Add Sonic.AI: Sonar Signal Classification Project (Mine vs. Rock)

[shirshanag]

This PR adds the Sonic.AI project, a machine learning model that classifies sonar signals as either "Rock" or "Mine", a task crucial in submarine object detection and naval defense systems.

🚀 Project Overview:
Dataset: Sonar signals dataset from UCI Machine Learning Repository

Goal: Predict whether the object detected by sonar is a metal mine or a rock

Algorithm: Logistic Regression (baseline model)

Evaluation Metric: Accuracy, Confusion Matrix

🧠 Key Features:
Preprocessing of 60-feature sonar frequency data

Binary classification using logistic regression

Train/test split and model validation

Confusion matrix visualization

CLI prediction script for custom input

📁 Files Added:
sonic_classifier.py — Model training and evaluation script

predict.py — Command-line prediction script

requirements.txt — Dependencies

README.md — Project overview and usage instructions

data/sonar.all-data.csv — Dataset file (or script to download if not committed)

✅ Future Improvements:
Explore advanced models like Random Forests or SVM

Deploy via Streamlit for interactive prediction

Add grid search and cross-validation