Multi-level aspect based sentiment classification of Twitter data: using hybrid approach in deep learning

• DOI: 10.7717/peerj-cs.433
• Publication Date: 2021-04-13T11:40:07Z
• Authors: Sadaf Hussain Janjua; Ghazanfar Farooq ...; Muddassar Azam Si...; Umer Rashid
• Keywords: Sentiment Analysis; Perceptron

Social media is a vital source to produce textual data, further utilized in various research fields. It has been considered an essential foundation for organizations get valuable data assess the users’ thoughts and opinions on specific topic. Text classification procedure assign tags predefined classes automatically based their contents. The aspect-based sentiment analysis classify text challenging. Every work related approached this issue as current usually discusses document-level overall sentence-level rather than particularities of sentiments. This aims use Twitter perform finer-grained at aspect-level by considering explicit implicit aspects. study proposes new Multi-level Hybrid Aspect-Based Sentiment Classification (MuLeHyABSC) approach embedding feature ranking process with amendment selection method comprising Artificial Neural Network; Multi-Layer Perceptron (MLP) used attain improved results. In study, different machine learning methods were also implemented, including Random Forest (RF), Support Vector Classifier (SVC), seven more classifiers compare proposed method. implementation hybrid shown better performance efficiency system was validated multiple datasets manifest domains. We achieved results all validation purpose accuracy 78.99%, 84.09%, 80.38%, 82.37%, 84.72%, respectively, compared baseline approaches. revealed that functionality enhanced, it outperformed existing classification.