OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Data & Analytics

Alphagen

Alphagen is a Python research codebase for reinforcement-learning-based generation and evaluation of formulaic predictive stock factors.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Alphagen — OpenViST project profile in AI Data & Analytics.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it for research workflows that generate and assess stock-factor candidates, particularly when you can prepare the required stock data or implement the project’s calculator interface.

Who it is for

  • Quantitative-finance researchers exploring automated factor discovery.
  • Developers reproducing alpha-generation experiments or adapting the factor calculator to another pipeline.

Documented uses

  • Generate candidate formulaic stock factors using reinforcement learning.
  • Evaluate individual factors and weighted alpha pools with IC and Rank IC measures.
  • Compare the reinforcement-learning approach with included genetic-programming and deep-symbolic-regression baselines.
  • Explore iterative LLM-based alpha-generation routines included in the repository.
Before you choose

Requirements & limits

  • The built-in calculation pipeline requires Qlib and local stock data; its preparation instructions use Qlib metadata and data retrieved from baostock.
  • An external data or calculation workflow requires an adapter implementing the AlphaCalculator interface, including factor and pool IC and Rank IC calculations.
  • The README recommends normalizing factors before combining them because their evaluated values may have substantially different scales.
  • The repository includes LLM-generation routines and modified gplearn and DSO baselines in addition to its reinforcement-learning experiments.
  • Trading strategies appear as experimental Qlib-based demos, distinct from the central factor-generation workflow.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare stock data for AlphaGen’s built-in Qlib pipeline

Prepare Qlib metadata and local Baostock stock data for use with AlphaGen’s built-in pipeline.

Before you start

  • Qlib library and its metadata, prepared using Qlib’s data-preparation process.
  • Access to Baostock stock data.

Steps

  1. Follow Qlib’s data-preparation process to obtain the required metadata.

  2. Run data_collection/fetch_baostock_data.py to download the stock data.

  3. If you customize the destination, pass the matching path to Qlib in StockData._init_qlib using qlib.init(provider_uri=path). The default destination is ~/.qlib/qlib_data/cn_data_baostock_fwdadj.

Expected outcome

Qlib metadata and locally stored Baostock stock data are prepared for the built-in alpha-calculation pipeline.

Guide limitations

  • The built-in pipeline requires Qlib and locally stored stock data.
  • Qlib metadata is required, but the README says to use Baostock for the actual stock-price and volume data.
  • This guide covers data preparation, not running experiments.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
ICT-FinD-Lab/alphagen · Official homepage title · retained in its source language
How we review sourcesSuggest a correction