keneandita/sentrix repository overview

⁠Project Summary

This repository contains a containerized version of SentriX, a cryptocurrency intelligence platform that combines machine learning models to analyze and predict market movements. SentriX uses LSTM-based models for price forecasting and Isolation Forest for anomaly detection, enabling users to identify unusual trading behaviors in major cryptocurrencies such as Bitcoin, Ethereum, Cardano, and Ripple.

The platform provides an interactive Streamlit dashboard that visualizes predicted price movements, market anomalies, and trends in a clean and accessible interface. By running the application in a Docker container, users can deploy the platform quickly without manual dependency installation or configuration.

SentriX is designed for experimentation with cryptocurrency price prediction, anomaly detection, and interactive data visualization in a self-contained environment.

⁠Architecture Overview

SentriX follows a modular architecture composed of a machine learning layer, a preprocessing pipeline, and a dashboard interface:

The data preprocessing module handles historical market data, including scaling, sequence generation, and feature extraction. This ensures that input data is compatible with both the LSTM price prediction model and the anomaly detection algorithm.

The prediction layer uses an LSTM network built with TensorFlow to forecast the next-day closing prices of supported cryptocurrencies. Parallelly, an Isolation Forest model analyzes volume and spread patterns to detect abnormal market behaviors.

The dashboard interface, built with Streamlit, provides interactive visualizations of predicted prices, anomaly alerts, and historical trends. Users can explore insights in real-time without leaving the application environment.

The Docker container environment packages the application, Python runtime, machine learning libraries, and dashboard dependencies, providing a consistent and isolated runtime for seamless deployment.

⁠Key Features

Next-Day Price Prediction

Forecasts the following day’s closing prices for major cryptocurrencies using LSTM sequence models.

Anomaly Detection

Identifies unusual market activity such as abnormal volumes and price spreads using Isolation Forest.

Historical Data Preprocessing

Processes and scales cryptocurrency historical data to prepare it for model inference.

Interactive Dashboard

Provides visual insights on predictions and anomalies through a Streamlit interface.

Machine Learning Integration

Combines TensorFlow LSTM models and Scikit-learn Isolation Forest models in one platform.

Containerized Deployment

The Docker container encapsulates all dependencies, ensuring the platform can run consistently across systems without manual setup.

⁠How to Run

First, pull the Docker image.

docker pull keneandita/sentrix

Run the container while exposing the dashboard port.

docker run -p 8501:8501 keneandita/sentrix

Once the container is running, open the Streamlit dashboard in a browser:

http://localhost:8501

The dashboard allows users to view cryptocurrency price predictions, detect anomalies, and interactively explore historical and forecasted market data.