A novel meta-embedding technique for drug reviews sentiment analysis

<p>Traditional word embedding models have been used in the feature extraction process of deep learning for sentiment analysis. However, these ignore properties words while maintaining contextual relationships and inadequate representation domainspecific words. This paper proposes a method to develop meta model by exploiting domain polarity adverse drug reaction (ADR) features render more suitable medical The proposed lexicon is developed from blogs corpus. scores existing lexicons are adjusted assign new score each word. neural network utilizes ADR refined embedding. obtained approach concatenated with original vectors, form meta-embedding which maintains both sentimental properties. final acts as extractor assess effectiveness reviews experiments conducted on global vectors (GloVE) skip-gram word2vector (Word2Vec) models. empirical results demonstrate outperforms traditional different performance measures.</p>