Using psycholinguistic features for profiling first language of authors

Abstract

This study empirically evaluates the effectiveness of different feature types for the classification of the first language of an author. In particular, it examines the utility of psycholinguistic features, extracted by the Linguistic Inquiry and Word Count (LIWC) tool, that have not previously been applied to the task of author profiling. As LIWC is a tool that has been developed in the psycholinguistic field rather than the computational linguistics field, it was hypothesized that it would be effective, both as a single type feature set because of its psycholinguistic basis, and in combination with other feature sets, because it should be sufficiently different to add insight rather than redundancy. It was found that LIWC features were competitive with previously used feature types in identifying the first language of an author, and that combined feature sets including LIWC features consistently showed better accuracy rates and average F measures than were achieved by the same feature sets without the LIWC features. As a secondary issue, this study also examined how effectively first language classification scaled up to a larger number of possible languages. It was found that the classification scheme scaled up effectively to the entire 16 language collection from the International Corpus of Learner English, when compared with results achieved on just 5 languages in previous research.

Learning Stylometric Representations for Authorship Analysis

“…We argue that the division of the whole feature space according to linguistic modalities is necessary because different application scenarios have different requirements of the features. Moreover, even for manual feature engineering, writing styles that correspond to different linguistic dimensions are constructed differently by humans [Solorio et al 2011;Torney et al 2012]; therefore, they need to be grouped together.…”

“…Usually the frequency value of the function words are used to represent the features [Baron 2014;Halvani and Steinebach 2014;HaCohen-Kerner and Margaliot 2014]. They are effective for identifying the first language of the authors [Torney et al 2012;Argamon et al 2009], identifying the actual author of French literature [Boukhaled and Ganascia 2014], and characterizing the gender of e-mails [Corney et al 2002]. Especially for the first language detection, the effect of language transfer affects the use of function words in the secondary language [Torney et al 2012].…”

“…Features of the character-level modality are concerned with the morphology and phonemes biases in the process of constructing/spelling a single lexical word [Torney et al 2012]. The typical feature used for character modality is the character n-gram feature.…”

The Role of Emotions in Native Language Identification

“…While emotion-based features have been used in other NLP tasks, such as sentiment analysis (Sidorov et al, 2013), classification of documents into the corresponding emotion category (Wen and Wan, 2014), deception detection (Newman et al, 2003), among others, they are an underexplored area of second language writing. Torney et al (2012) use psycholinguistic features extracted by the Linguistic Inquiry and Word Count (LIWC) tool (Pennebaker et al, 2007) to identify the first language of an author, where emotion-based features are included as part of the feature vector, e.g., percentage of positive/negative emotion words. The LIWC feature set used in the paper also contains other types of features, e.g., personal concern categories (work, leisure), paralinguistic dimensions (assents, fillers, nonfluencies), which obscure the contribution of the actual emotion features.…”

Ngrams and Engrams: the use of structural and conceptual features to discriminate between English translations of religious texts

“…LIWC has been used in a number of applications, but particularly in psychology; for example, by comparing the scores of the writing of mental health patients over time, Pennebaker (1997) was able to track changes in cognitive and emotional processes and pinpoint potential triggers for recovery. It has also been used in author profiling (Mairesse and Walker, 2006) and first-language profiling (Torney et al, 2012), among others. Figure 1 gives examples of some of the words that make up the different LIWC dimensions.…”