Crypnosys

• 

  Asset Selector

• 

  Data Visualization

• 

  Closing Price vs Time chart

• 

  Closing Price vs Time chart with 100MA and 200MA

• 

  Stream The Live Data

• 

  Prediction of the prices of the selected asset in next 5 days

Inspiration

The intricacy and ever-changing dynamics of financial markets often surpass the capabilities of conventional financial forecasting methods. Our goal is to enhance the accuracy and reliability of predictions for financial asset prices by employing XGBoost, a powerful algorithm proficient at recognizing temporal dependencies in data. This innovative approach aims to enhance decision-making processes within the financial industry, offering a creative strategy to address market complexities.

What it does

Crypnosys stands at the forefront of technological innovation, integrating cutting-edge artificial intelligence to revolutionize how investors navigate the dynamic worlds of cryptocurrency and stock markets. Through advanced machine learning algorithms, Crypnosys offers invaluable predictions, enabling investors to anticipate market trends with precision and confidence. In the domain of ML + Accessibility, Crypnosys sets itself apart by not only delivering sophisticated predictive analytics but also ensuring accessibility for all users. Beyond predictions, Crypnosys provides an immersive experience, allowing investors to visualize live asset prices seamlessly. This comprehensive approach empowers investors of all backgrounds to make informed decisions, optimize their portfolios, and seize opportunities in the ever-changing landscape of financial markets.

How we built it

Collect and clean historical financial data: Gather and preprocess pertinent historical financial data. Design XGBoost model: Develop a tailored XGBoost architecture for accurate forecasting.

Train and validate: Partition the data, optimize parameters, and validate the XGBoost model for robust performance.

Evaluate performance: Assess model accuracy through metrics such as MSE, MAE, and RMSE.

Generate forecasts: Utilize XGBoost to predict asset prices for the next five days.

User Interface: Create a user-friendly interface to enhance user convenience and accessibility.

Challenges we ran into

Integrating a variety of data sources proved challenging, as did the delicate task of refining AI models to make accurate predictions amidst market volatility. Even in the process of fine-tuning, we experimented with numerous combinations before achieving satisfactory results. Initially, we employed an LSTM model, which demonstrated commendable accuracy at 97.5 percent, albeit at a slower pace. Subsequently, we transitioned to XGBoost, which not only enhanced accuracy but also significantly improved user experience due to its speed. Throughout this transition, we encountered numerous obstacles. Nonetheless, despite the challenges, we persevered and delivered our best effort.

Accomplishments that we're proud of

Achieving reliable predictions in a highly volatile market and creating a user-friendly interface for investors are our major accomplishments. Real time data showing in our website .

What we learned

We learned the intricacies of cryptocurrency markets, honed our AI modeling skills such as LSTM , XGBoost , and gained insights into user-centric design.

What's next for Crypnosys

In the future, we aim to enhance Crypnosys with advanced machine learning techniques, expand its coverage to more cryptocurrencies, and offer personalized investment strategies. Because there are some assets for which we cant fetch the data . So we will try to add that historical data .

Built With

• keras
• lstm
• matplotlib
• numpy
• pandas
• python
• streamlit
• tensorflow
• xgboost