Skip to main content
. 2024 Jan 23;10:e1745. doi: 10.7717/peerj-cs.1745

Table 4. Performance of techniques evaluation in predicting helpful reviews.

Comparison of the correlation coefficient and SelectKBest scores based on the features.

Features Correlation coefficient scores with helpful SelectKBest scores with helpful Feature category
Rating 0.279878 1455.2 Metadata
Reviewer Location Existence 0.199417 405.4025 Metadata
Reviewer Name Existence 0.183377 316.0375 Metadata
Subjectivity 0.143916 276.8457 Subjectivity
Polarity 0.141497 311.0764 Polarity
Length 0.118247 112.8417 Linguistic
Adjective 0.117354 118.9454 Linguistic
IAV 0.11467 117.5278 Linguistic
Noun 0.108986 95.1823 Linguistic
Adverb 0.107725 96.45396 Linguistic
Sentences in a Review 0.107396 90.1899 Linguistic
Capitalized Words 0.105275 95.15608 Linguistic
SAV 0.09265 69.18848 Linguistic
ARI 0.083694 57.38236 Readability
RIX 0.080336 30.78826 Readability
LIX 0.076998 32.57184 Readability
FK 0.076388 31.86075 Readability
DCI 0.066897 34.63997 Readability
FKGL 0.060843 6.485602 Readability
SV 0.057572 36.28841 Linguistic
GFI 0.056262 5.642237 Readability
SMOGI 0.04897 49.93567 Readability
DAV 0.048794 28.46312 Linguistic
CLI 0.038327 34.69642 Readability
Reviewer Location Existence Validation 0.030635 10.22768 Metadata
Reviewer Name Existence Validation 0.022298 7.191998 Metadata
FRE −0.07262 36.80358 Readability

Notes.

The bold scores indicate the selected features in this research.