Skip to main content
. 2017 Mar 19;17(3):631. doi: 10.3390/s17030631

Table 9.

Prediction results of SVR model using our features for one_per_user and Personal_collection settings.

One_Per_User Personal_Collection
Features Views Interaction Views Interaction
Low_Level BoVW 0.029 0.036 0.211 0.143
Color 0.01 0.018 0.249 0.071
Gist 0.031 0.024 0.252 0.142
LBP 0.078 0.1003 0.329 0.228
Middle _Level Attribute 0.056 0.059 0.308 0.163
Classemes 0.077 0.076 0.269 0.172
SentiBank 0.089 0.099 0.381 0.175
Deep Learning FC7 0.059 0.074 0.349 0.246
FC8 0.115 0.137 0.207 0.103
Aesthetic 0.064 0.068 0.328 0.231
Contextual features 0.381 0.503 0.323 0.263
Textual feature TF 0.184 0.215 0.709 0.492
TF-IDF 0.177 0.202 0.629 0.416