Text classifier 2

Service for text classification using several classifier algorithms including neural networks, naive bayes etc. Source of inspiration was taken from https://github.com/RusZ/TextClassifier.

Currently project is being actively developed.

Project use Gradle 4.2 build platform with 2 submodules:

• textclassifier2_core - almost pure java app incapsulates core logic layer
• textclassifier_service - spring boot 2.0. async app incapsulates service layer and infrastructure (storage) layer

Requirements

• Java SE Development Kit 8 (jdk-1.8)
• Intellij 2017 (recommended) with lombock plugin
• MongoDB ('v.3.6.0')

Dependencies

• Encog Machine Learning Framework (org.encog:encog-core:3.3.0)
• Apache POI (org.apache.poi:poi-ooxml:3.16)
• Spring ('5.0')
• JUnit (junit:junit:4.12)

Quick start guide

1. When you compile and launch textclassifier_service application first time, it will start Netty web server on 127.0.0 .1:8080 and deploy REST interface. For API description look at sources

2. When you compile and launch textclassifier_Core it will automatically read xsl file texts for testing inside project:

After that application will build vocabulary, create and train neural network for each Characteristic.

Author

• ripreal

License

This project is licensed under the MIT License - see the LICENSE file for details.

TODO

• Mention about DDD paradygme in description
• Transfer network learning to Apache Ignite platform or S3 Cloud

Text Classifier