Data Mining is a process
of applying ML techniques to dig into large amounts of data can help to discover patterns that were not immediately apparent.
Types of Machine Learning System:
Machine Learning system can be classified according to the amount and
type of supervision they get during training.
Supervised learning: In supervised learning, the training data you feed to the algorithm includes the desired solutions, called labels.
Spam Filter of your Gmail Inbox is a good example of this, as it is trained with many emails (spam & ham) and it must have to learn how to classify new emails.
A typical supervised learning task is classification.
Some important supervised learning algorithm:
- K-Nearest Neighbors.
- Linear Regression.
- Logistic Regression.
- Support Vector Machines(SVMs).
- Decision Trees.
- Random Forest.
- Neural network.
Unsupervised Learning: In unsupervised learning, the training data
unlabeled and the system try to learn without the teacher.
Some important unsupervised learning algorithm:
- Clustering.
- K-Means.
- Hierarchical Cluster Analysis(HCA).
- Expectation-Maximization.
- Visualization and dimensionality reduction.
- Principal component analysis(PCA).
- Kernel PCA.
- Logical Linear Embedding(LLE)
- t-distributed Stochastic Neighbor Embedding(t-SNE).
- Association rule learning.
- Apriori.
- Eclat.
For example Grouping the visitor of your blog into a different group like
40% of visitors are male who likes cricket and 20% of men like football. You
don’t even have to tell which group visitors belongs to, the clustering algorithm
finds those connections by themselves.
Semisupervised
Learning: In this, the algorithm has to deal with lots of unlabeled data and a little bit of labeled data.
Google Photos are good examples of semisupervised learning. After
uploading the photos, it automatically recognize if a person is present in many
different photos now you just have to tell who these people are. One label per
person and later it will help you search for photos.
Most of the semisupervised learning algorithms are a combination of
supervised and unsupervised algorithms.
Reinforcement Learning: The best example of a reinforcement learning algorithm is a robot learning how to walk. The robot observes the environment, select and perform an action and get rewards or penalties, and base on this the robot decide what to do and what not to do.
Another criterion used to classify Machine Learning systems is
whether or not the system can learn incrementally from a stream of incoming
data.
Batch
Learning: In batch learning, the system is not capable of learning incrementally
with time. As it takes a lot of time and computing resources so the entire
process is done offline. First, the system is trained, and then it is launched
into production and runs without learning anymore. It is also known as offline
learning.
Online Learning: In this
algorithm, you train the system incrementally by feeding data continuously,
either individually or by small groups called mini-batches. The best use of
this learning is for stock prices where data change rapidly.
Machine Learning system is also categorized base on generalization:
Instance-based
learning: In this learning, the system learns the old examples by heart then
generalizes to new cases using a similarity measure.
Model-based learning: In model-based learning, we build a model of these examples then use that model to make predictions.
