Welcome to our prediction engine — a data-driven tool designed to analyze historical stock data and generate insight into potential future trends. Here’s a breakdown of how it works and what the results on our Stock Predict page represent.
📊 The Data
We analyze the most recent 3 years of daily stock data for each ticker. We include a select number (~2000) of ticker symbols derived largely from medium-cap stocks and up. Our models incorporate technical indicators such as:
- RSI (Relative Strength Index)
- Moving Averages
- Price Momentum
- Volume Trends
- And other market-derived signals
For every prediction, the dataset is split into:
- 80% Training Data – Used to teach the model
- 20% Testing Data – Used to evaluate model performance on unseen data
🔁 Two Types of Predictions
Our system generates two kinds of predictions per stock:
1. Binary Trend Prediction (Up / Down)
- Predicts whether the stock is likely to go up or down.
- Uses classification models.
- Evaluated using the Weighted F1 Score.
🧠 What is Weighted F1 Score?
The F1 score is a balanced metric that combines precision (how accurate the “up” or “down” calls are) with recall (how many of the actual “ups” or “downs” we correctly catch). The weighted version adjusts for imbalance in class distribution.
- 1.0 = Perfect accuracy
- 0.5 = Similar to random guessing
- 0.0 = No predictive power
This gives a more realistic measure of prediction quality in the real world.
2. Numeric Price Prediction (Regression)
- Forecasts the actual price level the stock may reach.
- Uses regression algorithms (e.g., linear, tree-based, ensemble).
- Evaluated with the R² (R-squared) Score.
📈 What is R² Score?
R² tells us how well the model explains variation in the stock’s price.
- 1.0 = Perfect prediction
- 0.0 = No better than guessing the average
- < 0.0 = Worse than guessing the mean
Higher R² values suggest the model has identified a meaningful pattern in the data.
📋 What the Results Mean
On the Stock Predict page, you’ll see:
- ✅ Prediction (Up/Down) – Our trend direction forecast
- 📉 Predicted Price – Our estimated numeric price
- 🔎 Accuracy Scores:
- F1 Score – For binary trend classification
- R² Score – For regression-based price prediction
These metrics are generated based on the model’s performance against a held-out 20% test set — data the model has never seen during training.
⚠️ Important Disclaimer
The predictions and data on this site are provided for informational and educational purposes only. They do not constitute financial, investment, or trading advice.
Stock markets are inherently unpredictable and influenced by countless external factors. While we strive to provide high-quality insights based on historical patterns, no model can guarantee accuracy or future results.
By using this site, you acknowledge and agree that:
- You are solely responsible for any investment or trading decisions.
- We are not liable for any financial losses or damages arising from the use of the content or predictions on this site.
Always consult with a licensed financial advisor before making any investment decisions.
🚀 Stay Informed, Not Misled
Our goal is to empower you with data-driven tools, not promise financial outcomes. Use our models as one of many inputs in your decision-making process — not as a substitute for independent research or professional guidance.
Thanks for being here, and stay smart with your trades.