Information Technology Reference
In-Depth Information
•
In
step
3
EA-Miner searches for words, that were not found in the previous step,
and whose semantic category
3
belongs to semantic classes to which several
NFRs belong (e.g., Permission related category such as S7.4, Time related
category such as T1.1.2). Moreover, some NFRs can also be identified by
analyzing the suffix of the words (e.g., words that end in “ilities” such as
availability and “ness” such as correctness). At the end of this step the list of
candidates identified is added to the list from step 2 and shown to the user.
•
In
step
4
the user has the opportunity to see the list of candidates and then
update the actual lexicon. The semantic tag enables the user to make an
informed decision if it is appropriate to include that word in the lexicon (in
Sect. 3.1.1 we explained that some words might have multiple meanings such as
performance).
The process is repeated every time a new file is mined. From our experience the
larger the lexicon gets the need to update it becomes rare. For the latest requirements
documents on which we have used the tool recently, just two or three entries were
added.
2
2
1
1
EA-Miner
EA-Miner
Search NFR Knowledge Base
Search NFR Knowledge Base
Requirements
documents
Requirements
documents
3
3
Analyze semantic class
Analyze suffix of word
Analyze semantic class
Analyze suffix of word
4
4
List of New Lexicon Entries:
List of New Lexicon Entries:
Athorised SEM = S7.4+ Meaning: Permission
…
Athorised SEM = S7.4+ Meaning: Permission
…
Fig. 9.
Lexicon update process
Regarding the analysis of functional early aspects for the viewpoint-based model,
EA-Miner utilizes the Fan-in based aspect mining code analysis technique [36] as
explained before in Sect. 3.1.1. For each action verb word its fan-in is analyzed
related to the viewpoints that make a reference to the word. If the word has a high
fan-in (e.g., fan-in ≥ 4) then it is considered a candidate functional early aspect.
4.1.4 Crosscutting Relationship Identification (Activity 2.4)
As mentioned previously, the collection of requirements of viewpoints and early
aspects are important for identifying the relationships. Each requirements sentence is
understood as a join point and the collections of requirements are compared based on
a set intersection operation. If the resulting set is empty it means that there is no
crosscutting relationship between the viewpoint and the early aspect. If the resulting
3
Remember that at this stage the file has already been tagged by WMATRIX and the semantic
tag for each word has already been allocated.
Search WWH ::
Custom Search