Implementation Details

The implementation for our approach leverages several mature technologies to realise different steps in our approach.

The first step, Identifying Candidate Terms, has been implemented using GATE NLP Workbench and the GATE's scripting language.
The second step, Calculate Similarity between Candidate terms, has been implemented using:
- Syntactic Similarity Calculation:: a JAVA package for string similarity calculation, i.e, Simpack.
- Semantic Similarity Calculation:: Approach followed in [1] using Wordnet::Similarity package.
  [1] Nejati, S.; Sabetzadeh, M.; Chechik, M.; Easterbrook, S.; Zave, P., "Matching and Merging of Variant Feature Specifications," Software Engineering, IEEE Transactions on , vol.38, no.6, pp.1355,1375, Nov.-Dec. 2012
The third step, Clustering, has been implemented using the EM algorithm in mclust package in R statistical programming language.

Fig: Overview of implementation for our approach

Improving Requirements Glossary Construction via Clustering:
Approach and Industrial Case Studies