Skip to main content

Table 1 The statistical information of all datasets

From: Attention-based Sentiment Reasoner for aspect-based sentiment analysis

Datasets Overall Pos. Neg. Neu. Len.
Restaurant (2014)
 Train 3578 2148 800 630 78
 Test 1110 721 195 194
Laptop (2014)
 Train 2317 991 865 459 82
 Test 637 341 128 168
Restaurant (2015)
 Train 1279 963 280 36 74
 Test 597 353 207 37
Restaurant (2016)
 Train 1880 1319 489 72 78
 Test 650 483 135 32
Camera
 Train 1635 1157 478 26
 Test 408 263 145
Car
 Train 885 668 217 34
 Test 221 166 55
Notebook
 Train 485 331 154 18
 Test 121 72 49
Phone
 Train 1885 1260 625 39
 Test 471 303 168
  1. We use “–” to indicate that the dataset doesn’t contain this kind of class. We use ‘Len.’ to represent the max length of the sentence. We respectively use “2014”, “2015” and “2016” to represent that the dataset is belong to SemEval 2014, 2015 and 2016