OpenGradient Model Highlight: SUI/USDT Short-Term Spot Forecasting Model A SUI/USDT model designed to predict short-horizon spot price movements using structured signals from market data. 🧵👇🏻

2/ In short-term markets, direction matters more than magnitude. We trained models across 30-minute and 6-hour horizons to estimate forward returns using structured market data. Each model produces a signal indicating expected movement over its timeframe.

3/ The model takes recent OHLCV candle data and performs classical feature-engineering for model inputs. We then apply Lasso-based feature selection to retain only signals that continue to show predictive relevance. This also prevents model overfitting.

4/ The model outputs a forecast return for the next interval, which can be mapped into simple execution logic: • Positive forecast → Long • Negative forecast → Short A simple backtest illustrate how this signal behaves when applied systematically over time.

5/ The two spot forecasting models demonstrate strong predictive capability by training on, identifying, and predicting market patterns.