Table 8 ACP extraction from natural language artifacts in the related work

ACP sentence identification

Semantic patterns matching

N/A

iTrust, IBMApp

927

Prec:88.7% Rec:89.4%

ACP elements extraction

Heuristics over the patterns

N/A

Access control (AC) sentences in: iTrust, IBM App, and collected ACP

241

Accu:86.3%

Transformation to formal model

Heuristics

N/A

N/R

N/R

N/R

ACP sentence identification

Majority vote of KNN, Naive Bayes (NB) and SVM classifiers

Words,synonyms,POS, named entities, & Levenshtein distance in the case of KNN

iTrust

1159

Prec:87.3% Rec:90.8%

ACP elements extraction

RE using bootstrapping; seeding patterns are derived from dependency tree

N/A

AC sentences in: iTrust

409

Prec:46.3% Rec:53.6%

NB classifier of candidate instances

were not clearly reported

ACP sentence identification

KNN

Levenshtein distance

iTrust,IBM App, Cyberchair, collected ACP

2477

Prec:81% Rec:65%

ACP elements extraction

RE using bootstrapping; seeding patterns are derived from dependency tree

N/A

AC sentences in: iTrust,IBM App, Cyberchair, collected ACP

1390

N/R

NB classifier of candidate instances

Pattern itself, relationships to resource and subject, POS of subject and resource

ACP sentence identification

NB and SVM classifiers

A total of 821 features categorized into: pointwise mutual information, security,syntactic complexity, and dependency features

Trust, IBM App, Cyberchair, collected ACP

2477

Prec:90% Rec:90%

ACP sentence identification

Deep recurrent neural network

Words embeddings

ACPData, iTrust,IBM App, Cyberchair, collected ACP

5137

Prec:81.28% Rec:74.21%

ACP elements extraction

SRL

N/A

AC sentences in: iTrust, IBM App, Cyberchair, collected ACP

726

Prec:58.3% Rec:86.3%

ACP elements extraction

SRL

N/A

AC sentences in: iTrust, IBM App, Cyberchair, collected ACP

841

Prec:63.5% Rec:86.25%