Article Preview
TopIntroduction
Authorship Attribution aids to recognize the right author of a specified unnamed article from a set of contender authors. Authors have presented a rigorous exploration of state-of-art methods in this regard already (Rajesh Shardanand Prasad, Midhun Chakkaravarthy, 2020). The applicability of this task can be originated in numerous areas, for example law enforcement interventions and data storage and retrieval. These application domains are not restricted to a explicit language, communal, or culture. However, most of the prevailing solutions are intended for English, and a slight consideration has been rewarded to regional languages.
Figure 1 shows the major challenges to apply this author attribution research to efficiently and effectively to regional languages or languages other than English. They are listed as follows:
- •
Handle outliers in the dataset;
- •
Scale when the size of the candidate authors set increases; and
- •
Perform well when the number of writing samples for each candidate author is low.
The first and foremost step includes identification of stylometric feature space. Further tasks of feature extraction and identification may employ state-of-art classifiers.
Most of the research literature review and their evaluation emphasize on the effectiveness of each stylometric features. The major findings of this study state that combination of all groups of the stylometric features outclasses the contemporary combinations. To conclude, authors cross equates the feature spaces and classification approaches of all systems. This literature evaluation concludes that Hybridization of stylometric features enhances the performance of the system increases with the increase in the number of candidate authors. Furthermore, such hybridized feature space provides improved efficiency than the feature space used by the non-hybridized feature space.
Figure 1. Major challenges before author attribution researchers
TopAuthors hereby begin the review with state-of-the-art and vigorous method to stylometric investigation starved of explanation and grasping etymological and sub-etymological evidence (Hou, Renkui & Huang, Chu-Ren., 2019). Especially, authors suggest to influence the lingual information of qualities and rimes in Mandarin Chinese mechanically pulled out from annotated texts. The texts from dissimilar writers were characterized by manners, manner themes, and expression span themes as well as rimes and rime themes. This methodology ensures effective results in author stylometric analysis with the use of Support vector machines and random forests.