Pattern Learner

A lightweight Java library for detecting and predicting patterns in sequential data using simple machine learning algorithms. Enables developers to easily identify recurring patterns in time series data, user behavior sequences, or any ordered datasets without complex ML framework dependencies.

Features

• Sequential pattern mining with configurable window sizes
• Real-time pattern prediction with confidence scoring
• Lightweight implementation with no external ML dependencies

Installation

Maven

<dependency>
    <groupId>com.patternlearner</groupId>
    <artifactId>pattern-learner</artifactId>
    <version>1.0.0</version>
</dependency>

Gradle

implementation 'com.patternlearner:pattern-learner:1.0.0'

Usage

Basic Pattern Mining

import com.patternlearner.PatternLearner;
import com.patternlearner.model.Pattern;
import java.util.*;

// Create pattern learner with window size of 3
PatternLearner learner = new PatternLearner(3);

// Train with sequential data
List<String> sequence = Arrays.asList("A", "B", "C", "A", "B", "C", "D");
learner.train(sequence);

// Get discovered patterns
List<Pattern> patterns = learner.getPatterns();
patterns.forEach(System.out::println);

Real-time Prediction

// Predict next element with confidence
List<String> currentSequence = Arrays.asList("A", "B");
Prediction prediction = learner.predict(currentSequence);

System.out.println("Next element: " + prediction.getElement());
System.out.println("Confidence: " + prediction.getConfidence());

Requirements

• Java 8 or higher
• No external dependencies required

License

MIT License