Identifying shared functionality using latent semantic indexing

Identifying shared functionality using latent semantic indexing

Learbuch, D.L.R.

Gross, H.G. (mentor)
Gerbrandy, J. (mentor)

Getronics PinkRoccade

Electrical Engineering, Mathematics and Computer Science

2008-09-04

There are numerous systems of which parts, at least in the minds of designers, have the same functionality. If the degree of this shared functionality can be measured some how, it would prove to be very valuable. This information can be used to modularize these systems and improve their maintainability. To identify this functionality a technique known as Latent Semantic Indexing (LSI), an information retrieval technique, is proposed. We applied LSI to two systems at Getronics PinkRoccade. Our main conclusion is that we can not recover 100% of the shared functionality automatically, however, the search space can be reduced. This means the experts, recovering the shared functionality, have less documents to evaluate.

lsi
latent semantic indexing
shared functionality

http://resolver.tudelft.nl/uuid:e74c11d4-b9be-4b89-ae5b-b7e5bc853717

TU Delft, Electrical Engineering, Mathematics and Computer Science, Information and Communication Technology (ICT)

Student theses

master thesis