ai-ml-clustering
ai-ml-clustering copied to clipboard
azampagl
Implementation of multiple clustering algorithms (K-means, Bisecting K-means, Agglomerative Hierarchial Clustering with Intra-Cluster Similarity (IST), Centroid Similarity (CST), and UPGMA) for perfor...
A Comparison of Document Clustering Techniques for performance comparisons.
Original algortihms are based on:
Michael Steinbach, George Karypis, Vipin Kumar Department of Computer Science and Egineering, University of Minnesota Technical Report #00-034 {steinbac, karypis, kumar}@cs.umn.edu @see http://www.cs.fit.edu/~pkc/classes/ml-internet/papers/steinbach00tr.pdf
The style guide follows the strict python PEP 8 guidelines. @see http://www.python.org/dev/peps/pep-0008/
"Modules".
- Preprocess: occurs in the init method.
- Cluster: occurs via the classes "execute" method.
- Evaluate: occurs during the classes "evaluate" method.
============================================ Arguments for python main.py
The following are arguments required:
-t: the topic file. -a: the clustering algorithm (agg-upgma-k-means, agg-cst, k-means, bi-k-means-size, agg-ist, bi-k-means-sim, agg-upgma). -k: the number of clusters -o: the TFIDF file. -r: the result file.
The following arguments are required for bisecting k-means algorithms:
-i: number of iterations.
============================================ Execution
Execution is straightforward. After choosing a topic file (-t), a clustering algorithm (-a), and the number of clusters (-k), the program will spit out the TFIDF vectors for each document (-o) and the results (-r).
====================== Usage
The following are some example use cases.
python main.py -t "../data/toy/toy-topics.txt" -a "k-means" -k 3 -o "../results/toy/tfidf.dat" -r "../results/toy/k-means.txt"
azampagl