malfusion / document-clustering

A Java experiment on clustering of textual data, dimensionality reduction and visualization

Textual Data Clustering In Java

A Java experiment on clustering of textual data, dimensionality reduction and visualization

Running the project

• Import the project into Eclipse and install maven dependencies. There are no external JAR dependencies or references.
  • You can do so by creating a new workspace and then File->Open Projects from File System, point to this folder.
  • Run App.java which is the entrypoint.

About

• The program will output a topics.txt within the same folder for viewing.
• The plot for Original vs (K-Means++ / Cosine Distance) Predictions are also present in the folder for reference.
• The output of the program is detailed and gives all the required details of accuracy, etc.
• Running the program with different Enum values for DistanceMeasure and CentroidInitializer will result in different results in the SVD and PCA visualizations.
• The program achieves the highest accuracies (sometimes 100%) with K-means++ and Cosine distance.