Implementation of Multimodal Sentiment Analysis using Dense Fusion Multimodal Residual(DFMR) in Malayalam with DravidianMultiModal Dataset. Textual (Video Transcripts), Audio, Visual (Video Frames) are the modalities considered for Sentiment Analysis. This Deep Learning Model could classify a video into Highly Negative, Negative, Neutral, Positive and Highly Positive Sentiments. Tokenization of Textual Data in Malayalam is done using pre-trained model from IndicBERT.

Python3 | Tensorflow | Keras | Jupyter Notebook | Natural Language Processing (NLP)

Chatbot that can handle general queries, chitchat, FAQs and out of scope queries. Other features include table reservation and food ordering. Chitchat involves handling insults, telling jokes etc. FAQs involve enquiry regarding restaurant opening and closing time, parking facility and other restaurant policies.

Python3 | RASA | REST API | Natural Language Processing (NLP)