Exploring KSOM model with word2vec, doc2vec and SBERT models, comparison were performed with LSTMs (bidirectional and unidirectional), conv1d, Supervised BERT models
- The polarity of associated sentiments was visualized by Tableau dashboards
- Read Project_Report.pdf for comprehensive understanding of dataset used, data preprocessing, data modelling, experimented techniques, and the data visualization obtained.
- A typical sentiment plarity plot obtained by training Kohononn's self organizing map using SBERT model looks like this : 
- The performace table of supervised approaches used for sentiment classification task looks like this :
- The performace table of Un-supervised Kohonon's approaches used for sentiment classification task looks like this :
