price_breakout / README.md
sovai's picture
Add price_breakout.parquet and README.md
307738d
---
icon: chart-line-up
description: >-
A dataset with daily updated predictions of price breaking upwards for US
Equities.
---
# Price Breakout
> **Data Notice**: This dataset provides academic research access with a 6-month data lag.
> For real-time data access, please visit [sov.ai](https://sov.ai) to subscribe.
> For market insights and additional subscription options, check out our newsletter at [blog.sov.ai](https://blog.sov.ai).
```python
from datasets import load_dataset
df_breakout = load_dataset("sovai/price_breakout", split="train").to_pandas().set_index(["ticker","date"])
```
Daily predictions arrive between 11 pm - 4 am before market open in the US for 13,000+ stocks.
`Tutorials` are the best documentation — [<mark style="color:blue;">`Price Breakout Prediction Tutorial`</mark>](https://colab.research.google.com/github/sovai-research/sovai-public/blob/main/notebooks/datasets/Breakout%20Prediction.ipynb)
<table data-column-title-hidden data-view="cards"><thead><tr><th>Category</th><th>Details</th></tr></thead><tbody><tr><td>Input Datasets</td><td>Historical Stock Prices, Trading Volumes, Technical Indicators, Order Book.</td></tr><tr><td>Models Used</td><td>Classification Algorithms, Regression Models, Conformal Predictors</td></tr><tr><td>Model Outputs</td><td>Price Movement Predictions, Probability Scores, Confidence Intervals</td></tr></tbody></table>
## Description
This datasets identifies potential price breakout stocks over the next 30-60 days for US Equities. This dataset provides daily predictions of upward price breakouts for over 13,000 US equities.
The accuracy is around 65% and ROC-AUC of 68%, it is one of the most accurate breakout models on the market. It is retrained on a weekly basis.
Several machine learning models are trained using the prepared dataset:
* **Calibrated Classifier**: A classification model trained on the engineered features to predict the binary target.
* **Proprietory Regressor**: A proprietory regression model is used to predict the probability of a price increase.
* **Conformal Regressor**: Used to provide calibrated confidence intervals around the predictions, offering an additional measure of uncertainty.
## Data Access
### Retrieving Data
#### Latest Data
```python
import sovai as sov
df_breakout = sov.data("breakout")
```
<figure><img src="https://raw.githubusercontent.com/sovai-research/sovai-documentation/main/.gitbook/assets/price_breakout_1 (2).png" alt=""><figcaption></figcaption></figure>
#### Full history
```python
import sovai as sov
df_breakout = sov.data("breakout", full_history=True)
```
#### Specific Ticker
```python
df_msft = sov.data("breakout", tickers=["MSFT"])
```
## Plots
### **Line Predictions**
```python
df_breakout.plot_line(tickers=["TSLA", "META", "NFLX"])
```
<figure><img src="https://raw.githubusercontent.com/sovai-research/sovai-documentation/main/.gitbook/assets/price_breakout_2 (2).png" alt=""><figcaption></figcaption></figure>
### Breakout Predictions
Visualize breakout predictions using the SDK's plotting capabilities:
```python
sov.plot("breakout", chart_type="predictions", df=df_msft)
```
<figure><img src="https://raw.githubusercontent.com/sovai-research/sovai-documentation/main/.gitbook/assets/price_breakout_3 (2).png" alt=""><figcaption></figcaption></figure>
### Prediction Accuracy
Assess the accuracy of breakout predictions:
```python
sov.plot("breakout", chart_type="accuracy", df=df_msft)
```
<figure><img src="https://raw.githubusercontent.com/sovai-research/sovai-documentation/main/.gitbook/assets/price_breakout_4 (2).png" alt=""><figcaption></figcaption></figure>
## Data Dictionary
<table><thead><tr><th width="237">Column</th><th>Description</th><th>Type</th><th>Example</th></tr></thead><tbody><tr><td><code>ticker</code></td><td>Stock ticker symbol.</td><td>object</td><td>"AAPL"</td></tr><tr><td><code>date</code></td><td>Date when the data was recorded.</td><td>datetime64[ns]</td><td>2023-09-30</td></tr><tr><td><code>target</code></td><td>Target variable for predictions.</td><td>float64</td><td>0.05</td></tr><tr><td><code>future_returns</code></td><td>Future returns of the stock.</td><td>float32</td><td>0.10</td></tr><tr><td><code>prediction</code></td><td>Predicted probability from the model.</td><td>float64</td><td>1.25</td></tr><tr><td><code>bottom_prediction</code></td><td>Lower bound of the prediction interval.</td><td>float64</td><td>1.20</td></tr><tr><td><code>top_prediction</code></td><td>Upper bound of the prediction interval.</td><td>float64</td><td>1.30</td></tr><tr><td><code>standard_deviation</code></td><td>Standard deviation of the predictions.</td><td>float64</td><td>0.02</td></tr><tr><td><code>bottom_conformal</code></td><td>Lower bound of the conformal prediction interval.</td><td>float64</td><td>1.18</td></tr><tr><td><code>top_conformal</code></td><td>Upper bound of the conformal prediction interval.</td><td>float64</td><td>1.32</td></tr><tr><td><code>slope</code></td><td>Slope derived from the rolling regression of predictions over a window.</td><td>float64</td><td>0.003</td></tr></tbody></table>
***
## Use Case
Understood. I'll focus on the use cases that would be most relevant to professional investors. Here's the refined list:
• Portfolio optimization:
* Identify potential new additions to diversified stock portfolios
* Rebalance existing holdings based on breakout predictions
• Risk management:
* Use confidence intervals and standard deviations to assess potential downside risk
* Implement more precise hedging strategies based on predicted price movements
• Sector and market analysis:
* Identify trends across industry sectors or the broader market
* Compare breakout potentials across different stock categories (e.g., large-cap vs. small-cap)
• Market timing:
* Use aggregate predictions across multiple stocks to gauge overall market sentiment
* Time entry and exit points for broader market positions