📊 Anomaly Detection Software

Welcome to Anomaly Detection Software, a C/C++ based tool that helps you detect anomalies in numerical datasets using statistical methods. The software uses a straightforward statistical anomaly detection algorithm that can be integrated into any system for identifying outliers in a given dataset.

⚙️ Overview

This software implements a statistical approach for detecting anomalies based on the assumption that the data follows a normal distribution. It identifies outliers using the mean and standard deviation, flagging data points that are outside a configurable threshold.

✨ Key Features

• 🧠 Statistical Detection: Simple and effective detection using mean and standard deviation. Note: These simpler methods may not be useful for your use-case.
• 📊 Normal Distribution Assumption: Detects anomalies assuming the data follows a normal distribution.
• 💾 Memory-efficient: The algorithm processes the dataset in two passes to calculate accurate statistics while minimizing memory use. Note: to be more precise: O(n) time and space complexity...
• 💡 Modular: The software is split into modular components for preprocessing, detecting, and printing anomalies.

🛠️ How It Works

1. Input: The software takes an array of float values as input.
2. Preprocessing:
   • The dataset is first preprocessed (if necessary) to clean and normalize the data.
3. Detection:
   • It computes the mean and standard deviation of the dataset.
   • Any data point that lies outside a specified number of standard deviations from the mean (default is 2) is flagged as an anomaly.
4. Output: The software returns a 2D array where each row contains:
   • The original data point.
   • A flag (1.0 for anomaly, 0.0 for normal data point).

🧑‍💻 Code Breakdown

Core Files

1. detect_anomalies.c: Implements the core algorithm for anomaly detection.
   • Computes the mean and standard deviation in two passes.
   • Allocates memory for results and returns a 2D array with the original values and flags for anomalies.
2. extract_features.c: Handles any feature extraction from the dataset, preparing the data for anomaly detection.
3. preprocess_data.cpp: Prepares and normalizes the dataset before feeding it into the anomaly detection algorithm.
4. print_anomalies.cpp: A utility to print the detected anomalies to the console and store them in a structured format.