Unsupervised machine learning

Unsupervised machine learning Tutorial: The unsupervised ML is the technique that is used to supervise the model. We allow the model to work on its own, discover the information and deals with the unlabelled data. The algorithm will allow you to perform the complex processing task compared to supervised learning. The unsupervised ML is not directly applied to regression as unknown values and makes it impossible to train algorithm. The easy way to understand is the thinking of what to test. The task of machine is group unsorted information according to the similarities, patterns and differences without prior training of data.
Need:-

• We use unsupervised machine learning to find all unknown patterns in data.
• The methods help you to find features which can be useful for categorization.
• It is easy to get unlabeled data from a computer than labeled data which needs the manual intervention.

Types of unsupervised learning:-

Clustering:-
It is the important concept as when it comes to unsupervised learning deals with finding a structure of uncategorized data.
The clustering algorithms will process data and find natural clusters if they exist in the data.
 You can modify how many clusters your algorithms should identify and allows you to adjust the granularity of these groups.
Different types of clustering are:-

Exclusive:-

The data are grouped in such a way that one data can belong to one cluster only.
Example: -
            K-means
Agglomerative:-
The unions between the two nearest clusters reduce the number of clusters.
Example:-
 Hierarchical clustering
Overlapping:-
The fuzzy sets are used to cluster data and each point may belong to two or more clusters.
Data will be associated with a membership value.
Example:-
                  Fuzzy C-Means
Probabilistic:-
The technique uses a probability distribution to create the clusters
Example:-

• "women's glove."
• "man's glove."

Applications of Unsupervised Learning

• They help in understanding pattern which can be used to cluster data points based on many features.
• Understanding various defects in dataset would not be able to detect initially.
• They will help in mapping various items based on the dependencies of each other.
• Cleansing the datasets by removing features which are not required for the machine to learn from.

Disadvantages of Unsupervised Learning

• There is no way of obtaining the way which the data is sorted as dataset is unlabelled.
• They are less accurate as the input data is not known and labeled by the humans.
• The information obtained by the algorithm may not correspond to the output class that we required.
• The user has to understand and map the output obtained with the labels.