PrakritiGPT

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

Idea/Approach

The objective is to predict the Prakriti of the user based on a few pre-determined questions.

• We can achieve that by using pre-existing data and training a machine learning model to predict the ‘Prakriti’ of the user.

• This has been implemented into a chatbot that runs on NLP-like model made by using Deep Learning, specifically a feed-forward neural network made by using the Tensorflow module.

• The chatbot is then hosted on a webpage using a ‘jinja2’ template using the flask module in Python.

• A Discord bot has also been made for brief information pertaining to the Prakriti assessment using an async ready API wrapper for discord written in Python.

Technology Stack:

• Frontend: HTML, CSS, JavaScript

• Backend: Python, flask

The project uses the following:

Pandas, numpy, json, tensorflow, sklearn and flask modules in Python.

Discord module in Python for the discord bot.

HTML CSS and JavaScript for the frontend

Credits

prakriti-analysis/ml/mlml.py at main · puter-nyaani-Soma/prakriti-analysis

prakriti-analysis/ml/deeplearning.ipynb at main · puter-nyaani-Soma/prakriti-analysis

Contribution

Team Lead - Ritesh Das:

• Developed the discord bot • Edited README.md

Ankan Das:

• Developed the backend of the chatbot model

Arka Nandi:

• Developed the frontend of the chatbot model

Saptarshi Nath:

• Assisted Ritesh Das in making the discord bot

Aishwaryya Bhattacharyya:

• Made the powerpoint presentation • Gathered questionnaire for chatbot model/ discord bot • Assisted Ankan Das in making changes in intents.json file

Nirban Das: