Problem Statement
With the start of the covid pandemic everyone was stranded at their homes and this decreased the social interaction that we used to have when we were in campus. It is very crucial to bond with a community having similar interests, as it leads to getting new opportunities.
This project bridges this gap of lack of interaction by providing the users a topocentric platform to connect with each other on the basis of their personal interests.
This project bridges this gap of lack of interaction by providing the users a topocentric platform to connect with each other on the basis of their personal interests.
It uses machine learning and AI to bring people with the same interest together where they can interact and talk about their own interests and explore all the opportunities they have missed before with the help of each other.
Features
> Interactive UI to input User's interests
> Recommendation algorithm in the backend to calculate correlation between 2 users based on their profile
> Displaying a list of top 10 users that match the most with the current user
> Providing an option to talk to the top matched user using MS Teams
Tech Stack
Python, Joblib, Pandas, Numpy, Streamlit
Source Code
Key Steps/Algorithms
> Tokenizing and Lemmatizing dataset bios
> Finding most popular bigrams in user bios
> Scaling categories and using PCA to reduce dimensionality
> Fitting dataset into Hierarchal Agglomerative Clustering Algorithm
> Finding optimum clusters using Calinski Harbasz, Silhouette Coefficient, and Davs Boudin algorithms
