India’s largest professional platform VakilSearch where hundreds of new companies and entrepreneurs register to use their services everyday, we are creating a chatbot to assist those potential customers with their various doubts and questions which may inevitably arise while navigating Vakil Search for getting help for their Legal, Tax and Compliance Services.
We will be feeding outr model with dataset in the JSON format which is very easy to comprehend and edit (just like a simple word document).
Then we tokenized the text content from probabilistic prediction (for which we have used porter stemmer) during this we have ignored the stop words/common words like is,am,was etc.
Then with these tokenized words we create a bag of words which will be used to identify from the given query by user.
Then we use of simple feed forward neural network with activation fiunction as relu to get the probabilistic output as a result.
- Our chatbot uses minimal computer resources and computer memory in contrast to its competition which hogs computer heavy resources.
- Our chatbot can be scaled to a very large extent and can be used on any website and it is completely free of cost whereas our competition only provides basic services for free and charges more for their premium services.
Arushi Gupta 👩💻
Aditya Singh 👨⚖️