Aspect Based Sentiment Analysis on Students' Feedback Using Deep Learning

  Atharv Anand,  Akansha Naidu,  Sumit Kumar,  Satyajeet Kumar Sinha,  Supriya Bhosale

Feedback from students is crucial for academic institutions to assess the performance of the faculty. Efficient handling of students' qualitative opinions while generating automatic report is a challenging task. Nonetheless, most institutions deal successfully with quantitative feedback, while qualitative feedback is either manually interpreted or ignored altogether. Therefore, we have proposed a supervised deep learning-based method which strives to improve the quality of feedback system by considering the sentiments along with their respective aspects. The proposed system is based on two-layer LSTM model which uses BERT for word embeddings. The pre-processed data is fed to the BERT embedding layer whose output is further passed on to the LSTM layers for aspect extraction and polarity classification. The task of aspect extraction is conducted by the LSTM layer1 and that of polarity classification by the LSTM layer 2.

Keywords- Aspect Extraction, Sentiment Analysis, Deep Learning, LSTM, BERT, Student Feedback

Unique Identification Number - IJEDR2002015
Page Number(s) - 69-72
Pubished in - Volume 8 | Issue 2 | June 2020
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Atharv Anand,  Akansha Naidu,  Sumit Kumar,  Satyajeet Kumar Sinha,  Supriya Bhosale,   "Aspect Based Sentiment Analysis on Students' Feedback Using Deep Learning", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.8, Issue 2, pp.69-72, June 2020, Available at :http://www.ijedr.org/papers/IJEDR2002015.pdf