Features level sentiment mining in enterprise systems from informal text corpus using machine learning techniques

ABSTRACT

This study explores feature-level sentiment analysis of Hindi blog reviews in enterprise systems, a significant area in the Indian context yet understudied. By applying machine learning techniques like SVM across unigram, bigram, trigram, and n-gram models, and combining Lexicon-based methods with machine learning algorithms, we aim to enhance sentiment classification for better customer relationship management and product development. Contrasting with document-level approaches, our method focusing on bigrams achieves a test accuracy of 75%, offering a scalable model for enterprises to extract detailed customer insights from informal text, thereby aiding informed decision-making in a multicultural environment.

KEYWORDS:

• Sentiment mining
• natural language processing
• Hindi language
• Hindi blog reviews
• knowledge discovery
• machine learning techniques

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. http://members.unine.ch/jacques.savoy/clef/index.html

2. http://users.aber.ac.uk/rkj/