Abstract
A Python data/ML experiment for Binance and stock price prediction, structured as a repeatable pipeline for iterative model comparison. The project is archived and kept deliberately separate from the public Binance trading-bot application.
1. What This Is
This was a 2026 project-set entry focused on data and ML automation for market prediction. The work involved preparing market and stock data into a pipeline suitable for repeated ML experiments, defining prediction targets, and running model comparisons during development.
It is distinct from the separate public binance Flask trading-bot (RSI/MACD indicators, exchange automation). The scopes differ: this project is prediction research; the bot is automated order execution. Keeping them separate prevents conflating the two.
2. How It Works
The workflow followed a standard ML experimentation loop: load and prepare the market dataset, clean and transform it into model-ready inputs, define the prediction target and experiment configuration, train or run the selected model, then evaluate predictions and iterate on the pipeline. No specific algorithm, target horizon, or accuracy figure is documented in the source material, so the page intentionally avoids inventing those details.
3. Constraints
-
No public code artifact
No matching repository was found in the reviewed GitHub account. The experiment exists only as a portfolio-listed entry with no inspectable source.
-
Undocumented model and metrics
The source does not specify which algorithm was used, what the prediction target was, or what evaluation metrics were applied. No accuracy or performance claim is made.
-
Closed and archived
The project is marked CLOSED in the 2026 project set. No further development is planned without a concrete restart scope.
-
Scope bounded by source evidence
All claims on this page are limited to what the portfolio PDF confirms. No technologies, results, or architecture details are inferred beyond that.
4. Next
- a. Pin down a concrete prediction target (e.g. next-day direction) and a baseline model before any further iteration.
- b. Publish the pipeline code to a dedicated repository so the experiment is reproducible and inspectable.
- c. Define evaluation metrics (directional accuracy, Sharpe ratio, etc.) and a backtesting protocol before comparing models.
— end of report —