method. Based on email logs, email social networks are built through the extraction of entities and relations from the email records using the MapReduce model. By combining behavior features from social networks and time features from email sending intervals, we adopt a Support Vector Machine based classifier to identify spammers and non-spammers. Compared with the current email classification methods, the advantages of our method are: (1) in addition to the behavior-based features, our method integrates the time feature to facilitate email classification; (2) to efficiently handle the vast number of emails, we employ the MapReduce model to extract the behavior- and time-based features on the email social network. Experiments on real email data of three years show that the proposed method achieves better classification accuracy.
Research On Advanced Filtering Algorithm For Anti-Spam Based On A Bayesian Classification Model

L Zhen,T Liang,S Kun,Mt Zhou

DOI: https://doi.org/10.1142/9789812701695_0012

2005-01-01

Abstract:Currently, naive Bayesian algorithm for Email filtering has been accepted widely and been applied in many commercial Email systems. However, facing with the still severe situation of spam Email, people are realizing gradually that depending on such a simple algorithm can't meet the practical needs for anti-spam. On the other hand, Bayesian network, as an important branch in the knowledge discovery area, has been researched and explored for a long period. The question of Email filter can be projected to a Bayesian network model. By doing so and utilizing Bayesian parameter estimation for some key nodes in the model, we can accomplish Email classification and discrimination based on probabilistic conditions. According to the experimental results, new algorithm has much higher rate to converge and more stable than the naive Bayesian one.
Detecting Spam E-mails with Content and Weight-based Binomial Logistic Model

Richa Indu,Sushil Chandra Dimri

DOI: https://doi.org/10.13052/jwe1540-9589.2271

2024-02-03

Journal of Web Engineering

Abstract:Spam e-mails are continuously increasing and are a serious threats to a network and its users. Several efficient methods are available regarding this context, but still, it is evolving randomly. Considering this, the proposed approach addresses the problem of spam detection by combining traditional content-matching criteria with the modified version of the binomial logistic algorithm. The work generates seven categories for content-matching, which begins from three basic categories, namely: special words, adult content, and specific symbols and digits. The remaining four categories are derived from various possible combinations of these basic categories. The words selected for each category are carefully curated based on the human psychology of action and reaction. Then, a weight is assigned to each of the categories to signify their importance and a threshold criterion is deployed before implementing the binomial logistic algorithm, which not only increases the efficiency of the proposed algorithm but also reduces the rate of misclassification. The proposed model is tested on six separate datasets of Enron Spam Corpus, where 98.31% and 92.575% are the maximum and minimum accuracies achieved, respectively, in spam e-mail classification. The AUC_ROC scores for the entire Spam Corpus range between 0.927 and 0.983. A comparison is also carried out between the proposed algorithm and the other methods of spam detection that have logistic regression. Finally, the suggested method can adequately handle a large sample size without compromising the efficacy, which is measured using accuracy, precision, recall, F-measure, and AUC_ROC score.

computer science, theory & methods, software engineering
Simplified Chinese spam mail filter:design and performance evaluation

Xu Yong

DOI: https://doi.org/10.3321/j.issn:1002-8331.2007.25.039

2007-01-01

Abstract:Paths to solving and methods of filtering unsolicited bulk e-mails,also known as spam,have been analyzed.And the method based on keyword and the statistical learning have been analyzed.Then a new method which is a combination of the two methods have been proposed.The method to filter spam using the na ve Bayesian decision theory,the nearest-neighbor classification,and the linear classification based the perceptron criterion function which is used in pattern classification has been introduced.The feature set used in the three theories have been gotten by mutual information.By comparied the three decision theories,the advantages and disadvantages of them has been presented.At same time,a good idea to filtering spam using mutual information has been pointed out in the paper.
A counting-based method for massive spam mail classification

Hao Luo,Binxing Fang,Xiaochun Yun

DOI: https://doi.org/10.1007/11689522_5

2006-01-01

Abstract:The past research works have explored the effectiveness of machine learning classifiers for filtering spam email, and the results have shown that machine learning classifiers can obtain a high degree of precision and recall. However, these methods cannot avoid classifying normal mail as spam mail for probability characteristics. The evident difference between spam mail and normal mail is that one spam mail will be delivered to many users, while most normal mails have only one single receiver. Based on this observation, this paper presents a server-based massive mail classifier incorporating counting-based classifier, bitmap-based white list (BWL) and grey list to filter massive spam mails. Results show that the spam mail classifier using our method can filter spam with a very low degree of false positive and also preserves performance while coping with large volumes of spam mail. With optimized parameter configuration, our method achieves a precision of 100% and recall of 75.3% in spam mail classification.
Classify E-mails by Support Vector Machine

WANG Qing-Xiang,GUANG Kai,PAN Jin-Gui

DOI: https://doi.org/10.3969/j.issn.1002-137X.2007.09.024

2007-01-01

Computer Science

Abstract:With the rapid explosion of unsolicited bulk e-mail,also known as"spam",has generated a great need for reliable anti-spam e-mail filters.Mail filtering technique is to used to detect the spare and find what’s useful.We dis- cuss the theorem to implement anti-spare,and propose an improved anti-spam mechanism based upon Support Vector Machine,mutual information and Z-test.We demonstrate the performance and positive results of the algorithm on a challenging problem.
Effective spam filter based on a hybrid method of header checking and content parsing

Ko‐Tsung Chu,Hua‐Ting Hsu,Jyh‐Jian Sheu,Wei‐Pang Yang,Cheng‐Chi Lee

DOI: https://doi.org/10.1049/iet-net.2019.0191

2020-11-01

IET Networks

Abstract:In recent years, hazardous e‐mails arose, such as the e‐mails infected with ‘viruses’ or ‘worms’ spreading destructive programs and the ‘Phishing Mails’ defrauding e‐mail accounts of the users. The number of spams continue to grow. With the related problems of spam coming to be more severe, the spam topics have become significant in various research domains. The common filtering methods include black/white list, rule learning, and those based on text classification, such as Naïve Bayes, support vector machine, and boosting trees, multi‐agent and genetic algorithm. Among these, the methods based on text classification are most widely applied. Moreover, some efficient methods were proposed to consider only the e‐mail's header section, based on which both operating efficiency and classification efficiency could be improved. By applying machine learning technique and decision tree data mining algorithm C4.5, this study aims to propose an efficient spam filtering method with the following features: (i) proposing a two‐phase filtering mechanism to scan mainly e‐mail's header and auxiliary content. (ii) Reducing the problem of false positive. The experimental results show that the authors’ method has a considerably high accuracy rate of 98.76%. Compared with some other methods of using the same spam data sets or of deep learning‐based, their method obviously has an excellent performance.
Research on Behavior Statistic Based Spam Filter

Meizhen Wang,Zhitang Li,Ling Xiao,Yunhe Zhang

DOI: https://doi.org/10.1109/etcs.2009.413

2009-01-01

Abstract:The spread of spam made a lot of interference to Internet users, and wasted network bandwidth. Anti-spam technology has been developed to the third-generation technology, behavior recognition technology. Many studies have focused on the detection of abnormal behavior in the period of SMTP conversation. In this paper, user's typical sending behavior mathematical model is concerned. To find the abnormal behavior different from the typical behavior, similarity is computed between behaviors of separate users, or between the same user's behavior in long-time and short-time. Combing the common features of email messages, especially most of spam presence of hyperlink, the URL model is also built. Lastly, by Naiive Bayes model, several behavior recognizing models can cooperate to detect spam.
Adaptive Email Spam Filtering Based on Information Theory

Xin Zhang,Wenyuan Dai,Gui-Rong Xue,Yong Yu

DOI: https://doi.org/10.1007/978-3-540-76993-4_14

2007-01-01

Abstract:Most previous email spam filtering techniques rely on traditional classification learning which assumes the data from training and test sets are drawn from the same underlying distribution. However, in practice, this identical-distribution assumption often violates. In general, email service providers collect training data from various public available resources, while the tasks focus on users' individual inboxes. Topics in the mail-boxes vary among different users, and distributions shift as a result. In this paper, we propose an adaptive email spam filtering algorithm based on information theory which relaxes the identical-distribution assumption and adapts the knowledge learned from one distribution to another. Our work focuses on the content analysis which minimizes the loss in mutual information between email instances and word features, before and after classification. We present theoretical and empirical analyses to show that our algorithm is able to solve the adaptive email spam filtering problem well. The experimental results show that our algorithm greatly improves the accuracy of email filtering, against the traditional classification algorithms, while scaling very well.
Junk Mail Content Detection Using Logistic Regression Algorithm

Hai Tao,Shaoyi Li,Ezinne C. Maxwell-Chigozie,CU Eze,Zunhai Gao,Celestine Iwendi,Zakaria Boulouard

2023-01-01

Abstract:In contemporary times, things have moved away from traditional method to sophisticated way of communication via social media. One of the common ways information is disseminated amongst people is in the use of emails. Emails are very effective, easy and less costly to use by the sender but invariably costly to the recipient. This is due to the effect unwarranted messages which are thrown in tons are being received daily. This paper focus is on developing an effective junk mail content detector to effectively detect the content of messages and properly classify them thereby eliminate spurious emails. Logistic regression and Random Forest algorithms were employed and the result showed thar our model Logistics regression proves a superior performance.
Application of Natural Language Processing and Machine Learning Boosted with Swarm Intelligence for Spam Email Filtering

Nebojsa Bacanin,Miodrag Zivkovic,Catalin Stoean,Milos Antonijevic,Stefana Janicijevic,Marko Sarac,Ivana Strumberger

DOI: https://doi.org/10.3390/math10224173

IF: 2.4

2022-11-09

Mathematics

Abstract:Spam represents a genuine irritation for email users, since it often disturbs them during their work or free time. Machine learning approaches are commonly utilized as the engine of spam detection solutions, as they are efficient and usually exhibit a high degree of classification accuracy. Nevertheless, it sometimes happens that good messages are labeled as spam and, more often, some spam emails enter into the inbox as good ones. This manuscript proposes a novel email spam detection approach by combining machine learning models with an enhanced sine cosine swarm intelligence algorithm to counter the deficiencies of the existing techniques. The introduced novel sine cosine was adopted for training logistic regression and for tuning XGBoost models as part of the hybrid machine learning-metaheuristics framework. The developed framework has been validated on two public high-dimensional spam benchmark datasets (CSDMC2010 and TurkishEmail), and the extensive experiments conducted have shown that the model successfully deals with high-degree data. The comparative analysis with other cutting-edge spam detection models, also based on metaheuristics, has shown that the proposed hybrid method obtains superior performance in terms of accuracy, precision, recall, f1 score, and other relevant classification metrics. Additionally, the empirically established superiority of the proposed method is validated using rigid statistical tests.

mathematics
Active learning based spam filtering method

Zhang Wei,Feng Gao,Lv Di,Feng Xue

DOI: https://doi.org/10.1109/WCICA.2010.5553918

2010-01-01

Abstract:Internet security is seriously threatened by spam spreading, and content-based spam filtering has become one of effective spam-filtering methods. Aiming at the practical problems, we propose an active learning based method which takes naive Bayesian means as basic classifiers. This method randomly initialize a small training set to generate basic classifiers, and then use them to classify mails, which add the most uncertain mail to training set each time to improve the classifier performance. The simulations based on the CCERT mail set show that this method not only reduces the number of mails to be labeled, but also improves classifier accuracy. © 2010 IEEE.
Spammer detection on Sina Micro-Blog

Tian Xian-yun,Yu Guang,Li Peng-yu

DOI: https://doi.org/10.1109/ICMSE.2014.6930212

2014-01-01

Abstract:The popularity of Sina Micro-Blog has made it a great target for spammers. The spreading spammers reduce the political, economic and social value of the Sina Micro-Blog and leads to a bad user experience, which attracts many researchers' attention. Here we introduce the background of the spammer detection and present a study of spammer detection on Micro-Blog. We collected a dataset that includes 1,900 users and more than 340,000 tweets. And different from previous studies which mainly focus on shallow machine learning algorithms, we also used the deep learning model to build a classifier which can be used to discriminate spammers from legitimate users. However, different from deep learning's good performance in other areas, the experimental results suggest that the deep learning model can't learn more information from the user profiles and tweets than the Support Vector Machine although it outperform the Naïve Bayes. The experimental results suggests that we can build a classifier which can correctly discriminate most of spammers from the normal ones, and the deep learning model is not necessarily better than the shallow ones.
Combining Neural Networks and Semantic Feature Space for Email Classification

Bo Yu,Dong-hua Zhu

DOI: https://doi.org/10.1016/j.knosys.2009.02.009

IF: 8.139

2009-01-01

Knowledge-Based Systems

Abstract:Email is one of the most ubiquitous and pervasive applications used on a daily basis by millions of people worldwide, individuals and organizations more and more rely on the emails to communicate and share information and knowledge. However, the increase in email users has resulted in a dramatic increase in spam emails during the past few years. It is becoming a big challenge to process and manage the emails efficiently for and individuals and organizations. This paper proposes new email classification models using a linear neural network trained by perceptron learning algorithm and a nonlinear neural network trained by back-propagation learning algorithm. An efficient semantic feature space (SFS) method is introduced in these classification models. The traditional back-propagation neural network (BPNN) has slow learning speed and is prone to trap into a local minimum, so the modified back-propagation neural network (MBPNN) is presented to overcome these limitations. The vector space model based email classification system suffers from a large number of features and ambiguity in the meaning of terms, which will lead to sparse and noisy feature space. So we use the SFS to convert the original sparse and noisy feature space to a semantically richer feature space, which will helps to accelerate the learning speed. The experiments are conducted based on different training set size and extracted feature size. Experimental results show that the models using MBPNN outperform the traditional BPNN, and the use of SFS can greatly reduce the feature dimensionality and improve email classification performance.
Text Spam Neural Network Classification Algorithm

Qiuming Ma,Zhiguang Qin,Fengli Zhang,Qiao Liu

DOI: https://doi.org/10.1109/icccas.2010.5581954

2010-01-01

Abstract:In order to effectively filter spam e-mails, this paper introduces a neural network-based spam filtering method, described text feature extraction, text vector generation, neural network classifier model building and testing. Verify through experiments that, choose different learning and training function, the network classification results are very different.
Spammer Detection Based On Hidden Markov Model In Micro-Blogging

Qi Dang,Few Gao,Yadong Zhou

DOI: https://doi.org/10.1109/WCICA.2016.7578397

2016-01-01

Abstract:Over the past decade, social network becomes the main way for people to obtain information and communicate with their online friends. With the tremendous increase of social network, spammer also develops rapidly and spam information posted by spammer seriously affects people's daily life. Although the official executes some strict strategies for prevention, spammer evolves dynamically to adapt these strategies at the same time. They disguise as common users with posting interesting tweets and posting spam with irregular time. To deal with this problem, we present a method for spammer detection on the basis of Hidden Markov Model in micro-blogging. The proposed method classifies the user into spammer and non-spammer according to their major count-based features and behavior-based features. The usage of HMM allows us to describe the user behavior with hidden states dynamically, which can be decoded from several observed features. The experimental results based on real data collected from Sina micro-blogging site demonstrate the effectiveness of our method.
A Two-Stage Spam Email Filtering Method Based on Naive Bayes and Hierarchical Clustering

LIAO Ming-tao,ZHANG De-yun,LI Jin-ku

DOI: https://doi.org/10.3969/j.issn.1000-7180.2007.08.001

2007-01-01

Abstract:To reduce misclassification rate of legitimate emails,proposed a two-stage spam email filtering method based on naive Bayes and hierarchical clustering. This method classifies emails as Legitimate,Unsure and Spam. At first stage,it classifies email as Legitimate and Unsure by using naive Bayesian classifier. At second stage,a hierarchical clustering method is used to find similar email in the pre-collected spam emails set. The experiment showed that,this method can increase the precision of spam detection,lower the misclassification of legitimate emails,which is more viable in practice.

A comparative study for content-based dynamic spam classification using four machine learning algorithms

A Comparative Study for Content-Based Dynamic Spam Classification Using Four Machine Learning Algorithms

Content-based Spam Email Detection Using N-gram Machine Learning Approach

Precision in Classification: A Comparative Study of Logistic Regression, Naive Bayes, LSTM, and CNN for Spam Email Detection

Email Classification Using Behavior and Time Features

Research On Advanced Filtering Algorithm For Anti-Spam Based On A Bayesian Classification Model

Detecting Spam E-mails with Content and Weight-based Binomial Logistic Model

Simplified Chinese spam mail filter:design and performance evaluation

A counting-based method for massive spam mail classification

Classify E-mails by Support Vector Machine

Effective spam filter based on a hybrid method of header checking and content parsing

Research on Behavior Statistic Based Spam Filter

Adaptive Email Spam Filtering Based on Information Theory

Junk Mail Content Detection Using Logistic Regression Algorithm

Application of Natural Language Processing and Machine Learning Boosted with Swarm Intelligence for Spam Email Filtering

Active learning based spam filtering method

Spammer detection on Sina Micro-Blog

Combining Neural Networks and Semantic Feature Space for Email Classification

Text Spam Neural Network Classification Algorithm

Spammer Detection Based On Hidden Markov Model In Micro-Blogging

A Two-Stage Spam Email Filtering Method Based on Naive Bayes and Hierarchical Clustering