Author Identification from Literary Articles with Visual Features: A Case Study with Bangla Documents

Author identification is an important aspect of literary analysis, studied in natural language processing (NLP). It aids identify the most probable author articles, news texts or social media comments and tweets, for example. can be applied to other domains such as criminal civil cases, cybersecurity, forensics, plagiarizer, many more. An automated system this context thus very beneficial society. In paper, we propose a convolutional neural network (CNN)-based from articles. This uses visual features along with five-layer authors. The prime motivation behind approach was feasibility distinct writing styles through visualization patterns. Experiments were performed on 1200 articles 50 authors achieving maximum accuracy 93.58%. Furthermore, see how different volumes data, experiments partitions dataset. outperformed standard handcrafted feature-based techniques well established works publicly available datasets.