Employing Thematic Variables for Enhancing Classification Accuracy Within Author Discrimination Experiments

Tambouratzis, George; Vassiliou, Marina

doi:10.1093/llc/fqm003

Cited by 9 publications

(9 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…They are syntactic, lexical and structural features. The various lexical features are content words, letter frequency, special characters , character n-grams, misspellings in [10,11], special use words in [12], 8 punctuation marks in [13], most frequent types [14], spelling errors, word form errors in [15], syntactically classified punctuation , syntactic structure [16], function word frequencies, POS trigrams or sequences of 3 in [17], word n-grams in [18], POS bigrams or sequences of 2 in [19], unigrams/types shared by training and testing samples in [20], 1024-character sequences in [21], content words, frequent words in [22], words bigrams or sequences in [23], emoticons, netabbrevs in [24], character n-grams or sequences in [25], non-function words in [26], frequency of lemmas (dictionary entry headwords), frequency of negative words in [27]. The various syntactic features are punctuation, function words [10], POS tags [11], syntactically classified punctuation, verbal phrases [15], syntactically classified punctuation [16], function word-token ratios [28], POS trigrams or sequences of 3 [17], punctuation frequency [18], POS bigrams or sequences of 2 [19], PCFG-obtained POS [26], syntactically classified punctuation [29], verbal phrases [30], phrase types, words per phrase type [31].…”

Section: E-mail Address: Nagkanna80@gmailcommentioning

confidence: 99%

Influence of Lexical, Syntactic and Structural Features and their Combination on Authorship Attribution for Telugu Text

Prasad

Narsimha

Reddy³

et al. 2015

Procedia Computer Science

View full text Add to dashboard Cite

Authorship attribution (AA) is the task of identifying author of an unknown text from the known author set. Authorship Attribution can be viewed is a problem of text classification. AA is based on the classification of documents on author writing style rather than the topic of the text. In this paper experimental evaluations were carried out on Telugu text for Authorship Attribution using various types of features and their combinations. Feature vectors were formed for the training set using lexical, syntactic and structural features and their combinations. Learned model was generated for each these vectors and performance of the learned model is calculated using F1 metric and accuracy. More number of features can slow down the model performance. Features which are not relevant or not more relevant were eliminated from the feature vectors using chi-square metric. Support Vector Machine (SVM) algorithm is used as a classifier to generate the learned model for each dimensional feature vector. This learned model is used to assign the anonymous text to one of the known authors.

show abstract

Section: E-mail Address: Nagkanna80@gmailcommentioning

confidence: 99%

Influence of Lexical, Syntactic and Structural Features and their Combination on Authorship Attribution for Telugu Text

Prasad

Narsimha

Reddy³

et al. 2015

Procedia Computer Science

View full text Add to dashboard Cite

show abstract

“…Como esta lista es dependiente del corpus del que es extraída, el tamaño de la misma varió entre 1,402 tipos para el corpus más pequeño, hasta 13,089 para el corpus más grande. Cabe aclarar que la puntuación fue removida de las unidades léxicas a las que se unía y los signos separados fueron utilizados como unigramas léxicos independientes, un procedimiento común en la atribución de autoría [9], [20], [25] y [26]. En cuanto a los rasgos sintácticos, se utilizó una lista previamente recabada (para otra tarea clasificatoria) con elementos léxicos funcionales pluriverbales, es decir, con más de una palabra.…”

Section: Rasgos Clasificatorios De Autoríaunclassified

Técnicas de reducción, algoritmos resistentes al ruido o ambos. Opciones para el manejo de rasgos clasificatorios en la atribución de autoría

Sulayes¹

2014

RCS

View full text Add to dashboard Cite

Resumen. Un problema en las tareas clasificatorias es el manejo de los rasgos que caracterizan las clases. Cuando la lista de rasgos es larga, se puede usar un algoritmo resistente al ruido de rasgos irrelevantes, o reducir dichos rasgos. La atribución de autoría, tarea que asigna un texto anónimo a un sujeto en una lista de posibles autores, ha sido ampliamente abordada como una tarea de clasificación automática de texto. En dicha tarea, los n-gramas pueden producir listas largas de rasgos incluso en corpus pequeños. A pesar de esto, falta una investigación que exponga los efectos de usar algoritmos resistentes al ruido, reducir los rasgos, o combinar ambas opciones. Este trabajo responde a esta carencia utilizando contribuciones a foros de discusión relacionados con el crimen organizado. Los resultados obtenidos muestran que los clasificadores evaluados, en general, se benefician con la reducción de rasgos, y que gracias a dicha reducción, incluso algoritmos clásicos superan a clasificadores de punta considerados altamente resistentes al ruido.Palabras clave: Atribución de autoría, Rasgos clasificatorios, Algoritmos resistentes al ruido, Reducción de rasgos.

show abstract

“…Generation of multiple clusterings. The notion that text collections may be clustered in multiple independent ways has been discussed in the literature on computational stylistics (see Lim, Lee, & Kim, 2005;Biber & Kurjian, 2006;Grieve-Smith, 2006;Tambouratzis & Vassiliou, 2007;Gries, Wulff, & Davies, 2010, for example). In machine learning, there have been attempts to design algorithms for producing multiple clusterings of a dataset.…”

Section: Related Workmentioning

confidence: 99%

Which Clustering Do You Want? Inducing Your Ideal Clustering with Minimal Feedback

Dasgupta¹,

Ng²

2010

jair

View full text Add to dashboard Cite

While traditional research on text clustering has largely focused on grouping documents by topic, it is conceivable that a user may want to cluster documents along other dimensions, such as the author's mood, gender, age, or sentiment. Without knowing the user's intention, a clustering algorithm will only group documents along the most prominent dimension, which may not be the one the user desires. To address the problem of clustering documents along the user-desired dimension, previous work has focused on learning a similarity metric from data manually annotated with the user's intention or having a human construct a feature space in an interactive manner during the clustering process. With the goal of reducing reliance on human knowledge for fine-tuning the similarity function or selecting the relevant features required by these approaches, we propose a novel active clustering algorithm, which allows a user to easily select the dimension along which she wants to cluster the documents by inspecting only a small number of words. We demonstrate the viability of our algorithm on a variety of commonly-used sentiment datasets.2. We follow Ng et al. (2001) and employ a normalized dual form of the usual Laplacian D − S.

show abstract

Employing Thematic Variables for Enhancing Classification Accuracy Within Author Discrimination Experiments

Cited by 9 publications

References 18 publications

Influence of Lexical, Syntactic and Structural Features and their Combination on Authorship Attribution for Telugu Text

Influence of Lexical, Syntactic and Structural Features and their Combination on Authorship Attribution for Telugu Text

Técnicas de reducción, algoritmos resistentes al ruido o ambos. Opciones para el manejo de rasgos clasificatorios en la atribución de autoría

Which Clustering Do You Want? Inducing Your Ideal Clustering with Minimal Feedback

Contact Info

Product

Resources

About