The last sub-column of “Labels” represents three different labels according to the value of the variable <time>: “flaring start time” when <time> =“start time”, “flaring peak time” when <time> = “peak time” and “flaring end time” when <time> = “end time”.

For regression methods, the "Multitask" algorithms can use more than one label (merge cell across all the labels), the other ones can just use one label.

The clustering methods, in the prediction phase, assign the data to one of the (un-labelled) clusters estimated in the training phase.

Algorithm

Type

Code

Labels

 

 

 

flaring

N flare

flare index

largest flare

duration flare

flaring

<time>

 

 

 

 

K-Means

"clustering"

Python

yes

     

Fuzzy C-Means

"clustering"

Python

yes

     

Possibilistic C-Means

"clustering"

Python

yes

     

Simulated annealing

K-Means

"clustering"

Python

yes

 

 

 

 

 

Simulated annealing

Fuzzy C-Means

"clustering"

Python

yes

 

 

 

 

 

 

 

 

 

Hybrid Lasso

"classification"

Python

yes

     

Hybrid Logit

"classification"

Python

yes

     

SVC CV

"classification"

Python

yes

 

 

 

 

 

Random Forest"classification"Pythonyes     

MLC Perceptron

"classification"

Python

yes

 

 

 

 

 

Neural Network

"classification"

R

yes

 

 

 

 

 

Support Vector Machine

"classification"

R

yes

 

 

 

 

 

Random Forest

"classification"

R

yes

 

 

 

 

 

Linear Models

"classification"

R

yes

 

 

 

 

 

Probit

"classification"

R

yes

 

 

 

 

 

Logit

"classification"

R

yes

 

 

 

 

 

 

 

 

 

 

MLR Perceptron

"regression"

Python

 

yes

SVR CV

"regression"

Python

 

yes

yes

yes

yes

yes

Poisson Multitask

"regression"

Python

 

yes

Adaptive Poisson

Multitask

"regression"

Python

 

yes

Adaptive Lasso

"regression"

Python

 

yes

yes

yes

yes

yes

Multitask Lasso

"regression"

Python

 

yes

Adaptive Multitask Lasso

Regr.

Python

 

yes