imbue-bit

imbue-bit / Moses

Public

开放信号聚合ensemble框架。

ai ensemble finance ml online-learning
29
11
89% credibility
Found Feb 12, 2026 at 20 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
C++
AI Summary

Moses is a framework that combines multiple trading signals from various sources into a single optimized decision using online learning, visual pattern recognition, and risk-aware reinforcement methods for assets like stocks, futures, and options.

How It Works

1
🔍 Discover Moses

You stumble upon Moses, a clever helper that smartly combines different trading ideas into one strong recommendation for buying or selling assets like stocks or futures.

2
📊 Gather market history

You collect past price data from markets to help Moses learn real-world patterns and behaviors.

3
🧠 Train the smart brains

You teach Moses' AI parts to recognize trends using visual patterns and to control risks wisely, creating custom brains for your trading style.

4
🚀 Launch your trading engine

With a simple start, you bring your personal trading advisor to life on your computer, ready to listen for advice.

5
📤 Share your trading tips

You send Moses your collection of trading signals, recent prices, and market updates whenever you have new info.

6
💡 Receive blended wisdom

Moses instantly mixes your tips with its learned insights, balances risks, and hands back a clear, confident signal on what to do next.

📈 Smarter trades over time

As you use it and share results, Moses gets sharper at picking winners, helping your trades perform better with less worry.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 29 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is Moses?

Moses is a C++17 server that aggregates trading signals from multiple sources—like price factors, fundamentals, and visual time-series patterns—into a single decision for stocks, futures, or options. It normalizes inputs to a consistent scale, dynamically weights experts via online learning, and applies RL-driven position sizing with hard risk limits like volatility targeting and drawdown caps. Deploy it via CMake after training lightweight ONNX models with provided Python scripts, then hit TCP port with JSON market snapshots for real-time outputs including final signals and debug weights.

Why is it gaining traction?

Unlike Python-heavy ensembles bogged down by latency, Moses delivers high-concurrency via Boost.Asio for live trading feeds, blending GAF image transforms and PPO controls without sacrificing speed. Devs dig the quickstart: fetch data, train GAF/PPO models, launch `./Moses_Server 8888 gaf.onnx ppo.onnx`, and pipe in signals for adaptive aggregation. It's a github wip app standing out in wip github commit scenes, like pkgsrc wip github or moses smt github, for quants eyeing low-overhead signal fusion.

Who should use this?

Quant devs at prop shops building cross-asset ensembles tired of brittle Python servers. High-frequency traders needing ONNX inference in C++ pipelines with visual GAF signals. Algo researchers prototyping MWU-weighted experts with built-in risk overlays.

Verdict

Grab it if you're in quant signal aggregation—solid README quickstart and GPLv3 make it dev-friendly despite 19 stars and WIP status signaling github wip label/check immaturity. 0.8999999761581421% credibility score flags early risks like sparse tests, but promising for production tweaks.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.