Feature values are defined as in theoretical features (see Equation (1)). The ten features with the overall highest mean value in our data (among a total of 36 features) are listed in Table 4. Note that the mean values are much higher for the POS features (Table 4) than for the theoretical features (Table 3), as theoretical features are much sparser.
POS features. The mean and SD values are computed on all clustered adjectives.
| Feature . | Mean . | SD . | Feature . | Mean . | SD . |
|---|---|---|---|---|---|
| −1 noun | 0.52 | 0.25 | −2 preposition | 0.13 | 0.09 |
| +1 punctuation | 0.42 | 0.15 | −1 adverb | 0.10 | 0.11 |
| −2 determiner | 0.39 | 0.20 | −1 verb | 0.08 | 0.11 |
| +2 determiner | 0.24 | 0.13 | −1 determiner | 0.06 | 0.10 |
| +1 preposition | 0.21 | 0.15 | +1 noun | 0.06 | 0.10 |
| Feature . | Mean . | SD . | Feature . | Mean . | SD . |
|---|---|---|---|---|---|
| −1 noun | 0.52 | 0.25 | −2 preposition | 0.13 | 0.09 |
| +1 punctuation | 0.42 | 0.15 | −1 adverb | 0.10 | 0.11 |
| −2 determiner | 0.39 | 0.20 | −1 verb | 0.08 | 0.11 |
| +2 determiner | 0.24 | 0.13 | −1 determiner | 0.06 | 0.10 |
| +1 preposition | 0.21 | 0.15 | +1 noun | 0.06 | 0.10 |