decentralized pseudo-anonymous browser-based user and content recommendation(Youtube, Spotify, Twitter, medium, Reddit) with the click of a button on chrome extension upon submitting your browsing history using a model with privacy-focused.
User signs up and signs in with a meta mask wallet. Login Signup with - ENS Ethereum General address. Users have an option to choose a pseudo-anonymous username. Post signup user gets a prompt of chrome extension of fetching associated data to create the profile features For web2 based data, it will be consent-based only being local encrypted data storage Example data for the analysis (via user permission): youtube hyperlinks: Spotify All data Medium Quora question Reddit posts Also possible to fetch the corresponding web3 data : theGraph Dune Covalent . Rarible multichain And then, using the lens protocol we will search for the given graph parameters. Hallelujah, the integration of the lens protocol Integration of the schemas of the different events in the webspace. User first gets its data stored in the local instance of the gunDB, which then is being used by the ML model in order to find the similarity with the current onchain detail parameters, and then based on its functionality, there will be the score of the profile predicted by the model. This browsing history data is being matched with other users' data(history) for matching with other users' history and finding a likely match. Collaborative recommendation for getting feeds(content embedded youtube videos and open sea art collection images and embedded quora question, embedded tweets) in other users based on their match with other user browsing history(content).
How it's made
this project uses metamask for signing user on the platform uses chrome extension powered by JS to record browsing history based on the permission of the user, this history will be saved on-orbit-dB in key-value format on his localhost instance. uses content-based and collaborative implicit matrix factorization-based machine learning models to predict and recommend the relevant user and content to users based on the numbers in matrix-vector/feature with help of tensorflow.js and browser's compute for recommendation without giving any data, which will be integrated into the browser extension. recommendation system: Currently, we are using a python based simple tf-idf based matrix factorization recommendation on the backend, but eventually will be done ideally by the local using browser compute. uses lens protocol to write smart contracts for building SOCIAL PLATFORM Lens Protocol is a user-owned, open social graph that any application can plug into. Mint a profile, follow others, create and collect any publications, including posts, comments, and mirrors, completely on-chain.