Importing, preprocessing, and clustering of textual data

The goal of this workflow is to cluster a set of newsgroup documents into their corresponding topic. The data is taken from the 20 newsgroups dataset. The workflow starts with a data table containing some newsgroup documents, divided into two categories, politics.guns and sport.baseball.

First, the data are converted into documents, whose category is the class politics or sport. The documents are then preprocessed by filtering and lemmatizing. After that, the documents are transformed into a bag of words, which is filtered again. Only terms that occur at least in 1% of the documents (at least in 2 documents) will be used as features and not be filtered out. Then the documents are transformed into document vectors.

The document vectors are a numerical representation of documents and are in the following used for hierarchical clustering based on Manhattan, Euclidean, and Cosine distance measures.

External resources

• Newsgroup dataset