Skip to main content

Advances, Systems and Applications

Table 2 Prediction results of deep learning models on Yonsei and Citi Bike datasets

From: Real-time trajectory privacy protection based on improved differential privacy method and deep learning model

Model

Yonsei

Citi Bike

MAE

MSE

BCE

MAE

MSE

BCE

CNN

0.00508

0.00156

0.01831

0.00018

0.000034

0.00048

LSTM

0.00346

0.00158

0.01491

0.00016

0.000037

0.00016

ConvLSTM

0.00293

0.00147

0.00978

0.00014

0.000031

0.00012

ST-LSTM

0.00286

0.00138

0.00901

0.00012

0.000029

0.00011

T-GCN

0.00271

0.00122

0.00885

0.00011

0.000023

0.00009