Scikit-Learn

intercept_scaling
: float, optional, default 
= 1 
This parameter is useful when

the 
solver ‘liblinear’
is used 

fit_intercept
is set to true
class_weight
: 
dict or ‘balanced’ 
optional, 
default = none 
It represents the weights associated with classes. 
If we use the default option, it means all the 
classes are supposed to have weight one. On the 
other hand, if you choose 
class_weight: balanced
, 
it will use the values of y to automatically adjust 
weights. 
random_state
: int, RandomState instance 
or None, optional, default = none 
This parameter represents the seed of the pseudo 
random number generated which is used while 
shuffling the data. Followings are the options: 

int
: in this case, 
random_state
is the seed 
used by random number generator. 

RandomState instance
: in this case, 
random_state
is the random number 
generator.

None
: in this case, the random number 
generator is the RandonState instance used 
by np.random.
solver
: 
str, {‘newton
-
cg’, ‘lbfgs’, ‘liblinear’, 
‘saag’, ‘saga’},
 optional
, default = ‘liblinear’
 
This parameter represents which algorithm to use 
in the optimization problem. Followings are the 
properties of options under this parameter: 

liblinear
: It is a good choice for small 
datasets. It also handles L1 penalty. For 
multiclass problems, it is limited to one-
versus-rest schemes. 

newton-cg
: It handles only L2 penalty. 

---

Scikit-Learn
36 
