NNs predicting Diabetes and propensity to it (TF.JS)
Train
Input Layer:
None
tanh
elu
hardSigmoid
linear
relu
relu6
selu
sigmoid
softmax
softplus
softsign
glorotNormal
None
constant
glorotUniform
heNormal
identity
leCunNormal
ones
orthogonal
randomNormal
randomUniform
truncatedNormal
varianceScaling
zeros
Bias:
None
glorotNormal
constant
glorotUniform
heNormal
identity
leCunNormal
ones
orthogonal
randomNormal
randomUniform
truncatedNormal
varianceScaling
zeros
Internal Layers:
Size:
Activation and Initializer:
tanh
None
elu
hardSigmoid
linear
relu
relu6
selu
sigmoid
softmax
softplus
softsign
glorotNormal
None
constant
glorotUniform
heNormal
identity
leCunNormal
ones
orthogonal
randomNormal
randomUniform
truncatedNormal
varianceScaling
zeros
Bias:
None
glorotNormal
constant
glorotUniform
heNormal
identity
leCunNormal
ones
orthogonal
randomNormal
randomUniform
truncatedNormal
varianceScaling
zeros
Output Layers:
tanh
None
elu
hardSigmoid
linear
relu
relu6
selu
sigmoid
softmax
softplus
softsign
None
glorotNormal
constant
glorotUniform
heNormal
identity
leCunNormal
ones
orthogonal
randomNormal
randomUniform
truncatedNormal
varianceScaling
zeros
Bias:
None
glorotNormal
constant
glorotUniform
heNormal
identity
leCunNormal
ones
orthogonal
randomNormal
randomUniform
truncatedNormal
varianceScaling
zeros
Loss function:
meanSquaredLogarithmicError
meanSquaredError
logcosh
meanAbsoluteError
meanAbsolutePercentageError
cosineProximity
squaredHinge
hinge
kullbackLeiblerDivergence
Optimizer:
adamax
adagrad
adam
sgd
adadelta
rmsprop
Batch Size:
Train Iterations:
Data imputation type:
k-NN
Median imputation
Train Model
Stop Model Training
Predict
Pregnancies:
<
Glucose:
<
Blood Pressure (mm Hg):
<
Skin Thickness (mm):
<
Insulin (mu U/ml):
<
BMI (kg/(m)^2):
<
DPF:
<
Age (years):
<
Predict