4.3 Anomaly Detection

Anomalies

Techniques for Anomaly Detection

1. Statistical Methods: Statistical techniques like mean, standard deviation, and probability distributions to detect anomalies.

2. Machine Learning Methods:
1. Isolation Forest: Uses decision trees to isolate anomalies by recursively partitioning the data.
2. Local Outlier Factor (LOF): Compares the local density of a point to its neighbors. Points with significantly lower density are considered anomalies.
3. One-Class Support Vector Machine (SVM): A regular support vector machine algorithm tries to find a hyperplane that best separates the two classes of data points. In an SVM that has one class of data points, the task is to predict a hypersphere that separates the cluster of data points from the anomalies.

Practical Application of Anomaly Detection

• Cybersecurity: Identifying unusual network traffic indicating potential attacks.

• Finance: Detecting fraudulent transactions or abnormal market behaviors.

• Healthcare: Monitoring patient data to spot irregularities that may indicate health issues.

• Manufacturing: Predicting equipment failures by monitoring sensor data.

Practice Problem

Instructions:

1. Input: A list of integers representing scores.

2. Output: Print each score that qualifies as an anomaly.

3. Criteria: An anomaly is a score that differs from the average score by more than a specified threshold value.

Example:

[85, 90, 82, 88, 92, 95, 75, 98, 80, 70]

75
98
70

Explanation:

• Average score:

• Threshold: ±10 points from the average (75.5 to 95.5)

• Anomalies: Scores 75, 98, and 70 fall outside this range.

Code:

import java.util.ArrayList;
import java.util.List;

public class AnomalyDetection {

    public static void main(String[] args) {
        int[] scores = {85, 90, 82, 88, 92, 95, 75, 98, 80, 70};
        
        detectAnomalies(scores);
    }

    public static void detectAnomalies(int[] scores) {
        // Calculate average score
        double sum = 0;
        for (int score : scores) {
            sum += score;
        }
        double average = sum / scores.length;

        // Define threshold (e.g., ±10 points)
        double threshold = 10.0;

        // Detect anomalies
        List<Integer> anomalies = new ArrayList<>();
        for (int score : scores) {
            if (Math.abs(score - average) > threshold) {
                anomalies.add(score);
            }
        }

        // Print anomalies
        System.out.println("Anomalies detected:");
        for (int anomaly : anomalies) {
            System.out.println(anomaly);
        }
    }
}

Explanation:

• detectAnomalies method: Calculates the average score from the input array, defines a threshold, and identifies scores that deviate from the average by more than the threshold. Anomalies are stored in a list and printed at the end.

About Me

Previous Section