IJERT-Study of Different Methods for Author Identification

International Journal of Computer Applications, 2013

Trends in data mining are increasing over the time. Current world is of internet and everything is available over internet, which leads to criminal and malicious activity. So the identity of available content is now a need. Available content is always in the form of text data. Authorship analysis is the statistical study of linguistic and computational characteristics of the written documents of individuals. This paper describes review of various methods for authorship analysis and identification for a set of provided text. Surely research in authorship analysis and identification will continue and even increase over decades. In this article, we put our vision of future authorship analysis and identification with high performance and solution for behavioral feature extraction from set of text documents.

stat.unipg.it

In the text classification field, great relevance is given to the classification task. In classical contest the SVM's classifier shows good performance. However, if a new document, belonging to an unknown author, has to be classified, the above method fails, because of the inability to detect the text as an example of a new category. In order to overcome this problem, a new method, based on the combined employment of decision tree and Kruskal-Wallis test, is proposed. The methodology applied is evaluated in terms of correspondence analysis.

2008

In this work we discuss author identification for documents written in Portuguese. Two different approaches were compared. The first is the writer-independent model which reduces the pattern recognition problem to a single model and two classes, hence, makes it possible to build robust system even when few genuine samples per writer are available. The second is the personal model, which very often performs better but needs a bigger number of samples per writer. We also introduce a stylometric feature set based on the conjunctions and adverbs of the Portuguese language. Experiments on a database composed of short articles from 30 different authors and Support Vector Machine (SVM) as classifier demonstrate that the proposed strategy can produced results comparable to the literature.

Journal of Computing and Security, 2015

Due to increasing criminal activities by anonymous E-mails in the cyber world, it is a challenging task to extract beneficial knowledge from E-mail systems. This problem in cyber world attracts many researchers in cyber-crime domain. Recent studies in this area concentrate on traditional classification approaches such as Decision Tree and Support Vector Machines (SVM). These approaches are employed to identify the author. The main goal of these researches is increasing the accuracy of identification, but the quality of evidence is ignored and also it is hard to be traced. So, in this paper, we propose a new approach based on data mining methods for improving the quality of evidence which leads to boost the accuracy of identification. We use writeprints as the evidence and extract them from each E-mail of individuals. The next step for author identification, is matching the writeprints with anonymous E-mails by applying Earth Mover Distance (EMD) criterion to identify the plausible author. In addition to high accuracy, EMD can help cybercrime investigators in making decision about anonymous intruder. Experiments with real data in both English and Persian languages, demonstrate the proposed approach can effectively identify the author and capture strong evidence to prove the identification.

Journal of The American Society for Information Science and Technology, 2006

With the rapid proliferation of Internet technologies and applications, misuse of online messages for inappropriate or illegal purposes has become a major concern for society. The anonymous nature of online-message distribution makes identity tracing a critical problem. We developed a framework for authorship identification of online messages to address the identity-tracing problem. In this framework, four types of writing-style features (lexical, syntactic, structural, and content-specific features) are extracted and inductive learning algorithms are used to build feature-based classification models to identify authorship of online messages. To examine this framework, we conducted experiments on English and Chinese online-newsgroup messages. We compared the discriminating power of the four types of features and of three classification techniques: decision trees, backpropagation neural networks, and support vector machines. The experimental results showed that the proposed approach was able to identify authors of online messages with satisfactory accuracy of 70 to 95%. All four types of message features contributed to discriminating authors of online messages. Support vector machines outperformed the other two classification techniques in our experiments. The high performance we achieved for both the English and Chinese datasets showed the potential of this approach in a multiple-language context.

International Journal of Multidisciplinary Studies and Innovative Technologies, 2022

Author identification is one of the application areas of text mining. It deals with the automatic prediction of the potential author of an electronic text among predefined author candidates by using author specific writing styles. In this study, we conducted an experiment for the identification of the author of a Turkish language text by using classical machine learning methods including Support Vector Machines (SVM), Gaussian Naive Bayes (GaussianNB), Multi Layer Perceptron (MLP), Logistic Regression (LR), Stochastic Gradient Descent (SGD) and ensemble learning methods including Extremely Randomized Trees (ExtraTrees), and eXtreme Gradient Boosting (XGBoost). The proposed method was applied on three different sizes of author groups including 10, 15 and 20 authors obtained from a new dataset of newspaper articles. Term frequency-inverse document frequency (TF-IDF) vectors were created by using 1-gram and 2-gram word tokens. Our results show that the most successful method is the SGD with a classification performance accuracy of 0.976% by using word unigrams and most successful method is the LR with a classification performance accuracy of 0.935% by using word bigrams.

2015

In the present research work, we deal with the problem of authorship attribution of ancient Arabic text documents, which were written by several ancient philosophers. For that purpose, we conducted several authorship attribution experiments applied with different text sizes. A special dataset, called "A4P" (Authorship Attribution for Ancient Arabic Philosophers), has been constructed by extracting texts of different sizes from the books of those 5 ancient Arabic philosophers, where the genre and topic are quite similar. The size of the texts varies from 100 words to 3000 words per text. In our approach two types of features are employed; character N-grams and words and several classifiers are used, namely: SMO based SVM, Multi Layer Perceptron, Linear Regression, Stamatatos distance and Manhattan distance. Results show that the minimum required text size (for getting good authorship attribution performances) depends on the used features and classification technique, but in the overall the performances of the proposed techniques are quite interesting.

As a result of Internet's anonymity, openness and freedom, there have been two kinds of security problems existing extensively in the online virtual space. They are user's ID theft and article counterfeit that may cause rampant cybercrime and online deception. Therefore, application of authorship or article identification techniques across online message content is increasingly necessary. In this paper, we proposed a feature-based identification method which can automatically identify the authenticity of online authorship or article according to their unique writing style. This technique is a hypothesis testing based method that creates a function words feature set specially tailored towards Chinese online messages. To evaluate the effectiveness of this automatic recognition method, we conducted several experiments on datasets extracted from a Chinese forum web, and statistically analyzed the content of registered uses' post and reply messages. Our experimental results showed that this method has an excellent identification performance.

SoutheastCon 2016, 2016

Authorship Identification is the task of identifying who wrote a given piece of text from a given set of candidate authors (suspects). The increasingly large volumes of texts on the Internet enhance the great yet urgent necessity for authorship identification. For this purpose, a large amount of work has already been done for the English language. Comparatively, less research has been carried out for Indian regional languages such as Tamil, Telugu, Bengali and Punjabi whereas no such experiment is available for Marathi. In this study presented a strategy for authorship identification of the documents written in Marathi language. Moreover, we adopted a set of fine-grained lexical and stylistic features for the analysis of the text and used them to develop two different models (statistical similarity model and SMORDT-Sequential minimal optimization with rulebased Decision Tree approach). Then, we validated the feature extraction method to show consistent significance in every model used in this experiment. The performance of the proposed approach has been evaluated based on the values of Recall,Precision, F-measure and Accuracy.

2015

In this paper we focus on the identification of the author of a written text. We present a new hybrid method that combines a set of stylistic and statistical features in a machine learning process. We tested the effectiveness of the linguistic and statistical features combined with the inter-textual distance "Delta" on the PAN’@CLEF’2015 English corpus and we obtained 0.59 as c@1 precision.