Wassap-ML-KM (Machine Learning - K-Means)

Wassap is a university project on Artificial Intelligence subject.
It is basically a text summarizer built using Python and NLTK as its library.
Code is a bit messy and under-documented.

Method

We use this paper as a reference on how to do K-Means algorithm. We recommend you to read the paper to understand how K-Means algorithm works. Searching for a visualizer will be helpful too.

Configurations

There are 7 settings that we can change in config.txt file:

• docsNum define how many from 1 to N documents you want to summarize. Documents are in the format of docsN.txt and are expected to be inside data folder.
• idkwgo basically will print out all of the process going on.
• sleepInterval will pause after each process so we can read for a moment.
• genSim we provide a simulator to see visually how the program works. When set to true simN.txt will be generated in sim folder.
• hashNum this is the number of characters from the start of a sentence that we use to store the sentence itself. Low numbers mean an overlap will likely to happen, high numbers may introduce error when sentence length is shorter.
• outputSum
  • 1 will show the final summary right on the console.
  • 2 will save the summary to a file sumN.txt in sum folder.
  • 3 will do both what 1 and 2 do.
• useGlobalIDF
  • false will make Wassap score words per document.
  • true will make Wassap score words by using all documents.