Please enable JavaScript.
Coggle requires JavaScript to display documents.
UnityModel, Different models used, Balancing techniques Attempted - Coggle…
UnityModel
Feature Selection/Engineering
All Features
Balancing
TF.Class_weights
Modeling
Simple NN
AUC=0.57
Complex CNN
aUC=0.52
Manual (under/SMOTE)
Modeling
Simple NN
AUC=0.54
Complex CNN
aUC=0.51
Azure AutoML : AUC=0.67
Not All
PCA, Spearman, and Chi-square
Drop (ViewCount, DeviceType,timeStamp)
Encoding (country,software)
Encoding(onehot encoding timestamp
Balancing
TF.Class_weights
Modeling
2 more items...
Manual (under/SMOTE)
Modeling
2 more items...
labelencoding timestamp
Balancing
TF.Class_weights
Modeling
2 more items...
Dropped timestamp
Balancing
TF.Class_weights
Modeling
2 more items...
One_hot encoding timeStamp
log device type
Balancing
TF.Class_weights
Modeling
Simple NN
1 more item...
Deep_NN
1 more item...
Log timeStamp
Group device type
Balancing
TF.Class_weights
Modeling
Simple NN
1 more item...
Deep_NN
1 more item...
One_hot encoding timeStamp
Label Encode device type, and country and software
Balancing
Under_Sampling
Modeling
Simple NN
1 more item...
Deep_NN
1 more item...
Different models used
Autoencoders
CNN
LogisticRegression
SVM
DNN
Balancing techniques Attempted
Random oversampling
SMOTE
TF Class weighting
UnderSampling