Machine Learning Pros and Cons

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Best Machine Learning Algorithms for Classification:

PROS CONS

Accepts every feature


Simple, easy and fast
is independent. This isn’t
Naive Not sensitive to irrelevant features
Works great in practice
always the truth.

Bayes Needs less training data


For both multi-class and binary
Classifier classification
Works with continuous and discrete
data

PROS CONS

Easy to understand Might suffer from overfitting


Easy to generate rules Does not easily work with non-
Decision There are almost null hyper- numerical data
parameters to be tuned Low prediction accuracy for a
Trees Complex Decision Tree models can be dataset in comparison with other
significantly simplified by its algorithms
visualizations When there are many class labels,
calculations can be complex
PROS CONS

Fast algorithm Doesn’t perform well with large


Effective in high dimensional spaces data sets
Support Great accuracy Not so simple to program
Power and flexibility from kernels Doesn’t perform so well, when the
Vector Works very well with a clear margin data comes with more noise i.e.
of separation target classes are overlapping
Machines Many applications

PROS CONS

The overfitting problem does not exist Complexity


Random Can be used for feature engineering
i.e. for identifying the most important
Requires a lot of computational
resources
Forest features among the all available Time-consuming
features in the training dataset Need to choose the number of
Classifier Runs very well on large databases trees
Extremely flexible and have very high
accuracy
No need for preparation of the input
data
PROS CONS

Computationally expensive testing


Simple to understand and easy to
phase
implement
Can have skewed class
KNN Zero to little training time
distributions
Works easily with multiclass data
Algorithm sets
The accuracy can be decreased
when it comes to high-dimension
Has good predictive power
data
Does well in practice
Needs to define a value for the
parameter k

http://intellspot.com

You might also like