Inspiration
Blockchain technology and cryptocurrencies are an incredible advancement that is disrupting established financial industries with its decentralized system. However, this decentralized system that we praise raises a significant issue. In the traditional financial system, the centralized system allows financial institutions like banks to defend your rights in fraud cases. Let's contrast it with blockchain technology. Imagine, you purchased a laptop from an online marketplace, you sent your bitcoins, and the seller decided to stop responding to you and leave you with misery.
What it does
Blockbook is a cryptocurrency wallet that protects you from transaction fraud using credibility evaluation.
How we built it
We built it with react.js for the website integrating Robust Covariance Estimation model and Network Proximity evaluation to measure user credibility.
Challenges we ran into
In order to build a valid and credible machine learning model, we have to use a large amount of real-world training data. In the prototype, we incorporate over 70,000,000 Bitcoin transaction data from 2013 to 2014. Due to the lack of resources and time, computing the model with the large data becomes especially difficult. We tackled that problem by improving the program structure to reduce the computation time.
Accomplishments that we're proud of
Successfully integrating machine learning model within the cryptocurrency wallet that is able to provide practical insight into the user financial safety.
What's next for BlockBook
Builds the function that enables real-time update of the model based on the concurrent Bitcoin transaction data.
Built With
- bcoin
- javascript
- networkx
- python
- react.js
- scikit-learn
Log in or sign up for Devpost to join the conversation.